Sales & GTM for AI Scale-ups · Andrew Wesbecher

01 · Why this exists

01 / 14

AI changed the job in November 2022. Sales and GTM leadership is now adapt or die.

I have carried a bag and built GTM engines for 26 years: TIBCO, Meraki, ThousandEyes, Lacework, Contrast, Traceable. The last three years, I have run AI and AI security scale-ups as a fractional CRO and advisor, with an AI-first GTM stack doing the ops work.

This Sales Playbook is what survived the collision: the enterprise foundations that still close seven-figure deals, and the modern layer that AI-era buyers now expect.

Nothing in it is aspirational. It is what I run now.

Who this is for

Founders and boards scaling $1M → $20M ARR who sell subscription or usage into enterprises that need a committee and a C-level signature. Every framework here has shipped inside a real plan. $1M to $25M ARR, AI tooling or AI security, is the full-time seat.

Who this is not for

A product an individual or a small team can evaluate and buy on a company card. High-volume, low-ACV, self-serve PLG is a different job. This playbook is not that job.

The exception

Open source can still be the land. Snyk got individual users in. Converting a small OSS deployment into an enterprise contract still takes a sales methodology: champion, economic buyer, paper process. That is in scope.

Next · 02 THE SHIFT

Contents

02 · The shift

02 / 14

Your buyer moved. Three years ago.

01Research happens in the answer box. Buyers ask ChatGPT, Claude and Perplexity who the leaders are before your SDR ever calls.
02Adoption starts at the practitioner. Engineers install the OSS and form opinions first. The exec meeting only ratifies it.
03Proof means production. A slideware demo earns nothing. Buyers expect a POV in a prod-like environment inside days, not a quarter.
04Procurement grew an AI checklist. Model risk, data boundaries, agent governance. Meet it in week 2, or lose to it in week 11.
05Categories form in months. A model release can commoditize your moat mid-quarter. Speed of iteration is the durable advantage.

The job

Get into the model's answer, the practitioner's terminal, and production, in that order. The rest is execution.

02.1

The GTM delta, in one table.

Same disciplines. Different mechanics. The right-hand column is where AI-native companies actually operate.

Discipline	The old playbook	The 2026 playbook
Demand capture	SEO, SEM, gated PDFs	AEO: be the answer the model gives. Ungated technical depth
Entry point	Top-down exec outreach	Bottoms-up practitioner adoption; exec alignment engineered later
Proof	Custom demo, 90-day pilot	Self-serve trial + scoped POV with signed success criteria
Technical selling	SE demos over Zoom	FDE builds alongside the customer, in their environment
GTM operations	SDR armies and ops headcount	Headless agent workflows; humans on judgment and calls
Content	Quarterly gated whitepaper	Continuous technical content practitioners actually share

The tell

The left column still works in legacy categories. In AI, it reads as a vendor who has not used their own product.

02.2

What 26 years says still wins.

Six disciplines that predate the model era and survive it intact.

01

MEDDPICC

Metrics, economic buyer, decision and paper process, pain, champion, competition. Deals get inspected, not narrated.

02

Command of the Message

Value framing the economic buyer signs, and discovery that earns the right to deliver it. AI buyers still buy outcomes.

03

Manufactured pipeline

PG quotas next to ARR quotas, activity math, coverage inspected weekly. Hope is not a pipeline source. It never was.

04

POV discipline

Pre-agreed success criteria, frozen scope, readout to the economic buyer. The POV got faster, never optional.

05

Coaching on cadence

Call review, scorecards, 60-day ramp gates on every new hire. AI does not fix a rep who cannot run discovery.

06

Forecast as arithmetic

Stage exit criteria in CRM. Commit, upside, closest-to-pin. Judgment sits on top of math, never instead of it.

The point

AI changed the how. It did not repeal the math. Every module after this one sits on top of these six.

02.3

MEDDPICC is old. Complex enterprise still dies the same eight ways.

The acronym is overused. The inspection is not. AI changed how buyers research and how reps log the answers. It did not repeal the committee, the paper, or the champion who has to carry the deal one floor up.

MEDDIC was built at PTC in the 1990s. Paper Process and Competition were added because deals that looked qualified still died in legal and in a bake-off. I run the eight letters as a deal-inspection board, not as a religion.

The eight letters

Letter	Inspection	The question	Red flag
M · Metrics	The number the economic buyer already tracks. Not a vendor ROI slide. In AI-native deals that number is usually usage, an outcome, or a platform fee with a governed user count, not seats.	What KPI shows up on the board report if this works, and what is the baseline?	"better efficiency."
E · Economic buyer	The person who can spend the money. In AI deals this is usually one floor up from the engineer who installed it.	Who signs at this dollar amount, and have we met them?	"my boss will approve it" and no meeting on the calendar.
D · Decision criteria	The scorecard they will use. In 2026 that list includes model risk, data boundaries, and agent governance. Get those written in week 2.	What do you have to prove, and who wrote the list?	Criteria only a practitioner cares about, with no economic-buyer language.
D · Decision process	The sequence after the technical yes. The practitioner installs. The platform team scopes a POV. The economic buyer contracts.	After the technical team says yes, what happens next, and who is in that room?	"we just buy it once we like it."
P · Paper process	Legal, security, procurement. The step most verbal-yes deals die in. Start it in parallel with the POV, not after the readout.	Whose paper, how long is security review, and is there a vendor-onboarding queue?	A close date set before anyone has asked legal.
I · Identify the pain	The cost of doing nothing, in their units. Production proof, not a slide.	What happens if you ship nothing this quarter?	Pain that cannot be numbered.
C · Champion	Four tests: influence, manager or above, can remove a barrier, puts their own name on the line. Fail any one and they are an advocate. Want a technical champion and a business champion. A fan who likes the product is a coach.	What did they do this week to move the deal, and which of the four tests do they fail?	They will not introduce you to the economic buyer, share the real process, or spend political capital.
C · Competition	The other vendor, the internal build, and doing nothing. Categories form in months. A model release can erase a feature mid-quarter.	Who else is in the bake-off, and what does doing nothing cost them?	"we are not looking at anyone else" on a six-figure platform deal.

Diagram 1 · Eight-letter inspect rail

The board. A deal is only as real as the letters you can prove.

Diagram 2 · Two floors

Same account, two floors. The practitioner installs. The economic buyer contracts.

Diagram 3 · Inspect board

Example inspect board

Proven Partial Empty

Not a real customer. Marks show how I inspect a deal, not a published score.

M

Metrics

Proven

E

Economic buyer

Empty

D

Decision criteria

Partial

D

Decision process

Empty

P

Paper process

Empty

I

Identify the pain

Proven

C

Champion

Partial

C

Competition

Empty

Empty letters keep a deal out of commit. This one is not ready for Stage 3.

How it runs

01One CRM field per letter. On every Stage-2+ opportunity. Module 13 already inspects the score weekly.
02No Stage 3 without an economic buyer meeting on the calendar. Already a tripwire in 13.1.
03The AI stack parses calls and logs notes. Module 05. The eight letters are the schema. A human still inspects. AI does not get to call a coach a champion.
04Paper process starts when the POV is scoped. Not when legal emails you.
05Velocity deals do not get the full eight. $30-60K, 30-45 days. A named buyer and a real initiative are enough. MEDDPICC is for the field engine.

The point

Score the deal. Do not narrate it. Why change and why now come first. The eight letters inspect what is already in motion.

Next · 03 MODERN MOTIONS

Enterprise and velocity run on different math. They land in the same forecast.

Contents

03 · Modern motions

03 / 14

Two engines, one capacity model.

Enterprise and velocity run on different math. They land in the same forecast.

Field engine · Enterprise + Growth

ACV $80-400K
by tier; 90-120 day cycles
MEDDPICC + POV
run through the nine stages below
Anchors
meeting → POV 4-5% · POV → win 75-85%
Staffing
AE + BDR pods, SE at 1:3, FDE on strategic

Velocity engine · PLG + Community

$30-60K entry
AI-native and mid-market; 30-45 days
PQL-gated
seats + prod-adjacent usage + findings
Anchors
PQL → win 15-25% · reverse trials, 14 days
Staffing
one velocity pod works the whole flow

Velocity, defined

Velocity here means OSS or community land that still has to become an enterprise contract. It does not mean credit-card self-serve. Snyk is the pattern.

Rule

One capacity model, one forecast. Separate staffing, separate math, one number. Module 10 builds the model.

03.1

The nine stages have not changed.

