David Appel

The win looks inevitable, not comparative.

• Their playbook: choose a system-of-record wedge—the first few objects a customer uses daily—then build a system of action, the fastest path from data entry to decision to execution.
• Price to match the workflow: where value scales with volume, add a usage or take-rate component, then measure contribution margin by product line to see which automations merit investment.
• For distribution, build where the work happens.
• Integrations are table stakes; partnerships with the two or three upstream or downstream vendors that can route customers provide leverage.

Appel identifies the objects as job, claim and transaction, and the volume-based pricing units as claims, closings, jobs and leads. Drawing on Euclid Ventures' roundup of vertical software moments, they cite ServiceTitan, Qualia, Broadlume, EvolutionIQ and Moxie as companies embedded deeply enough that a niche cannot easily switch.

Read original

Owning data used to be the defence, but large language models can reach much of it, too.

• What they cannot reproduce is outcome data: the record created when your agent does the job with your context.
• As Appel puts it, systems of record store what happened, while systems of action make it happen and keep the receipts.
• They say buyers now ask a model first, then text a friend who runs the same kind of business to validate the answer.
• That second step cannot be bought with a bigger booth or more advertising; it comes from delivering on the brand promise.

Appel opens with an apparent contradiction: a legal AI company reportedly displaced by a general-purpose model, then valued 40% higher at $15 billion.

• They argue the moat relocated rather than disappeared.
• Their test for any vertical AI founder is one sentence: what outcome does your agent drive, and would your customer's CFO put their name behind it?

The argument shifts across these posts.

• Earlier pieces stress owning proprietary data, embedding pattern layers in the product and improving predictions over time.
• A later post argues that is no longer enough: general models can reach the same data, making outcome data the new moat.

Read original · Original 2

They draw three conclusions for AI-native companies.

• First, a company's own pattern data matters more than running a model over customer data.
• Investors want proprietary insight layers with vertical depth a horizontal model cannot copy; otherwise, the company is a feature, not a company.
• Second, predictability beats growth.
• Investors want to know not just how fast a company is growing but whether it can say what quarter seven looks like.
• Retention signals value, rather than serving only as a metric.
• Third, seat-based pricing is the first casualty of AI performing the work: charging per seat is like charging per employee at a factory you just automated.

Appel's concrete case for the pricing shift: a durable model prices the value of the decision, not the human who used to make it. They say CFOs building models on seat count today are building on sand, and close by asking readers which of the three keeps them up at night.

Read original

Their advice is to fix that first, then set guardrails on what agents can see and do.

• Companies that skip this step, they report, end up with siloed agents fighting over conflicting data.
• Appel applies the same discipline to growth-equity sourcing: build a vertical ontology of the 30–50 job titles, workflows, integrations, compliance triggers and pricing units that define your niche.
• Feed it into research and CRM so searches draw on operator context, then turn signals into a weekly queue where humans still pick ten outreach targets every Monday.

The weekly queue is the mechanism: scrape hiring bursts, new integrations, pricing-page changes, regulatory deadlines and customer logo swaps.

• Have AI summarise each company into wedge, ideal customer profile, ACV, retention risk and likely buyer, then make a human choose.
• Every call note, win or loss and why-now answer becomes training data.
• If the model is not getting smarter each week, Appel says, you are producing faster noise.

The posts show a tension rather than a clean change: a salon write-up argues for slowing down on agents until data and governance are right, while growth-equity and sales posts urge companies to adopt AI now and move fast. Appel resolves the tension with a line from the room: "Slow down to go fast."

Read original · Original 2

They recommend per-account agents that track emails, calls and signals to show which deals need attention, who the right buyer is and what to do next—so pipeline reviews stop being theatre.

• Automate CRM updates after calls instead of hoping reps log them, and combine call recordings, emails and champion changes that currently sit in silos to inform decisions.
• Appel reports 20–30% productivity gains from making every rep smarter, rather than doubling headcount or tripling quota.
• Their description of the buyer is blunt: knowledgeable, impatient, uninterested in the founder story and looking for outcomes—as is the board.

At a go-to-market salon, Appel describes a C-level rep who was nearly let go but became a solid performer once AI handled research, next-best-action suggestions and call prep. Their team-level conclusion: train people as you train agents, with role play, call scoring against your real methodology and meeting preparation notes built from account history.

