You raised the round.
You're still not standing on your own.
You hit the targets. You hired ahead of plan. You bought the growth the board wanted to see — and the engine underneath it never got built.
So the growth costs more than it returns. The forecast keeps missing. The renewal comes back lower than your team predicted. And every quarter you're a little more dependent on the next round to cover the gap the last one left.
That's not a sales problem. It's structural — and you already suspect it. You're right.
Growth at any cost isn't a strategy. It's a loan against your company's future — it serves this round and quietly mortgages the next.
The founders who get free build the revenue engine in sequence with their ARR — the right architecture for the stage they're actually in. Build it in time, and each raise becomes optional fuel toward standing on your own, instead of life support for growth you couldn't hold. A, B, and C still come. They're fuel you choose, not a rescue you need.
That's what “Win and Last” means — the engine that lets you stand on your own is the same engine that compounds the value your investors underwrote. Built right, you don't have to choose.
If you're a VC or PE operating partner, it's the same read from the other side of the table. Growth that outruns the engine caps the return — capital spent on growth that can't compound. The disciplined build is the one that pays.
Structural problems don't announce themselves. They show up as patterns — pipeline you can't trust, retention you can't predict, deals that went to someone you don't respect.
AI is the newest of these patterns and the most exposing. Engines that compound AI as leverage were already coherent before AI arrived. Engines that industrialize the wrong work, faster, were already broken — AI just made it visible.
88% of organizations use AI.
Only 1% have reached operational maturity.
The gap is architectural, not technological. Your revenue engine either absorbs AI as leverage — or industrializes the wrong work, faster.
The Revenue Data Debt™ Diagnostic is 24 statements across your four revenue functions and the six seams that connect them. About ten to fifteen minutes. A defensible read on the gap between the client data you have and the data your AI needs — and exactly where it's accumulating.
24 statements · four functions, six seams · about 10–15 minutes · nothing to sign up for
Most of what looks like a sales problem at around $5M is structural.
The reps aren't broken. The team you hired isn't broken. You aren't broken.
The engine is broken — and the engine can be architected.
Marketing optimizes for one outcome. Sales for another. Client Success for a third. Client Support absorbs whatever falls through. Finance and Product, operating from outside the revenue organization, bring their own disciplinary orientations — unit economics and forecast precision for Finance, feature velocity and roadmap optionality for Product. Each function looks productive. The company doesn't compound. By the time it shows up — churn nobody saw coming, expansion that didn't happen, renewals that quietly downgrade — the operating choices that produced it were made months ago.
No tooling fixes this. No methodology fixes this. No fractional executive fixes this by stepping into the role and operating it.
That six-cylinder fragmentation is the dominant reason most companies can't operationalize AI. The 1% aren't running better AI — they're running engines where Marketing, Sales, Client Success, and Client Support inside the revenue organization, plus Finance and Product participating from outside it, are operating as close to a singular revenue architecture as a real company gets. AI compounds in engines that were already coherent. It exposes the ones that weren't.
It is fixed by architecting the engine, transferring it to the operators who will run it, and exiting before dependence forms.
The path is sequential. Skip a stage and the next one breaks.
Most revenue problems get solved out of order. Velocity tools deployed onto unaligned organizations. Alignment programs run on data nobody trusts. Architecture work attempted before clarity exists about what the engine is actually being built to produce.
The companies that escape the pattern do it in a specific order.
Clean, structured, timely data. A single source of truth across forecasting domains. Objectives specific enough to be falsifiable. Without clean data, your forecast is fiction and AI processes garbage faster.
All six cylinders of the revenue engine — Marketing, Sales, Client Success, and Client Support inside the revenue organization, plus Finance and Product participating from outside it — operating from one architecture, one source of truth, one shared definition of the client outcome. Accountability lines that survive when leadership changes.
The architecture running. Forecast accuracy you trust. Net retention compounding. The right operators in seat, holding the standard. The engine producing results without the founder as the bottleneck.
The sequence is not negotiable. Companies that try to skip stages typically end up paying for the same work twice — once in the wrong order, once in the right one.
The Future of B2B SaaS Under AI
SaaS isn't dying. It's evolving faster than your roadmap.
The loudest take is the wrong one. The verified read is more useful — and more unsettling: the ground under B2B SaaS is moving faster than most operators are repricing their assumptions. Here's where it's actually heading. Seven shifts. Every one sourced to a primary report, or cut. The rest is noise.
are resetting.