Exit criteria, not activities. A stage advances when the criterion is met, not when the rep feels good.

#	Stage	Close %	Exit criteria	Advance →
0	Qualify	0%	ICP fit, BANT, a real initiative	50-55%
1	Discovery	10%	Pain quantified, champion identified	~30%
2	Scoping	20%	Technical fit, budget confirmed, POV criteria drafted	~30%
3	EB Alignment	35%	Economic buyer met, reverse timeline agreed	~90%
4	POV	50%	Success criteria signed, scope frozen	~80%
5	Win the Decision	75%	POV readout to the EB, verbal to proceed	~90%
6	Negotiate & Close	90%	Proposal, mutual action plan, paper process mapped	~95%
7	Closing	99%	Signatures in motion	100%
8	Closed Won	100%	Order form executed	–

Stage 4 is shaded because that is where the deal gets decided.

3-4%

Meeting → win

Nothing structural changed. The POV now runs in days, not weeks, so the whole table spins faster.

Stages vs the CRM

Most people treat enterprise as a five-step CRM path. Intro, demo, proposal, contracting, close is a forecasting tool. It is not how a buyer decides. The table above is the actual run: about fifteen steps when you count entry, proof, paper, and the post-signature expand. Forecast in CRM. Run the deal on the nine stages.

03.2

The FDE motion: selling by building.

The forward-deployed engineer is a sales motion with an engineering cost structure. Scope it like one.

What it is, when to run it

Forward-deployed engineers embed with the customer, build the integration in their environment, and map expansion. The build is the proof, and the customer keeps it.

Deploy on agentic or complex products, design-partner phases, and lighthouse logos where the demo cannot carry the sale. Never on a deal an SE can close alone.

Palantir normalized it. OpenAI and Anthropic industrialized it. AI-native buyers now expect a builder in the room, and they can tell inside an hour if you sent one.

The rules that keep it a sales motion

·Scope FDE time like a POV: success criteria, exit date, EB readout
·FDE cost is CAC, not services revenue. Watch the margin line
·1 FDE per 2-3 strategic deals in flight; classic SE 1:3 for growth
·FDE builds twice, product ships once. Weekly roadmap loop
·Exit criteria: repeatability. Without it you built a consultancy

Scope

Like a POV

criteria, exit date, EB readout

Cost

CAC, not services

watch the margin line

Exit

Repeatability

without it you built a consultancy

03.3

Bottoms-up and top-down, same account.

The practitioner installs it. The economic buyer contracts it. Four steps, in order.

01Practitioner installs. OSS or free tier, no gate, generous limits. Treat the giveaway as marketing spend, not lost revenue.
02Champion forms. Usage becomes evidence: findings, saved hours, incidents avoided. The AI stack spots the signal.
03Platform team scopes a POV. Success criteria signed in writing, not implied on a call. Economic buyer aligned on a reverse timeline from go-live.
04The EB contracts the platform. Land $50-120K. Security review done once at the platform, then inherited by every team after.

Time-to-first-value is a commercial choice.

I watched cloud security split on this. Depth of the runtime graph versus how fast a CISO sees the first risk that matters. Agent versus agentless is a sequencing decision, not a religion. Show the risk that is already on fire, then earn the right to go deeper.

The POV clock in 03.5 is the same rule. Two or three days. If first value takes a quarter, you are selling architecture.

Expand vectors

Workloads → environments → modules → teams. Expansion runs on triggers, not on an account manager's mood. NRR 105-110% landing year, 115-120% at scale.

Pricing architecture

Value metric: per governed workload, not per seat

Floors by tier; exceptions are CEO-only

Annual prepay default; multi-year needs cash up front

Free tier generous; enterprise buys governance

The point

Free adoption is the demand gen line item. The contract is signed one floor up, by someone who never touched the product.

03.4

Qualify the company, then the committee.

Five company prerequisites. Three buying rings. Nobody prospects outside them.

ICP prerequisites. All five, or it is not ICP.

01AI agents or LLM apps in production, or within two quarters
02A named owner of the AI platform on the org chart
03Security budget with an AI line item, this fiscal year
04A compliance driver: EU AI Act, SOC 2, model risk management
05Meaningful model and inference spend, growing

1-2 people

Economic

CISO, with the CFO co-signing. Buys risk reduction and audit readiness

3-6 people

Product decision

VP AI Platform · Dir AI Infra. Owns the platform your product lands on

8-15 people

Evaluators

AI + security engineers · MLOps. Daily users. Opinions form in the free tier. Scored weekly by the AI stack. Nobody prospects off-list.

Primary axis

VP AI Platform → Dir AI Infra → AI Engineers. Sell where the pain lives. Contract where the budget lives. The POV bridges the two.

03.5

Running the complex deal.

Operator notes that sit under the nine stages. Field engine only. Velocity deals do not get this full run.

Open the door

01Pincer the first meeting. Two valid entry points only: the economic buyer and the person one step below. Founder works the top. AE works the N-minus-one. Same account, same week, both tracks live.
02Pitch the alpha, not the problem. Executives ignore generic pain. Lead with the unfair advantage they get only through you. Time edge, information edge, risk edge. Savings language dies in the inbox.

Entry rule

If the first touch is not the EB or the N-1, it is not a first touch. It is research someone else will ignore.

First call and proof

03The intro call is intel, not a pitch. Keep it short and informal. Let them go first. Ask what has to be different next year. No recording, no slides, no demo. What they say becomes the frame for every later call.
04Withhold the demo until you co-author it. Run a fifteen-minute prep with the champion: which features, which questions, which attendees. The group demo should feel built for that room.
05Demo the 20% they care about. Full-product demos unravel deals. Extra features become cost they will not use. Let curiosity pull you wider. Do not lead with the catalog.
06Text the champion inside five minutes. Raw debrief before the room settles on polite consensus. Who you lost, who you landed, who needs another twenty minutes. Find the deal-killer early.

Pilot and champion risk

07Plan for a 25% to 35% win rate on qualified field opps. Half is a fantasy. Enterprise needs time, maturity, and alignment you do not control. A quiet share of losses come back inside a year. Cutting price to raise win rate trains the market against you. Executives talk.
08Time-box the POV hard. Two or three days. Three or four hand-picked daily users, not the C-suite. Shared success definition. Longer only when deep integration is required, and then charge for it and credit the fee if they proceed.
09A quiet champion means the deal is slipping. Their job is internal navigation. If they go dark, something moved. Find it before procurement and legal finish the kill.

Win-rate note

Field engine, qualified opportunities. Velocity math is different. Raising win rate by discounting is not a strategy. It is a price leak.

Close and expand

10A technical yes is not a deal. The POV readout is a technical commit. It is not an economic commit. Do not send a quote as if you already won. The economic buyer still has to say yes, on a date, with a champion who can take the room.
11The champion presents, you do not. Build a short deck in their language. They walk it upstairs. If you have to present it to the economic buyer yourself, you do not have a champion.
12Sales is project management after a strong POV. Send the champion a forwardable package: agreed timeline, price tied to a signature date, and a kicker if they hit it. That is how procurement gets looped without you becoming the internal PM.
13Expansion starts the day you sign. Five minutes of celebration, then back to pipeline. The real prize is year-two: one team to several, six figures to mid six or seven. Land is the first invoice. Expand is the business.

Handoff

Signature is not the finish line. Module 11 prices the next stage. Module 10 staffs it. The mutual action plan already names the expand seats.

Discovery is a handshake, not a demo.

A discovery call exists to decide whether both sides should invest more time. Validate need, current state, pain severity, prioritization, and fit. Founders default to "let me show you the product." That is how you burn calendar on accounts that never had the problem, never felt it, or never ranked it. Frame the call up front: we are figuring out whether we can help. Hard sell is not the job on this call.

How I run discovery.

Prep before the call. Calendar invite says discovery, not demo. Light intake on the booking form is useful friction. Block thirty minutes to research the company and the person, hypothesize the pain, and queue only the content you might need if fit appears.

On the call, in order: two to three minutes of real rapport, then an agenda with permission (questions, possible short content, next step or part as friends), then discovery that validates research and goes two or three levels deep on pain. Content or demo only after you know what to show, and only the slice that maps to what they said. Close on a concrete next step or a clean no. "I will send some info" is not a close.

The next move is a record.

Every Stage 2+ opp has one card: who has the ball, the action, the document in motion, the date, the unresolved concern, and who sees it next. If you cannot name the person, the action, and the date, the deal is sitting in the pipeline while you tell yourself you are following up. Wednesday inspects that card.