Read original · Original 2

They recommend three moves for vertical SaaS operators: turn the best discovery calls into a written talk track, with objection handling organised by buyer persona rather than generic scripts; build a 30-60-90 onboarding plan for the first sales manager before hiring, so a system scales instead of tribal knowledge being delegated; and qualify for pain, urgency and technical fit early, because seven-figure deals die when teams mistake product excitement for budgeted necessity. The advice reflects a recurring finance-and-go-to-market theme: enterprise winners industrialise the middle of the funnel and inspect it rather than improvising.

The concrete artefacts are the talk track and pre-written onboarding plan: one captures how the founder actually sells; the other makes the first management hire a transfer of a system, not instincts. Appel credits Bessemer Venture Partners material on Recall.ai's Amanda Zhu for the argument.

Read original

Meaning, not features, builds attachment.

• They also note that most organisations are piloting AI but almost none have fully integrated it, leaving customers confused; more feature explanations make things worse.
• AI raises the floor of writing, not the ceiling: it predicts but does not feel, and you are still selling to humans.
• If your messaging starts with your roadmap, Appel says, you have already lost the room.

Appel retells the 2009 Salesforce Service Cloud launch: the product was not fully built, but its messaging named what customers were living through and showed a path to resolution.

• A slide from the launch was still circulating fifteen years later.
• They attribute the framework and story to Joe Lazar of Article Group.

Read original

The approach turns social learning into specific, testable actions rather than general impressions, and much of their advice reaches the page this way.

From the SaaStr AI vertical panel with Jeremy Kaufmann, Cecilia Ziniti and Alexander Hejnosz, Appel takes the lesson that context beats model quality, UI and pricing: undocumented workflows, compliance edge cases handled in their sleep, and workarounds power users built in year one and abandoned by year three.

• One detail they say they are still thinking about: sensors on nursing home walls to help keep people alive longer.
• The same pattern appears in write-ups of Demand & Expand, a London Stock Exchange CFO event, a VP Sales Forum led by Jeff Williams, and NetSuite alumni gatherings.

Read original · Original 2 · Original 3 · Original 4 · Original 5 · Original 6 · Original 7 · Original 8

It defines the terms they reuse across their writing — "systems of action," proprietary insight layers, outcome-based pricing — and shows what those ideas look like to the people judging whether a company can go public.

Reporting from a run of investment-bank conferences and the NYSE IPO Summit, Appel notes that one phrase came up more than any other: "systems of action," not systems of record.

• They draw three conclusions.
• First, an AI-native company's own pattern data matters more than running a model over customer data: investors want proprietary insight layers with vertical depth a horizontal model cannot copy, or the company is "a feature, not a company."
• Second, predictability beats growth: the investor question is not how fast you are growing but whether you can say what quarter seven looks like.
• Third, seat-based pricing is the first casualty of AI performing the work, because charging per seat is like charging per employee at a factory you just automated.
• The durable model, they argue, prices the value of the decision rather than the human who used to make it.
• They close by asking which of the three keeps the reader up at night.

For brands planning a campaign

Useful as a frame for an internal agent-readiness session with a revenue or finance leader: inventory the company's proprietary pattern data, test whether forecasts can withstand investor scrutiny, and pressure-test seat-based pricing before AI reduces the headcount the model assumed.

It names the exact question that separates a demo-ready agent from a defensible one and explains why they think the moat moved.

Appel opens with an apparent contradiction: a legal AI company reportedly displaced by a general-purpose model, then valued 40% higher at $15 billion.

• Their explanation is that the moat has moved.
• It used to be your data, but large language models can reach that too; now it is outcome data, which exists only after your agent does the job with your context.
• They put the distinction sharply: systems of record store what happened, while systems of action make it happen and keep the receipts.
• They also describe how buyers now behave — they ask a model first, then text a friend who runs the same kind of business to validate the answer, a second step that cannot be bought with a bigger booth or more advertising.
• They close with a question they would put to every vertical AI founder: what outcome does your agent drive, and would your customer's CFO put their name behind it?
• If the answer takes more than one sentence, they say, keep thinking.

For brands planning a campaign

A qualifying question for a GTM and agent-readiness brief: state the outcome metric, name the proof a customer CFO would sign, and map the peer-validation path before spending on category marketing.

Where the IPO and moat pieces argue about where value sits, this one shows the operating moves inside a revenue organisation, in the voice of the people running them.

Appel recaps the 60th Go-to-Market Leadership Salon — more than 160 senior leaders in one room — and reports three things that moved them.