The 80% margin that defined SaaS isn't coming back.
For two decades, near-zero marginal cost gave software its 80% gross margins. AI broke that — every model call is real, variable compute. The average AI-product gross margin now sits around 52%, and the reset is structural, not a bad quarter. This is the shift that forces all the others.
gross margin · classic SaaS → AI-product average
The seat is becoming the wrong thing to sell.
When one agent does the work of ten people, charging per seat punishes your client for getting more value — and punishes you for delivering it. The live proof is already shipping: Zendesk, Intercom, and Salesforce now bill by resolution and outcome, not by login. IDC expects pure seat-based pricing to be obsolete by 2028.
is changing.
Software is moving from where you record the work to where it gets done.
The dashboard was the product for twenty years. The next layer doesn't wait for you to log in — it reads context, acts across systems, and surfaces a human only when judgment is required. Bain frames it as routine tasks moving from "human plus app" to "agent plus API" within a few years.
The moat was never the model. It's your data.
A generic model can write an email. It can't know which accounts are about to churn, which contracts will slip, or which approvals are required — that lives in your data, your workflows, your domain. As models commoditize, the defensible thing moves from the interface to the depth underneath it.
is changing.
Your client can now build the thing you sell.
AI cut the cost and time of building software by two to three times. A third of teams have already replaced at least one SaaS tool with a custom build, and most plan to build more. But it's concentrated — workflow automation, internal admin, BI. Deep, regulated, system-of-record software is exactly what they won't build — and gets more defensible for it.
"SaaS is dead" is the wrong question. Which SaaS is the right one.
The smartest people in the room disagree in public — one CEO calls SaaS dead, another calls that the most illogical thing in the world. The verified read isn't extinction; it's a sorting. Disruption is mandatory; obsolescence is optional. And the dividing line isn't how old the company is — it's how deep the workflow goes.
nobody puts on the slide.
95% of AI projects produce nothing. The 5% aren't who you'd guess.
MIT studied 300 deployments. Only about 5% delivered real P&L impact; the rest stalled. The failure isn't the technology — it's the approach: integration, data, and where the work was pointed. Buying and partnering beat internal builds roughly three to one. Adoption is high. Value capture is rare.
None of it works on data you can't trust.
Every shift above assumes the data underneath is clean enough for AI to use. Mostly, it isn't. Data privacy and security top the enterprise concern list, governance hasn't caught up, and the AI projects that fail mostly fail on data readiness — not model quality. The least glamorous finding is the most reliable one.
Where this leaves you
SaaS isn't ending. It's compounding — or it's being left behind. And the clock runs faster every quarter.
The engines that absorb these shifts as leverage were coherent before AI arrived. The ones that industrialize the wrong work, faster, were already broken — AI just made it visible.
None of these directions is in doubt. The only open question is whether your revenue engine is built to move at the speed they're arriving.
This is the part of my work that doesn't fit on a services page, so I'll put it here.
For most of my career, I spent ten months a year on the road. Different industries. Different stages. Different problems. The companies were never the same. The work always was. Walk in, see what was actually happening underneath what the leadership said was happening, build what was missing, transfer it to the people who would run it after I left, exit.
Some companies returned victorious. Some didn't. The ones that did had something in common, and so did the ones that didn't.
What I learned across all of it is simple enough to say in one sentence: revenue problems are structural, not personal. The engine is broken — and the engine can be architected. The standard I hold the work to comes from twenty-plus years of watching what produces that outcome and what doesn't.
Know the client's operation better than they expect you to.
Sell the outcome they are actually buying, not the feature you happen to have.
Build systems that do not depend on you to keep working.
Hire operators who want to win and will not tolerate compromise.
Exit when the work is done. Do not stay for the fees.
This isn't a methodology. It's a posture. It's the posture that produced 100% retention in an earlier era of B2B selling, before the industry decided that depth was inefficient and replaced it with handoffs.
Every engagement is held to this standard. So is the CEO who hires the work.
I am currently writing Truly Built to Win and Last: How B2B SaaS Founders, Leaders, and Investors Architect Companies That Compound — the book that walks the architecture in full. Alongside the book I publish The Revenue Data Debt™ Discussion, with Issue 01 available now: "The Bill Has Arrived."
I am only victorious when the Hero is victorious. I try to limit myself to 2–3 clients at a time. That kind of victory requires presence I can't fake at scale.
If that sounds like the work you need,
I'd want to hear about your situation.