Build the price before you show it.

Most price objections are finished before the fee appears. The number needs three things under it: why you have a right to solve this, what the outcome or the delay is worth in their units, and what is in scope, out of scope, owned by whom, and how progress is governed.

A narrow promise on a consequential outcome is easier to defend than a catalog meant to justify the fee. A land package or a paid POV is designed before the objection. It is not a discount you invent when they push back. If the economics no longer work, the answer is later or no.

The champion marks up V1.

Do not send last quarter's POV with a new logo. Current condition, commercial outcome, in and out, milestones, their work, yours, timing, what happens next. V1 is not a yes-or-no. Ask what feels inaccurate, risky, hard to defend, or incomplete. Revise through their lens before it goes one floor up. Their fingerprints are how they stop forwarding it and start defending it when you are not in the room.

Signed is not collected.

When the buyer has produced enough signal, send the finish line as one package: final scope or order form, paper path, redline instructions, signature path, first invoice, dates for both, onboarding date. Track signature and first invoice as separate handoffs. They should never be ready to move and still be waiting on you for the next document.

How this sits on the nine stages

Open the door before Qualify. Intro and proof run Discovery through POV. Pilot and champion risk sit on POV. Technical commit is Win the Decision. Economic commit and the champion deck sit on Negotiate & Close, then the land-and-expand loop in Module 11.

03.6

How I run a patch.

One AE book, for a year. Written once. Updated when the rank moves. Field engine. Velocity books get a shorter list, same three artifacts.

Three artifacts

01Ranked account list. Every name in the book gets a rank. 1 is this-quarter focus. 2 is ICP this fiscal year. 3 is keep warm. Rank is a decision, not a CRM field. The list lives where the AE actually works.
02One-page plan on the top 15. Why this account, who is the economic buyer and the N-1, what land looks like, what expand looks like, the next meeting. One page. If it needs a second page it is not a plan.
03Handover note when the seat turns. Why we won or why we stalled. Who the champion is. Paper status. Next 90 days. The new AE should not reconstruct the deal from email.

Patch rule

No ranked list, no forecast you can defend. No one-pager on a Rank 1, you are hoping. The GTM engineer can own the system. The AE owns the rank.

When it gets written

Once a year, then when the book changes. New AE, new territory, or a Rank 1 that went dark. Module 10 sizes the seat. This page is how the seat is run.

03.7

Project mode.

The AE who hit 200% two years running as we went $1M to $10M had one skill the rest did not. He treated every live deal as a project. Field engine.

Two buckets

01Project management. A work-back plan built with the champion, starting from the value-delivery date. Revisited at every meeting. The buyer can see the finish line.
02Project formation. Re-listen to the calls. Read the notes. Tailor the next move. The job is to say why this must become a project, name the champion, or cut it.

Bucket rule

Most follow-ups exist to extract an answer for the forecast. His felt like project management. If a deal will not become a project, it is not a deal.

How he ran it

03Multi-thread for consensus, not a title. Decisions happen in committees. People below the line of power are often as important as the exec you want. Wide and up. Arm the champion so they look like the hero when that exec walks in. Going around them backfires.
04Follow-ups recap the project, not the next task. Every note lists completed and open milestones through value delivery. Buyers stare at the goal, not a task they will postpone. POC: done. Commercials: done. Executive review: a date.

Hire for it

Ask for a live deal and which bucket it is in. If they cannot say, they are chasing answers. The bar is on the AE profile.

How this sits on 03.5

03.5 already says sales is project management after a strong POV. This is the skill that AE ran underneath those beats. The champion deck in beat 11 is useless if the follow-up is still "just checking in."

Next · 04 PICK THE MOTION FROM ACV

The GTM motion is a function of average contract value. Same outbound motion, two different businesses.

Contents

04 · Pick the motion from ACV

04 / 14

Pick the motion from ACV.

The GTM motion is a function of average contract value. I do not start from the playbook the last company ran.

A booked meeting through an SDR costs $900 to $1,400 as a working band. Meeting cost divided by ACV is the close rate that meeting needs just to pay for itself. At $5k that is 18 to 28 percent of meetings. At $100k it is under 2 percent. Same outbound motion. Two different businesses.

The figures below are operating rules of thumb. They are not a client result and not a study.

Four ACV bands. Channel, personalization, team, metric.

Under $5k

Self-serve

Channel: Content, SEO, product-led, referrals
Personalization: None
Team: No human in the seat
Metric: Activation and inbound

$5k to $25k

Automated outbound

Channel: Mapped TAM, sequenced email
Personalization: One message per segment
Team: One SDR or an agency
Metric: Reply rate

$25k to $100k

Phone and multi-thread

Channel: Phone first, email as support
Personalization: Named accounts, 3 to 5 contacts
Team: AE plus a caller
Metric: Pipeline per interested reply

$100k and up

ABM

Channel: Every channel, including ads
Personalization: Named owner, research doc, months-long sequence
Team: Named owner per account
Metric: Coverage and engagement

Under $5k

Outbound barely pays. Close one meeting in three to five and you have covered the meeting, not the month. Run content, SEO, product-led, and referrals. Self-serve is the motion.

If you still send mail, it is volume, zero personalization, a one-line free offer, and no human in the seat. The math only works past 10,000 sends a day. The free offer is the whole email.

$5k to $25k

Automated outbound pays here. Map the TAM, segment by pain, write one message per segment, and launch 15 to 30 variants at once. After 1,000 sends each, kill the bottom two thirds. Warm-call every positive reply the same day.

One SDR or an agency. Not a team. Keep cost per meeting under $1,000 or the model breaks. Reply rate is the leading indicator.

$25k to $100k

The phone is the primary channel. Enterprise buyers sit behind Barracuda and Proofpoint. Email alone will not reach them. Multi-thread every account: 3 to 5 contacts, not one. Named accounts, buying signals, warm intros where they exist.

Reply rate stops mattering. Judge the campaign on pipeline per interested reply.

A no-show at this ACV is a five-figure miss. The pre-call sequence matters more here than in any other band.

$100k and up

This is ABM. TAM is usually under 5,000 accounts, and you know most of them by name. Every account gets a named owner, a research doc, and a sequence that runs for months, on every channel, including ads.

Personalization pays only here, because the contract can carry it. Cycles run 6 to 18 months. Half the book is locked in 36-month competitor contracts. Measure coverage and engagement, not meetings. You want to be the name they search the month that contract is up.

Close rate a booked meeting needs to pay for itself

Meeting cost divided by ACV. The light column is the close rate at $1,400. Blue is the $900 to $1,400 band. Illustrative math, not client data.

At $5,000 ACV the meeting must close 18 to 28 percent of the time to cover itself. At $10,000, 9 to 14 percent. At $25,000, 3.6 to 5.6 percent. At $50,000, 1.8 to 2.8 percent. At $100,000, 0.9 to 1.4 percent.

Close rate equals meeting cost divided by ACV. $900 is the low end of the band. $1,400 is the high end.
ACV	At $900	At $1,400
$5,000	18%	28%
$10,000	9%	14%
$25,000	3.6%	5.6%
$50,000	1.8%	2.8%
$100,000	0.9%	1.4%

The cost of the wrong tier

Get the tier wrong and you spend personalization the ACV cannot carry, or you send a six-figure buyer a two-line email. The matching engines sit on Pipeline Gen.

Next · 05 THE AI-NATIVE ENGINE

Every tool below is run by Claude, through official MCPs or Claude in Chrome. I prompt; the stack executes.

Contents

05 · The AI-native engine

05 / 14

The stack I pay for but never log into.

Every tool below is run by Claude, through official MCPs or Claude in Chrome. I prompt; the stack executes.

Know · Data, research, record

Attio
CRM + call recording. System of record

Clay + Apollo
Enrichment, ICP + persona sourcing

Parallel
Account research + web scraping

AthenaHQ
AEO: visibility in AI answers

One interface: a prompt

Claude

The control plane

Official MCPs · Claude in Chrome
Cowork agents · skills · projects

Slack + CRM
Where insights and coaching land

Act · Execution channels

Origami + Instantly
Outbound email sequencing

HeyReach
LinkedIn outbound campaigns

Nooks
AI autodialer + call sessions

AgentMail
Agent-owned inbox: inbound SDR

The point

Ten paid tools. Zero logins. A human approves every external send. Everything else runs headless.

Contents

05.1

A prompt is the new login.

Four prompts I actually run. Each one ends in a system I never opened.

“Move Acme to Stage 3 in Attio, log the EB meeting, set next steps.”