• Build versus buy is settled: buy, and build only what is uniquely yours, because someone is already thinking about your problem around the clock.
• AI is raising the floor, not just the ceiling: a C-level rep who was nearly let go became a solid performer once AI handled research, next-best-action suggestions, and call prep, because the middle of the team is moving up.
• The best teams do three things: know the buyer's journey cold and attack where it is collapsing (decision cycles are compressing, with fewer touches and faster noes); train people the same way they train agents (role play, call scoring against the real methodology, pre-meeting briefs); and fix data and governance before scaling agents, since scattered CRM, email, calendar, conversation, and product-usage data produces siloed agents fighting over conflicting records.
• Their quoted line of the night: "Slow down to go fast."

For brands planning a campaign

A concrete adoption sequence for finance or revenue-operations teams: start with pipeline, win rate, ARR forecast, and churn signals; bring enablement up to the same standard as the agents; then set guardrails on what agents can see and do.

It also states the productivity claim and buyer shift Appel repeats elsewhere, making it a quick test of whether you agree with their working assumptions.

Writing up a VP Sales Forum led by Jeff Williams, Appel offers three moves.

• Let per-account agents suggest the next move before the rep asks, tracking emails, calls, and signals to surface which deals need attention, who the right buyer is, and what to do next — so pipeline reviews stop being theatre.
• Kill manual CRM entry: the call ends, the rep hits a button, and the conversation is synthesised into the CRM, a welcome change for someone who lived through the era of hoping salespeople would log their calls.
• Turn unstructured data into decisions by aggregating Gong calls, emails, and champion changes that sit in silos; companies doing this, they report, are seeing 20–30% productivity gains rather than doubling headcount or tripling quota.
• Their description of the new buyer is blunt: knowledgeable, impatient, and uninterested in the founder story — they want outcomes, and so does the board.

For brands planning a campaign

A starting point for an agent-readiness framework, focused on internal data hygiene and rep-facing agents, with an argument for shifting budget from headcount to tooling.

The three actions are specific enough to copy, and the second — a forced Monday pick of ten targets — is the part most teams skip.

Responding to Euclid Ventures research on AI-powered growth equity, Appel reframes it as an "AI-qualified lead" engine that manufactures investable or acquirable targets before competitors notice the pattern, then lays out three actions.

• Build a vertical ontology: list the 30–50 job titles, workflows, integrations, compliance triggers, and pricing units that define your niche, and feed it into research and CRM so AI searches with operator context instead of generic keywords.
• Turn signals into a weekly queue: scrape for hiring bursts, new integrations, pricing-page changes, regulatory deadlines, and customer logo swaps; have AI summarise each company into wedge, ideal customer profile, ACV, retention risk, and likely buyer; then require humans to choose ten outreach targets every Monday.
• Close the loop with truth data, where every call note, win or loss, and "why now or why not" becomes training data — if the model is not getting smarter each week, the output is just faster noise.
• They connect this to earlier finance and revenue work on shared definitions and a single pipeline view.

For brands planning a campaign

A scoping template for a growth or agent-readiness plan: it shows whether a team has the CRM, research, and feedback plumbing in place before buying more AI, and puts a human decision point inside an automated workflow.

It offers a route from value metric to packaging to expansion, plus a distribution argument.

Working from Euclid Ventures' roundup of vertical software moments, Appel argues that category leaders win not on features but on inevitability — once embedded, they become the operating system for a niche that cannot afford downtime.

• They offer three tactical moves.
• Pick the system-of-record wedge, then expand into the system of action by mapping the first three objects the customer lives in daily (job, claim, transaction) and building the fastest path from data entry to decision to execution.
• Monetise the workflow, not just the seat: where value scales with volume — claims, closings, jobs, leads — add a usage or take-rate component, then instrument contribution margin by product line to see which automations are worth scaling.
• Build distribution where the work already happens: integrations are table stakes, partnerships are leverage, so design a co-sell motion with the two or three upstream or downstream vendors that can route customers to you.
• They again cite finance and go-to-market agreement on the value metric as the thing that sharpens packaging and reduces discounting.

For brands planning a campaign

A concrete pricing and packaging agenda for a finance leader developing an AI-era product line, with a partnership and distribution angle to test alongside it.

Drawing on Euclid Ventures' roundup of ServiceTitan, Qualia, Broadlume, EvolutionIQ, and Moxie, they say category leaders become inevitable once embedded in a niche.

• A separate post distinguishes systems of record (what happened) from systems of action (making it happen and keeping receipts).
• For AI agents, they say the new moat is outcome data that exists only after the agent does the job.