→CRM updated, tasks created. Attio never opened.

“Find 50 Series B AI companies hiring security. Enrich in Clay, draft N=1 copy.”

→List, enrichment and copy live in Origami within the hour.

“Launch the HeyReach campaign for the conference follow-up list.”

→LinkedIn touches running by lunch, replies triaged back to me.

“Parse today's call recordings. Coach the team in #sales, log notes to CRM.”

→Every recording lands in a Claude project; insights hit Slack.

Always on, no prompt needed

Meeting intelligence. Every call parsed. Coaching to Slack, notes and next steps to CRM.

Inbound agent. Claude plus AgentMail: replies in minutes, qualifies, books the meeting.

Copy engine. Cadence copy drafted by Claude skills, tuned on what converts.

The operating model

Claude is the interface. The stack is the execution layer. One operator runs what used to take a pod.

The requirement

An official MCP, or Claude in Chrome. That covers every tool above, today. No custom integration work, no engineering ticket.

05.2

AEO is the new demand capture.

Buyers ask an assistant before they run a search. Five moves, four measures.

01Entity-clean docs and llms.txt. Make the product legible to models: crisp category language, structured docs, consistent naming.
02Comparison and alternatives pages. Answer the questions buyers actually ask assistants: X vs Y, best tools for Z.
03Community surface. Reddit, HN, Stack Overflow, GitHub. Models source opinions where practitioners argue.
04Review corpora. G2 and peer reviews are retrieval fodder. A 30-review gap is a demand-capture gap.
05Ungated technical depth. Models cite what they can read. Every gated PDF is invisible to the answer box.

Measure it

Monthly share-of-answer audits across ChatGPT, Claude and Perplexity

AI-referral sessions tracked as their own channel

Branded-query lift measured after content ships

Percent of inbound that says an assistant recommended you

The shift

Rank #1 on Google and still lose the deal. The model handed your competitor a buyer who never searched.

05.3

The headless workflow catalog.

Eight jobs that used to be headcount. Every external send still passes a human.

Workflow	Stack	Cadence	Human gate
Signal sensing: funding, hiring, tech adoption	Claude Cowork	Daily	None: read-only
Enrichment waterfall + account scoring	Clay + Apollo	On signal	None
Outbound cadences, drafted N=1	Claude → Instantly	Daily	Every send approved
LinkedIn campaigns + reply triage	Claude → HeyReach	Daily	Every send approved
CRM hygiene + stage enforcement	Attio / Salesforce	Nightly	Exceptions flagged
AEO, SEM and content pipeline	Claude Cowork	Weekly	Publish approved
Event ops: lists, pre-books, follow-ups	Claude + CRM	Per event	Sends approved
Win/loss tags + forecast prep	Claude + CRM	Weekly	CRO judgment

The hire

One GTM engineer runs this entire table. That role replaces the ops admin as the first RevOps seat, by roughly $3M ARR.

05.4

What the engine buys you.

Not fewer people. The same people, spending their hours where judgment actually pays.

100%

Account coverage

Every tiered account researched continuously, not the top 20%.

Under 5 min

Speed to lead, 24/7

PQLs and inbound routed, researched and answered around the clock.

~3 FTE

SDR capacity added

Without the headcount. Pods spend their hours in live conversations.

15 → 2 min

Research time

Brief, angle and opening line waiting in the CRM before the rep dials.

N=1

Personalization

Copy drafted per account from live signals, never from merge tokens.

Nightly

CRM hygiene runs

Stages, next steps and close dates enforced before the forecast call.

Not a concept

My daily driver.

This is the production stack I run across portfolio companies today. Claude Cowork orchestrates Attio, Clay, Apollo, HeyReach and Instantly on headless daily jobs, with human-approved sends.

This Sales Playbook was researched, written and designed by that same stack.

Attio · Clay · Apollo · HeyReach · Instantly · Nooks · AthenaHQ

The trade

Headcount stays. The hours move to judgment, relationships and closing.

05.5

Find the revenue leak before you buy another AI tool.

Assessment first. Tools second. No vendor case study.

What usually breaks

·Teams get faster at tasks that do not move win rate, cycle time, or pipeline contribution.
·Sales and marketing still run on separate systems, so handoff loses the context that earned the meeting.
·Buying behavior often happens before identity. By the time the dashboards catch up, the cycle moved.
·Point tools multiply. Nobody shares one layer of intelligence that answers: why is this deal stalling, which accounts matter, where did conversion drop.
·Time saved is not a board metric. Win rate, deal velocity, and pipeline contribution are.

Operating cycle

01Reveal. Map the revenue leaks in the human workflow before you commit to a build. Name the leak in one sentence.
02Architect. Design around one revenue question and the minimum shared data that answers it. Prove a narrow use case before you connect everything.
03Mobilize. Habit, not launch day. One or two workflows until they stick. Put the assistant where the team already works. Reduce friction before you expand scope.
04Prove. Measure the leak you named, before and after, in numbers that hold in a board pack. Then run the cycle again. Each loop compounds.

AI does not replace the relationship, the creative call, or the leadership that aligns sales and marketing. It replaces the hours spent assembling context. Start with one leak you can name.

Next · 06 SALES QUALIFICATION AND PAIN DISCOVERY

Seven stages, one question set per stage. Each stage either qualifies a deal criterion or it doesn't belong on the call.

06 · Sales qualification and pain discovery

06 / 14

Sales qualification and pain discovery.

Seven stages, one question set per stage. Each stage either qualifies a deal criterion or it doesn't belong on the call.

Most discovery calls fail the same way. The seller asks about the problem, hears "yes, that's an issue," and goes straight to the demo. Nobody priced the pain, nobody asked who signs, and the deal shows up in the forecast with no date that means anything.

I run discovery in a fixed order. Situation, then pain, then what the pain costs, then whether it gets fixed this quarter. Only after that do we talk about the fix, the decision process, and a date. If a stage comes back empty, the deal stays where it is.

The questions below are adapted from Caliber's 50 Proven Sales Questions That Sell. I cut the list to the ones I'd actually ask, reworded them, and sorted them by stage.

Figure 1 · Discovery stage flow

01SituationContext and trigger
02PainIdentified pain
03ImpactMetrics and cost
04UrgencyPriority and timing
05VisionValue and champion
06DecisionDecision process and budget
07CommitmentClose plan

Conceptual sequence, not data. Highlighted stages are gates: no urgency, no proposal; no agreed date, no commit.

Figure 2 · Pain-to-impact ladder

Conceptual, not data. Each follow-up moves the buyer one rung up. Most deals that slip were left on rung two.

01

Situation

Qualifies: Context and trigger. Find out what changed before I say anything about the product.

What made this call worth your time today? You get plenty of these.Caliber p2 Pain #8
How are you handling this workflow today, step by step?Caliber p6 Demo #1
Since we last spoke, what's changed on your side?Caliber p5 Progression #1

02

Pain

Qualifies: Identified pain. Get one named problem, in the buyer's words, with an example.

What's the one problem with this that keeps coming back?Caliber p2 Pain #3
When you say that, what do you mean? Walk me through it.Caliber p2 Pain #9
Give me the last time it happened.Caliber p2 Pain #10
How long has this been going on? Every week, or now and then?Caliber p2 Pain #11
What have you tried already? What worked and what didn't?Caliber p2 Pain #12
Is this the one to focus on, or is something else bigger right now?Caliber p2 Pain #6

03

Impact

Qualifies: Metrics and cost. Turn the pain into a cost the business already feels.

What does this break downstream?Caliber p3 Impact #2
Which teams feel it, and how?Caliber p3 Impact #4
Rough guess: what has it cost so far?Caliber p3 Impact #5
And what does that lead to? How often does it repeat?Caliber p3 Impact #10-13
How much of this lands on you personally?Caliber p3 Impact #6

04

Urgency

Qualifies: Priority and timing. Find out whether this gets fixed this quarter or someday.

Where does this sit on your list? Top three? Top ten?Caliber p2 Pain #5 / p4 Solution #1
If nothing changes for six months, what hurts most?Caliber p2 Pain #1-2
Why fix it now instead of next year?Caliber p3 Impact #7

05

Vision

Qualifies: Value and champion. Let the buyer describe the fix and what it's worth.

What would it take to fix this, and what would that change for you?Caliber p4 Solution #2
Which outcome matters most to the business, and what's it worth?Caliber p4 Solution #6
How much of the problem you described does this solve?Caliber p6 Demo #6
This isn't landing. Where am I off?Caliber p6 Demo #10
Why does this matter to you personally?Caliber p4 Solution #7

06

Decision

Qualifies: Decision process and budget. Map who signs, in what order, and with what money.

Walk me through the steps your team takes to get to a confident yes or no.Caliber p5 Progression #2
Who's in those steps, and what does each of them care about?Caliber p5 Progression #3
How will this get funded?Caliber p5 Progression #4
What could stall this?Caliber p5 Progression #6

07

Commitment

Qualifies: Close plan. Agree on a date and the conditions for hitting it.

What has to be true for us to sign by [date]?Caliber p5 Progression #5
Is there any reason not to move forward now?Caliber p5 Progression #7

Figure 3 · Question-to-criteria matrix

Stage	Pain	Metrics	Urgency	Champion	Decision process	Budget	Close plan
01 Situation
02 Pain
03 Impact
04 Urgency
05 Vision
06 Decision
07 Commitment

My mapping, not data. Filled: the stage's main job. Outline: it also feeds that criterion. Every criterion has at least one filled cell, so a full run leaves no blank in the deal record.

Figure 4 · How Caliber's list is weighted

Business pain12

Negative impact13

Solution7

Deal progression7

Demo10

Count of numbered questions per category in Caliber's PDF (49 numbered; the title says 50). Over half sit on pain and impact. I kept 28 for the framework above.

The point

Discovery is qualification. If I can't fill the matrix from the buyer's own answers, the deal isn't real yet, however good the demo went.

Source: questions adapted and reworded from Caliber, 50 Proven Sales Questions That Sell. Stage order, criteria mapping and diagrams are mine. Figure 4 counts Caliber's list; no other figure on this section is data.

Next · 07 WHAT ACTUALLY PREDICTS QUOTA

Activity gets you in the game. These seven metrics tell you how to win it.

Contents

07 · What actually predicts quota

07 / 14

What actually predicts quota.

Dials, emails, and meetings booked matter. On their own they leave you flying blind.

Everyone tells new sellers to track activity. Dials. Emails. Meetings booked. Those matter. They get you in the game.

If that is all you watch, you are flying blind on what actually predicts whether you hit quota. Here is what I look at, quantitative and qualitative.

Two groups. Activity, then the seven that predict quota.

Activity: gets you in the game

Dials
Activity
Emails
Activity
Meetings booked
Activity

What predicts quota

Pipeline velocity
Quantitative
Stage conversion rate
Quantitative
Average deal size
Quantitative
Next step quality
Qualitative
Champion engagement
Qualitative
Win/loss patterns
Qualitative
Total time on calls
Quantitative

Pipeline velocity

How fast are deals actually moving? A full pipeline that is not moving is just a wish list.

Stage conversion rate

Where are deals dying? If you are great at discovery but cannot get to proposal, you have a gap.

Average deal size

Are you consistently underselling? Most new reps leave money on the table because they chase easy wins over right-fit deals.

Next step quality

Not just whether you have a next step. Is it meaningful? “I’ll follow up next week” is not a next step. A scheduled call with a new stakeholder is.

Champion engagement

Are the people inside your deals actually advocating for you? If the champion goes quiet, the deal is in trouble before the CRM knows it.

Win/loss patterns

Why did you lose? Track it and reflect on it. The reps who compound fastest treat every loss as a data point.

Total time on calls

How do I stack up against the rest of the org on time with customers and prospects? I want to be near the top. The time still has to be well spent.

How to win it

Activity gets you in the game. These metrics tell you how to win it.

Next · 08 AI CHANGES THE MATH

What I do with AI on qualification, discovery, and the deal. The stages above stay.

Contents

08 · AI changes the math

08 / 14

AI changes the math.

Qualification, discovery, and deal execution still have a person in the chair. The buyer, the prep, and the inspection changed.

The buyer already ran discovery on you. Per G2's 2026 research, 52% of buyers say AI leads their initial vendor discovery, 75% verify what they see by default, and 76% say a human still anchors negotiation. I treat the first call as a check against the stages in sales qualification and pain discovery, not a product lecture.

If Situation or Pain comes back empty, the deal stays where it is. AI does not fill those cells for me.

The Bridge Group's 2026 AE study is the quota picture I trust more than a vendor slide. 48% of AEs hit quota, down from 51% in 2024. Orgs they coded as high AI-engagement sat at 57% at quota. Low-engagement sat at 39%. That is a survey correlation, and they say so. A plurality also rated 10 of 11 AI use cases "hit or miss." I do not take 57% as a promise.

Bridge Group 2026 · AEs at quota, by AI engagement

High AI engagement57%

Low AI engagement39%

Observational survey, not a controlled lift. Bars scale to 60%. Source: The Bridge Group, AE Models, Motions & Metrics 2026.

On a live deal I use Gong Labs' three rungs, and I label them as Gong's. Across more than a million opportunities at 1,418 organizations, Gong Labs (updated March 6, 2026) measured win rates 26% higher when AI informed the deal, 35% higher when AI guided it, and 50% higher when the rep completed the AI to-dos. Inform, then guide, then make the rep finish the work. I run that against the same MEDDPICC inspection I already do.

01InformGong: +26% win rate when AI briefed the deal. I have the last three calls summarized before I sit down, then I ask the discovery questions myself.
02GuideGong: +35% when AI guided the deal. I use that as a missing-criterion prompt in the review, not as a score I read in the forecast.
03Finish the workGong: +50% when the rep completed the AI to-dos. The lift sat on work the rep actually did. A suggestion list that nobody closes is theater.

Those three numbers are Gong's, on Gong customers. Vendor data. Clari's 2026 discovery note is the practical habit underneath: discovery quality drives forecast accuracy, and you review the prior call before the next one.

Gong's State of Revenue AI 2026 press release says 7 in 10 enterprise revenue leaders now trust AI to make decisions, and US companies using AI to forecast rose 50% year over year. AI is an input to my commit call. It is not the call. What actually predicts quota still decides whether the deal is real: next step quality, champion engagement, stage conversion. A clean CRM row is not a next step.

Jason Lemkin wrote that 1.25 humans and 20 agents closed 140% of the prior all-human year, and then said he is not sure that is the real story. Read his own caveat: concentration of leads in top closers, and an AI market tailwind, sit next to the tools. I separate those before I credit a vendor.

SaaStr's later field report is the closer I keep. Agents can absorb a lot of SDR work. There still is not a good AI account executive. Emergence, cited there across 560-plus companies, saw 36% cut SDRs, 14% cut SEs, and 28% grow AEs. The closing layer stays human.

Pick the motion from ACV still picks the motion. A $5k deal does not become ABM because a copilot drafted the email. A $100k deal still needs a named owner and a research doc. AI can make the research cheaper. It cannot change the close rate a booked meeting needs to pay for itself.

Where this goes next

Once the motion is set, the next question is where the agents sit in outbound. I point them at work nobody is doing, and I keep a person on the send. That lives on Pipeline Gen.

Sources, all third-party: The Bridge Group, AE Models, Motions & Metrics 2026; Gong Labs, The ROI of AI in Sales (updated March 6, 2026); Gong, State of Revenue AI 2026; G2, AI Speeds Discovery, But Trust Still Decides; Clari, Which Discovery Questions Matter Most by Deal Stage?; Jason Lemkin / SaaStr, the 140% caveat and no good AI AE yet (Emergence cited there). None of these numbers are mine.

Next · 09 FORECAST MATH

I build the board number bottom up, one rep at a time, then check it three ways.

Contents

09 · Forecast math

09 / 14

The forecast is an equation, not a vibe.

I build the board number bottom up, one rep at a time, from tenure, ramp, activity, and pipeline. Then I check it three ways before I say it out loud.

Most forecasts I inherit are a roll-up of rep gut calls with a haircut on top. That works until the quarter it does not, and you find out in week eleven. I built a rep productivity model to stop guessing. Every rep gets an expected yield for the quarter based on how long they have been in the seat. The team number is the sum of those yields. When a rep is behind, the model tells me in week three which input broke: ramp, activity, or pipeline.

It sits on top of what actually predicts quota. Stage conversion and deal size from that list are the coefficients in the equation below.