For brands planning a campaign

For an AI-agent readiness framework, expect Appel to look beyond model quality and UI to ask which workflow the agent owns and what outcome data it produces. A campaign angle can ask prospects to name the decision their agent drives and whether a customer's CFO would sign off on it.

Examples: Source 1 · Source 2 · Source 3 · Source 4

Their advice: fix data and governance before scaling agents, then add guardrails on what agents can see and do. In a growth-equity post, they connect this to AI sourcing: build a vertical ontology of 30-50 job titles, workflows, integrations, compliance triggers, and pricing units, then turn signals into a weekly outreach queue.

For brands planning a campaign

A readiness assessment should cover data synthesis and agent governance, not just model selection. Campaign content can prompt operators to inventory scattered systems and agree on inputs before adding more agents.

Examples: Source 1 · Source 2 · Source 3

Recurring examples include per-account agents that surface the next best action before a rep asks, automated CRM updates after calls, and turning unstructured call and email data into decisions. One post cites 20-30% productivity gains from making every rep smarter rather than hiring more.

For brands planning a campaign

For a GTM framework, expect Appel to test whether AI is embedded in the daily sales motion or just generating reports. Campaign content can focus on enablement workflows such as call scoring, role play, pre-meeting briefs, and removing manual CRM entry.

Examples: Source 1 · Source 2 · Source 3

Their advice includes turning discovery calls into written talk tracks with objection handling by buyer persona, building a 30-60-90 onboarding plan for the first sales manager before hiring, and qualifying for pain, urgency, and technical fit early.

For brands planning a campaign

A campaign brief on scaling sales can ask which parts of founder-led sales have been written down and where the middle of the funnel is improvised.

Examples: Source 1

In IPO-focused posts, they argue predictability beats growth, investors ask what Q7 looks like, and seat-based pricing is the first casualty of outcome-based AI pricing. Durable models price for the value of the decision made, not the human who used to make it.

For brands planning a campaign

For a finance or revenue leader assessing Appel, expect a framework that ties pricing, retention, and forecast discipline to AI-agent output. Campaign messaging can benchmark whether the company prices on seats or outcomes.

Examples: Source 1 · Source 2 · Source 3

They quote an attendee: 'Someone is thinking about your problem 24/7.

• Buy the platform.
• Build only what's uniquely yours.' They frame this as a way to raise the floor for average reps.

For brands planning a campaign

A GTM framework can include a build/buy/partner decision checkpoint, with criteria for what is uniquely yours and what a platform already handles.

Examples: Source 1

They say the 2009 Salesforce Service Cloud launch succeeded because its messaging named what customers were living through before the product was fully built. They also note that AI raises the floor of writing, not the ceiling, and that starting with your roadmap loses the room.

For brands planning a campaign

Appel can help shape an AI-agent readiness narrative around the buyer's current workflow pain rather than the agent's features.

Examples: Source 1

Events include SaaStr AI, Demand & Expand, a London Stock Exchange CFO event, the NYSE IPO Summit, and NetSuite alumni gatherings. The method turns social learning into specific actions rather than general impressions.

Conference or panel takeaways distilled into numbered actions with named speakers and practical steps.

For brands planning a campaign

For a campaign, Appel's posts can be timed around event seasons to turn live conversations into structured takeaways.

Examples: Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7 · Source 8

They describe DOSS as the operating system for consumer goods brands across inventory, cost, orders, and procurement.

For brands planning a campaign

A brand evaluating Appel for a GTM framework can see hands-on experience scaling marketing teams inside a vertical SaaS company. Ask for examples beyond their company-specific hiring posts.

Examples: Source 1 · Source 2

They describe DOSS as an operating system for consumer goods brands and cite its Series B momentum.

View example

Ask whether they own, host, or are otherwise employed by it.

View example

The post points to a NetSuite alumni connection; ask about their role and dates.

View example

Both write for people deciding where AI changes revenue work, rather than explaining the models themselves, and credit the speakers and platforms behind the claims they repeat.

Singla focuses on AI search visibility: which platforms large language models pull answers from and how brands show up there.

• Their audience is CMOs and marketing leaders.
• Appel stays on the operating side of vertical software: what a company owns, how it prices, and how its sales team is built.

Read original · Original 2

Both examine what AI should own inside a revenue team and what stays with the person.

Wetzel draws on Pavilion survey data to cover revenue-leadership hiring, framing the work as building a new operating model from scratch. Appel focuses on vertical software and its economics—moats, outcome data, pricing—with reference points in investment-banking and CFO conversations rather than peer survey data.