The equation

Expected yield (rep, quarter) = Fully ramped productivity × Ramp % (tenure)

Team capacity (quarter) = Σ Expected yield across every rep, including hires not yet started

Activity required (rep) = Expected yield ÷ ASP ÷ Trial-to-close ÷ Discovery-to-trial

Board number = the lowest of
   Plan:     Team capacity
   Pace:     Σ (actual bookings per day × days in quarter)
   Pipeline: Σ (qualified pipeline × trial-to-close rate)

Fully ramped productivityWhat a tenured rep books in a quarter. I set it from tenured reps' actual bookings, not from the quota sheet. Quota is a comp number. This is a capacity number.
Ramp % by tenureA step curve in 90-day blocks from hire date. Each rep's quarter lands in one block, and a rep who crosses a block mid-month gets half of that month at the new rate.
ASP and conversion ratesAverage deal size, discovery-to-trial, and trial-to-close. These come straight out of qualification and discovery. If discovery quality slips, these rates drop and the activity number goes up.
Pace and pipelinePace is what the rep is actually booking per day this year, projected over the quarter. Pipeline is open, qualified opportunity weighted by the trial-to-close rate. Neither one is a rep's opinion.

The ramp curve

New reps produce nothing in their first 90 days and I plan for that. Two shapes are worth running side by side: a fast curve (zero, half, full) for the op plan, and a slower four-step curve as the conservative case. I keep both in the model and re-true them against cohort actuals every month. The average real bookings per 90-day block, across every rep who has finished that block, is the next hire's assumption.

Illustrative · Ramp % by 90-day block from hire date

Illustrative curve showing the model's structure, not any company's actuals. The board sees the conservative case.

Per-rep expected yield

Here is one quarter for an eight-rep team. Every number here is an illustrative example: a fully ramped rep books $250K a quarter, ASP is $50K, discovery-to-trial is 25%, and trial-to-close is 30%. The last three columns are the back-solve. That is the activity each rep owes me this quarter to hit their own yield, and it is the conversation I have in the weekly one-on-one.

Illustrative · One quarter, eight reps ($K)
Rep	Ramp block	Ramp %	Expected yield	Deals needed	Trials needed	Discovery mtgs
Rep A	5th+	100%	250	5.0	17	67
Rep B	4th	100%	250	5.0	17	67
Rep C	3rd	60%	150	3.0	10	40
Rep D	3rd	60%	150	3.0	10	40
Rep E	2nd	25%	63	1.3	4	17
Rep F	2nd	25%	63	1.3	4	17
Rep G	1st	0%	0	0	0	0
Rep H	1st	0%	0	0	0	0
Team			925	18.5	62	247

Two reps in their first block produce zero and still count as headcount. That is why an eight-person team rarely forecasts like eight people. How many seats you need to hire, and when, is the job of the capacity planner. This model takes the seats you have and tells you what they will produce.

Three checks, one number

Plan capacity says what the team should book if everyone ramps on the curve. It is the ceiling, not the forecast. I check it against pace and pipeline, and the board gets the lowest of the three.

Illustrative · Same quarter, three views ($K)
Rep	Plan yield	Pace	Qualified pipeline	Coverage	Pipeline-backed
Rep A	250	262	880	3.5x	264
Rep B	250	214	650	2.6x	195
Rep C	150	160	520	3.5x	156
Rep D	150	118	360	2.4x	108
Rep E	63	55	240	3.8x	72
Rep F	63	74	170	2.7x	51
Rep G	0	8	90	n/a	27
Rep H	0	0	30	n/a	9
Team	925	891	2,940	3.2x	882

Illustrative · Pipeline coverage by rep vs the 3.3x line

Coverage = qualified pipeline ÷ expected yield. The required line is the inverse of the close rate, so it moves when conversion moves. Ramping reps G and H have no yield to cover. Illustrative numbers.

Say the target is $900K. Plan says $925K, so it looks covered. Pace says $891K and pipeline says $882K. The number I commit to the board is $882K, and I tell them the $18K gap and where it sits: Reps B, D, and F, all under 3.3x coverage. The fix is pipeline for three named reps, and it goes into the pipeline engine that week. A blanket "everyone push harder" fixes nothing.

When all three views land within a few points of each other, I trust the number. When plan runs far above pace and pipeline, the ramp curve is wrong, and I fix the curve before anyone fixes the reps.

Quarter to annual

The annual number is the same equation run forward. Every rep moves one ramp block per quarter, and planned hires enter at 0%. Hires approved in Q2 add almost nothing until Q4. Boards forget that every planning cycle.

Illustrative · Ramp-adjusted team capacity vs target by quarter ($K)

Eight reps from the table above plus two hires in Q2 and two in Q3, on the conservative curve at $250K per fully ramped quarter. Plan capacity only; each quarter still goes through the three checks. Illustrative numbers.

The cadence

WeeklyEach rep's pace and coverage against their own yield. Actual discovery meetings and trials against the back-solve. A miss on activity in week three is still fixable. A miss on bookings in week eleven is not.
MonthlyRe-true the ramp curve and the conversion rates against actuals. If tenured reps are booking above or below fully ramped productivity, the input changes and every yield downstream moves with it.
QuarterlyThe board gets the lowest of plan, pace, and pipeline, the gap to target, and the named reps and inputs behind the gap. Then the next four quarters run forward with the current hiring plan.

Nothing in the board number depends on a rep telling me how they feel about a deal. Rep judgment still matters. I use it on the deals in deal inspection, not to build the total.

All figures in this section are illustrative examples I made up to show the model's structure: rep labels, ramp percentages, productivity, ASP, conversion rates, pipeline, and targets. None are a company's actual results.

Next · 10 SALES CAPACITY PLANNING

The model that turns a growth target into a hiring calendar, and a hiring calendar into a number you can commit.

Contents

10 · Sales capacity planning

10 / 14

Sales capacity planning, defined.

The model that turns a growth target into a hiring calendar, and a hiring calendar into a number you can commit.

A capacity model forecasts ARR from rep production, seat by seat. Ambition is not one of the inputs.

Each seat carries a ramp schedule and an expected first-year yield. The sum of scheduled seats, net of a failure haircut, is the number you can commit.

The fundamental advantage: the revenue plan and the headcount plan become the same document. You cannot change one without repricing the other.

How it works

01Define the unit. One seat's ramp schedule and first-year yield, at the real median deal size.
02Schedule the seats. Hire dates decide when capacity exists. The grid is the calendar.
03Haircut the grid. Some reps will miss. The model absorbs 30% before the board has to.
04Commit the output. What survives the haircut is the number, and it is already a hiring plan.

Why it predicts

Misses become seat-level variances you diagnose on Friday. Not annual surprises you explain in Q4. Executed well, the growth number is an arithmetic consequence of hiring dates.

10.1

A rep is a $600K asset in year one.

Example scenario: a company exits this year at $1.2M ARR, targeting $7.0M in 12 months. Start with one seat, one ramp.

Bookings ramp, one new seat

$0

$125K

$225K

$250K

Q1

onboarding + pipe

Q2

$41.7K / mo

Q3

$75K / mo

Q4

$83.3K / mo

The unit, summarized

$600K

Year-1 yield

$1.0M/yr

Steady, M10+

$120K

Median deal

$300-330K

OTE, 50/50

The rule

Plan capacity at production, not quota. Haircut the team 30%. Front-load hiring: late seats pay next year.

10.2

The activity math behind the $600K.

One ramping seat, first 12 months, locked at the $120K median deal. Drag the deal size and watch the same $600K reprice the year.

The funnel, compounded from the nine-stage system

147

Meetings held

16%

23

Qualified opps

27%

6

POVs started

81%

5 wins

= $600K booked

Median deal (ACV)$120K

$120K is the locked plan. The $75K sensitivity: 8 logos, ~229 meetings.

Activity, per ramped seat	Annual	Qtr	Mo	Wk
New discovery meetings (S0)	147	37	12	3
Qualified opportunities (S2)	23	6	2	0.5
POVs started	6	1.5	0.5	0.1
Closed-won new logos	5	1.3	0.4	0.1

Compounded from the nine-stage table: meeting → qualified 16% · qualified → POV 27% · POV → win 81%. 147 meetings = 3 a week × 49 working weeks.

The link

Three net-new meetings a week per seat: exactly the pod's dual PG quota (AE 1 + BDR 2). The capacity model and the activity system are the same machine.

10.3

From $1.2M to $7.0M: a hiring schedule.

Example scenario: a company exits this year at $1.2M ARR and wants to reach $7.0M total ARR in the next 12 months.

Failure haircut30%

Plan range 25-30%. The deck plans at 30.

Expansion, net of churn$0.93M

Workbook models $1.0M expansion less $70K churn.

Median deal (ACV)$120K

Prices the logo count, and the activity math above.

Advanced: starting ARR and the ramp

Starting ARR Ramp Q1 Ramp Q2 Ramp Q3 Ramp Q4

Bookings per quarter of tenure. The Q4 rate holds from month 10 on: the $1.0M/yr steady state.

Seat

Start

Q1

Q2

Q3

Q4

Year 1

Cell shading steps with the dollar figure: the capacity staircase, made visible.

Illustrative board-plan build: 12 seats, 8 new hires by July.

Gross capacity

$6.96M

12 seats

Failure haircut

$2.09M

30% of the grid

Net new logo

$4.87M

~41 at $120K

Starting ARR

$1.20M

the ARR carried in

Expansion, net

$0.93M

base, less churn

Exit ARR

$7.00M

total company ARR

Why the haircut

Three in ten new reps never reach $600K.

Much of that attrition is involuntary, and you make the call. They never build pipeline, or they never learn the pitch well enough to reach the $600K a ramped rep returns in year one. Plan gross, commit to net.

Companion page Where the self-sourced meeting comes from. This module sizes the number. The weekly pipeline plan is how a rep books the one meeting a week nobody else can book for them. Standalone tool Plan your own year. The capacity planner: eight inputs, the hiring schedule, the activity math, and the burn. It solves for minimum in-year coverage, so its schedule runs leaner than the board-plan build above. Share it by link.

Next · 11 THE SCALE PATH

You graduate on evidence, or you do not get to spend like the next stage.

Contents

11 · The scale path

11 / 14

Four stages, one question each.

You graduate on evidence, or you do not get to spend like the next stage.

How to read it

Stages are gates, not vibes. The next four panels are the playbook for each one.

Stage 1 · $0-1M: sell truth, not scale.

Nothing here scales. That is the point. You are buying evidence, not bookings.

01Founder sells. An FDE builds. Nobody hires a VP Sales yet. The founder cannot outsource learning what makes buyers move.
0210-20 design partners at real prices. Discounts trade for references and roadmap input, never for silence. Free pilots teach nothing about willingness to pay.
03Run pricing experiments on purpose. Three packaging tests before you claim repeatability. The metric that scales is tied to value delivered.
04Instrument everything from deal one. Win/loss on every deal, CRM discipline from the first opportunity, signal capture running even now.

The stage in one number

10-20

Design partners, real prices

Graduate when

Two net-new logos, sourced outside the network, live through a full renewal cycle.

Failure mode

Renting revenue from the founder's network and calling it repeatability.

The only proof

Two logos with no tie to the founder. Bought, deployed, renewed. Everything before that is anecdote.

Stage 2 · $1-5M: the first repeatable pod.

Two sellers who are not the founder have to hit a number. Everything here serves that test.

01Hire 2-3 AE athletes, not a big-company VP. Sellers who prospect, run their own POVs and thrive in ambiguity. The professionalize-it VP comes later.
02Process of record in CRM, week one. Stages, exit criteria, MEDDPICC fields. Forecast becomes arithmetic the day the fields exist.
03Pipeline-generation quotas from day one. AE sources one net-new ICP meeting a week, BDR two. Coverage inspected Fridays.
04Formalize the community funnel. PQL gates, reverse trials, DevRel that ships technical content. Bottoms-up stops being luck.
05Hire the GTM engineer by ~$3M. The AI stack scales before headcount does. One person runs the entire workflow catalog.

The stage in one number

~$3M

When the GTM engineer lands

Graduate when

Two non-founder reps at 70%+ attainment, two consecutive quarters.

Failure mode

Scaling outbound spend before message-market fit is proven outside the community.

Capital amplifies the model you already have. It does not invent a repeatable motion. I saw a company hire the org the raise implied, then cut it when growth normalized. The gate stays the same: two non-founder sellers at 70%+, two quarters. Until that is true, do not staff the platform company.

The test

Two sellers, 70% of quota, twice running. Anything less and you are still at Stage 1 with a bigger payroll.

Stage 3 · $5-10M: the engine, math shown.

Illustrative: $4.2M today, $10M planned. New logo carries $5.2M of that gap, net expansion carries the rest.

Field engine · $3.6M new logo

18 / week

AE 30% · BDR 35% · marketing 25% · partners 10%

× 48 weeks

864 meetings

ICP only, held across the year

× 4.3% to POV

37 POVs

scoped, success criteria signed

Velocity engine · $1.6M new logo

420 free-tier orgs

fed by the OSS and community funnel

× 38% to PQL

160 PQLs

3+ seats · prod-adjacent · findings in 14 days

× 40% to opp

64 velocity opps

worked by the velocity pod, not the field

The output

62 wins · $5.2M new logo. Field: 37 POVs × 81% win = 30 at $120K blended ACV. Velocity: 64 opps × 50% win = 32 at $50K ACV.

The haircut

Staff the seat grid against a 30% team failure rate. If capacity net of the haircut does not clear the number, change the plan, not the story.

Stage 4 · $10-20M: make growth compound.

The question stops being can you sell it and becomes does each new dollar cost less than the last.

15% → 35%

Channel share of new

Marketplaces first, since EDP and MACC budgets unlock stuck procurement. Then two security VARs.

6-8

Reps per first-line manager

Players stop coaching at this span. Managers get hired on coaching evidence, not on their old logo slide.

115%+

NRR, from playbooks

Whitespace maps, trigger-based plays, QBRs run with customers, not at them. Expansion stops being a surprise.

International

Rides evidence, never ahead of it. An EMEA pod once three or more organic logos already exist in region. Hiring into a region with no pull is the most expensive way to learn it has no pull. Open with one AE and one SE on local paper, local support hours and a reseller already selling into those same accounts.

The board forecast

Built bottom-up, seat by seat. Capacity per seat, coverage entering the quarter, a tripwire attached to every assumption. Ramp curves, start dates and the 30% failure haircut all sit in one model, so moving a hire date reprices the year on the spot. The Series B gets raised on evidence, not narrative.

Graduate when

Channel clears 25% of new business, two quarters running.

Failure mode

Scaling spend on an engine whose Stage 3 unit math never closed.

Next · 12 THE TEAM

Sixteen functions. Which exist by $5M, $10M and $20M is the hiring argument.

Contents

12 · The team

12 / 14

The functions you are actually hiring.

Sixteen functions. Which exist by $5M, $10M and $20M is the hiring argument.

By milestone

New logos & expansion

Sales function

RSMs5M

Enterprise Sales

AEs5M

Commercial Sales

BDRs5M

Business Development

SEs5M

Sales Engineering

Marketing function

Pipeline5M

Demand Generation

DevRel5M

Dev Community

PMM10M

Product Marketing

Content10M

Content Marketing

Channel & alliances

CAMs20M

Channel Sales

Partnerships20M

GTM Alliances

Field20M

Field & Channel Marketing

Co-marketing20M

Partner Co-Marketing

Renewals & success

AMs10M

Account Managers

CSEs10M

Customer Success Engineers

Advocacy20M

Customer Marketing

Brand20M

Brand & Corporate

The 17th function

It spans every function above. Systems, forecast, comp, territory, data. One person: the GTM engineer.

12.1

Who to hire, and when.

Four stages. The seat count is the plan; the role each stage unlocks is the argument.

Stage	Sellers	Support	Role the stage unlocks
$0-1M	Founder + 1 FDE	Contract SE help	FDE: the builder in the room
$1-5M	2-3 AEs + 1-2 BDRs	1 SE · DevRel	GTM engineer by ~$3M
$5-10M	6-8 AEs, dual-track	2-3 SEs · 2 BDRs	More AEs, SEs, FDE on strategic. No VP Sales until evidence.
$10-20M	10-14 AEs · EMEA pod	SE team · channel lead	More of the same five seats.

Pair every seller hire

Two at a time, shared onboarding, and an honest A/B on whether the bar held.

Hire ahead of the ramp

A month-7 hire contributes a stub this year. The ramp is the constraint.

Backfill on day 60

Two misses at the 60-day activity gate, open the req that same day.

The rule

Every seat on this table is a line in the capacity model. Change a hire date and the ARR number reprices itself.

Companion page Who to hire, and how to select them. Stage gates, scorecards, the selection loop, ramp gates, and the ungated hiring kit. This module names the functions; that page is how you fill the seats.

12.2

What the team costs.

2026 ranges for AI and AI-security scale-ups in US metros. Quota-to-OTE is the number that decides the model.