Read original · Original 2

Both take a specific external argument—a vendor piece, a CFO essay, or a conference slide—and turn it into conclusions for technology executives.

Lah writes in a formal framework voice tied to TSIA's models and its TECHtonic show, making the case from the services-and-margin side of the industry. Appel writes shorter posts anchored to events they attend, for vertical SaaS operators and go-to-market leaders rather than services and customer-success organizations.

Read original · Original 2

The need is a go-to-market and AI-agent readiness framework; Singla reframes its first question as discovery.

• He reports polling 100-plus CMOs on whether generative engine optimisation or answer engine optimisation was in their top three problems and getting an 80 percent show of hands, with 5 percent saying their in-house team had solved it and 60 percent saying they would rather have an outcome-based partner than a do-it-yourself dashboard.
• He also convenes the platform and investor side of AI search, listing LinkedIn, Reddit, NVIDIA, Lightspeed and Scale Venture Partners on an INDEX'26 ecosystem panel about what the platforms see from the inside.
• For a vendor selling an AI agent into mid-market revenue and finance teams, that is a useful starting frame: buyers ask a model first and then check the answer with a peer, so visibility is part of readiness.

The tradeoff is coverage of the operating questions a campaign aimed at evaluating a go-to-market and AI-agent readiness framework actually names.

• Singla's posts are about AI search positioning, CMO influence and who owns the outcome—not the internal checks the framework needs, such as where completed-job data lives, what guardrails sit on the agent, how the product is repriced when work is automated, or how reps are trained against the sales methodology.
• A vendor that wants a scorecard leaders can fill in would need to pair him with an operator voice.
• Ask Singla for a commercial offer, rate card and audience breakdown before planning a paid placement.

Examples: Source 1 · Source 2

He cites a Pavilion AI Pulse Survey showing 95 percent of GTM teams using AI more than six months earlier, 70 percent saying it already lets them do more with fewer resources, and 46 percent having no clear owner or strategy for any of it.

• He then proposes the model itself—agents handle internal work such as meeting prep, enrichment, pipeline reporting and workflow automation, while humans own the customer—and describes what implementation looks like in practice: one approach in a Pavilion operator dinner returned a 76 percent response rate and 41 percent executive meetings against a 1.5 percent baseline, and another replaced three full-time alert-correlation roles with $18K in tokens against $450K in headcount.
• He is direct about the failure case too, reporting a large enterprise with $50M invested, 200 use cases tested and no quantifiable ROI for the board, which he attributes to infrastructure rather than AI.

The tradeoff is depth where Appel is strongest: value-metric and pricing restructuring, vertical workflow ownership, and the finance-side conversation about what a decision is worth.

• Wetzel writes as an operator comparing notes with peers, and his pricing and margin arguments are thinner than his org-design ones.
• His second post also reads as a job search — he says he is looking for a leader role to implement the new operating model — so a vendor planning a sponsored series should confirm availability and labelling rather than assume a commercial format.
• Wetzel's two posts are dated within three days of each other in May 2026, which is a narrow window for judging how he sustains a series.

Examples: Source 1 · Source 2

His posts build the argument that implementation-heavy AI delivery should be funded through a technical professional services motion—the forward-deployed engineer pattern—and that a company should accept low services margin in year one to secure the data layer and the lock-in that produces durable recurring revenue.

• He also presses the forecasting side, describing TSIA's Model Driven Revenue Engine and pairing a TECHtonic episode with Stephen Messer of Collective[i], who argues the weekly forecast call is guesswork, against OpenAI's finance function publishing zero-day close and continuously updated forecasting goals days later.
• For a campaign aimed at revenue and finance leaders judging whether they can scale an agent, that is the version of readiness a CFO will recognise: can we evidence the outcome, can we price it, and can we forecast the business around it.

The tradeoff is that Lah's frame is the vendor's own economics and the services organisation, not the end buyer's workflow.

• He gives less on vertical workflow ownership, rep enablement and the live operator-session format needed for a campaign evaluating a go-to-market and AI-agent readiness framework.
• His posts lean on TSIA positions and other companies' published research and named authors rather than firsthand observation of one vertical.
• Unlike David Appel, who gives readers a single diagnostic question and a worksheet, Lah offers a strategy thesis that needs an operator to operationalise it.
• Ask for audience data, per-post performance and commercial terms before considering sponsorship; his posts do not state them.

Examples: Source 1 · Source 2