Role	OTE range	Split	Mechanics
Enterprise AE	$300-330K	50 / 50	Quota-to-OTE 3-3.5x. Accelerators 1.25x above plan, no cliffs.
FDE	$220-280K	80 / 20	Bonus on POV wins and on repeats that get productized.
Sales Engineer	$230-260K	70 / 30	Variable on POV win rate and POV cycle time.
GTM Engineer	$160-200K	Base + bonus	Owns the AI stack. Bonus on pipeline-per-pod efficiency.
BDR	$90-110K	75 / 25	Paid on held meetings that convert to Stage 2, never on dials.

55-60%

Year-1 yield vs quota

Nobody carries a full number before month seven. Plan year-one production against this yield, and gate 10% of variable comp on pipeline generation.

Next · 13 THE OPERATING SYSTEM

Leading indicators get inspected. Lagging indicators get judged. The calendar enforces both.

Contents

13 · The operating system

13 / 14

The operating rhythm.

Leading indicators get inspected. Lagging indicators get judged. The calendar enforces both.

Leading · Inspected weekly

Stage-0 meetings by pod and source, vs the weekly bar
PQL flow vs gate, velocity pod SLA compliance
Coverage entering quarter: 3.5x new · 1.2x expansion
Stale pipeline under 15% older than two quarters
MEDDPICC score on every Stage-2+ opportunity

Lagging · Judged monthly

Bookings vs plan, by segment and by source
Blended new-logo ACV vs the field-economics floor
Stage-2 to win 20%+ · POV-to-win 75-85%
NRR 115%+ · GRR 90%+
CAC payback under 18 months · burn multiple

The cadence

Mon

Commit and upside call

Wed

Pipeline scrub by stage

Fri

PG scorecard by pod

Monthly

Win/loss and pricing

Quarterly

QBRs and the board pack

The rhythm

Nothing on this page is a status meeting. Every session ends in a decision, an owner and a date.

13.1

Six ways AI scale-ups die.

Every risk gets a mitigation and a numeric tripwire. When the tripwire fires, the decision is already made.

01Adoption runs slower than the hypeUnder 40 PQLs in a quarter

Mitigation

Anchor on what is in production today. PLG keeps CAC low while the category matures.

Tripwire → action

Under 40 PQLs in a quarter → shift the mix to field, revisit the segment bets

02Priced as a tool, not a platformField ACV under $120K for two quarters

Mitigation

Pricing floors, platform packaging, POV outcomes the economic buyer signed.

Tripwire → action

Field ACV under $120K for two quarters → repackage and requalify up-market

03Rep failure runs past the haircutTwo reps below gates at day 60

Mitigation

Hold the hiring bar. Scorecards from day one, activity gates at day 60.

Tripwire → action

Two reps below gates at day 60 → reallocate pipeline, open the backfill req that day

04Paid tiers burn community goodwillDownloads down 20% for two months

Mitigation

Clean OSS and paid boundary: the community keeps the tool, enterprise buys governance.

Tripwire → action

Downloads down 20% for two months → packaging review with founders

05Enterprise deals go single-threadedOver 30% of late-stage pipeline single-threaded

Mitigation

MEDDPICC enforced: no Stage 3 without an economic buyer meeting on the calendar.

Tripwire → action

Over 30% of late-stage pipeline single-threaded → daily deal reviews on those accounts

06Right market, thin captureTwo non-founder AEs still under 70% while the board pack says category leadership

The category can be real and the tech can be real, and you still fail as an independent company. I was there for the rise at Lacework. Cloud security was the right market. The graph was real technology. What decided the outcome was whether the product sat in a weekly workflow, owned a budget, and had a distribution engine that could keep it.

Inspect the chain, in order: problem, product, workflow, control point, budget, distribution. A hole anywhere leaves you as a feature inside someone else's platform.

Mitigation

Sell the workflow you sit in this quarter. Do not sell the category you hope to own next year. Land on a named owner and a budget line, then expand.

Tripwire → action

Two non-founder AEs still under 70% while the board pack says category leadership → rewrite the land motion around the workflow you actually sit in, or cut the platform story.

Why this exists

A plan without failure conditions is a pitch. Wire each tripwire to a decision and review all six at every QBR.

13.2

Forecast inspect.

Commit, upside, pipeline. Three buckets. Evidence, not hope. The Monday call is this page, spoken.

The three buckets

01Commit. Economic buyer met. Paper process mapped. Champion moved this week. Miss any one and it is not commit.
02Upside. Two of the three. A real path exists. A date on a slide is not a path.
03Pipeline. Everything else. Coverage lives here. It does not live in commit.

Paper rule

Unknown paper process cannot sit in commit. A close date set before legal has been asked is hope with a date on it.

Name the break.

Do not write ghosted or price on the loss. Classify it, then change that part of the system.

01No gap. Qualified, and discovery found none. Qualification.
02No next meeting. Good call, nothing booked. Momentum.
03Scope ignored. They did not read it, or it was not theirs. The document.
04Revision dies. An unresolved concern, or a weak champion.
05Cannot go upstairs. The champion cannot take it. Approval map.
06Price on sight. Rejected before the paper. Positioning, anchor, or risk.
07Agreement stalls. Decision, legal, or timeline.
08Signed, unpaid. Payment handoff.

Repeated objections become scope language. Missing stakeholders become discovery questions.

How this sits on the week

Monday is commit and upside. Wednesday scrubs the rest. If you cannot say which bucket and why, the deal is pipeline. Coverage (3.5x new, 1.2x expand) is a leading indicator. It is not a substitute for this inspect.

Next · 14 THE OPERATOR

Enterprise revenue, rebuilt around an AI-native stack.

Contents

14 · The operator

14 / 14

The operator behind the Sales Playbook.

Enterprise revenue, rebuilt around an AI-native stack.

Andrew Wesbecher

26 years building enterprise security and infrastructure revenue.

VP Sales and CRO roles scaling early-stage SaaS.

Today: full-time chief revenue officer or VP of Sales search, AI or AI security.

Advisory continues across AI security: AI-SPM, identity, agent security, detection and response.

This Sales Playbook was researched, written and designed with the AI stack it describes.

Where the work happened

TIBCO · Meraki · ThousandEyes
Lacework · Contrast · Traceable

$11B+

Combined exit & peak value

Lacework $8.3B peak. Meraki $1.2B and ThousandEyes $929M to Cisco. Traceable $900M to Harness.

$1M → $25M

Early-stage ARR, scaled

The path this Sales Playbook maps, run more than once as first or second sales leader.

Next step

If this maps to your next twelve months, I would rather talk than pitch. andrew@wesbecher.llc

Companion page The operator, in full. Career history, the four exits, and how to work together.

Contents

Sales & GTM for AI Scale-ups.

Contents

AI changed the job in November 2022. Sales and GTM leadership is now adapt or die.

Your buyer moved. Three years ago.

The GTM delta, in one table.

What 26 years says still wins.

MEDDPICC is old. Complex enterprise still dies the same eight ways.

Two engines, one capacity model.

The nine stages have not changed.

The FDE motion: selling by building.

Bottoms-up and top-down, same account.

Time-to-first-value is a commercial choice.

Expand vectors

Pricing architecture

Qualify the company, then the committee.

Economic

Product decision

Evaluators

Running the complex deal.

Win-rate note

Discovery is a handshake, not a demo.

How I run discovery.

The next move is a record.

Build the price before you show it.

The champion marks up V1.

Signed is not collected.

How I run a patch.

Project mode.

Hire for it

Pick the motion from ACV.

Under $5k

$5k to $25k

$25k to $100k

$100k and up

Under $5k

$5k to $25k

$25k to $100k

$100k and up

The stack I pay for but never log into.

A prompt is the new login.

Always on, no prompt needed

The operating model

AEO is the new demand capture.

Measure it

The headless workflow catalog.

What the engine buys you.

Not a concept

Find the revenue leak before you buy another AI tool.

Sales qualification and pain discovery.

Situation

Pain

Impact

Urgency

Vision

Decision

Commitment

What actually predicts quota.

Pipeline velocity

Stage conversion rate

Average deal size

Next step quality

Champion engagement

Win/loss patterns

Total time on calls

AI changes the math.

The forecast is an equation, not a vibe.

The ramp curve

Per-rep expected yield

Three checks, one number

Quarter to annual

The cadence

Sales capacity planning, defined.

A rep is a $600K asset in year one.

The activity math behind the $600K.

From $1.2M to $7.0M: a hiring schedule.

Four stages, one question each.

Stage 1 · $0-1M: sell truth, not scale.

The stage in one number

Graduate when

Failure mode