This is The Agentic Commercial Model, Season 2: the build. This issue is the spine. Everything that follows this season hangs off it.
Let's define the thing properly, because most companies can't, and that's where the trouble starts.
Your commercial operating model is the machine that turns your product into cashflow. It's not your strategy deck and it's not your org chart. It's the working system underneath both: the motions your teams run, the cadences they run them on, the metrics that tell them how they're doing, and the incentives that tell them what to care about. Every company has one, whether it was designed or just accreted. And here is the problem this whole newsletter exists to name: most software companies have spent two years rebuilding the product and zero minutes rebuilding the machine.
The machine they're running was built as a line. A funnel. Awareness at one end, a close in the middle, a handoff, a quiet year, a renewal at the far end. Every motion, every cadence, every metric and every incentive in the classic SaaS operating model is a station on that line, and the line was a rational design for the world it was built in. Value was delivered annually, in the form of access. Information was scarce, so buyers needed educating. The close was the moment value was captured, so the close was the centre of gravity, and everything before it was expensive persuasion and everything after it was cheap maintenance.
Season 1 spent six issues establishing that every one of those conditions is now false. Value is delivered continuously, by a product that does the work. The buyer arrives pre-decided, educated by a machine. The revenue is no longer captured at the close; it's earned in the usage, realised in the outcomes, and lost in the silence between QBRs. Run a line-shaped operating model in that world and you are optimising stations on a railway your customer no longer rides.
So here is the replacement, in full. Not a framework with boxes. A machine with moving parts.
The new commercial operating model is a loop: land, realise, expand, advocate. It has no terminus, because the relationship has no terminus. Each turn of the loop deepens the customer's calibration into your product, which widens the moat; each turn generates proof, which feeds the machine that lands the next customer. The funnel emptied into a close. The loop compounds. Let me take you around it, stage by stage, with the operating cadence that makes each stage real, because a motion without a cadence is a intention, not a model.
Stage one: Land. The job is to start the loop, not to maximise the deal.
In the line model, landing was the whole game. The deal was the unit of success, so the motion optimised deal size, and the cadence was the pipeline review: weekly interrogations of stage progression toward a close that everything else waited on.
In the loop, landing has one job: get the customer into the loop as fast as possible, because nothing, not revenue, not calibration, not proof, starts until they're using the product on real work. The governing metric of the land stage is not deal size. It is time-to-first-value. A smaller land that activates in a week beats a triple-sized land that spends five months in procurement and two more in onboarding purgatory, because the loop compounds and the line does not. You are not maximising the entrance fee. You are minimising the distance to the flywheel.
If that sounds like theory, look at the fastest commercial ramp in software history. Cursor went from $1 million to $100 million of ARR in about twelve months, faster than any software company ever, and it did not hire an enterprise sales rep until well past the $200 million mark, running zero outbound until late 2025. The product landed itself, on real work, in days. Then watch what happened next, because this is the part that matters for enterprise operators: corporate buyers were roughly 25% of revenue at $400 million, 45% at $1 billion, nearly 60% at $2 billion, and about 75% of the $4 billion run-rate it passed this June. Cursor hired a President of Global Revenue in February 2026, when there was a $2 billion business to formalise. Read that sequence again. The enterprise motion arrived after the demand. Sales did not create the market; sales industrialised a loop that was already spinning. That is what the land stage looks like when it's built for the loop: the close isn't the entrance to the relationship anymore. Usage is.
The cadence: the weekly pipeline review survives, but its questions change. Not "what stage is the deal in," but "how many days from signature to first realised value, and what's blocking activation." The deal desk becomes a value desk: its job is not to approve discounts, it's to strip friction out of the first thirty days.
Stage two: Realise. The new centre of gravity of the entire model.
Here is the stage the line model never had, and its absence is where SaaS revenue now goes to die quietly.
The old model treated the period after the close as maintenance. Onboarding, a health score, a QBR if the account was big enough. The operating assumption was that value took care of itself because the customer had bought access and access didn't need managing. In a consumption and outcome world, that assumption is fatal, because unrealised value is unearned revenue, and unearned revenue is churn with a delay on it. The benchmark data on product-led funnels puts a number on the corpse pile: 40 to 60% of users who sign up never reach the activation milestone, the point of first real value. The industry calls them zombie users. And here is the operational scandal inside that statistic: only about a third of companies even track activation as a metric. Two-thirds of the industry cannot see the exact point where its revenue is being lost.
In the loop, realisation is the centre of gravity, the stage every other stage serves, and it runs as an operating discipline with the same rigour the old model reserved for pipeline. That means three things concretely. Outcomes are instrumented, which is the VRR work of the next issue: verifiable value events, defined with the customer, detected by the system rather than asserted in a slide. Value is reviewed on a monthly cycle, not a quarterly ceremony: the QBR, that theatre of retrospective slideware, is replaced by a shorter monthly value review built on live realised-value data, and its agenda is two questions: what value landed this month, and what's blocking the next tranche. And silence is treated as a fire alarm. In the line model, a quiet customer was a happy customer. In the loop, a quiet customer is a decaying one, because the product only earns while it's working.
The cadence: daily, the telemetry runs and flags stalls automatically. Weekly, a value standup inside the account team: activation blockers, usage anomalies, outcomes pending verification. Monthly, the value review with the customer. That rhythm is the heartbeat of the whole model, and everything in Issue 2.5's org design exists to protect it.
Stage three: Expand. Expansion is discovered, not sold.
The line model treated expansion as a sales event: an upsell motion, run against a renewal date, argued from a slide about unused entitlements. The loop treats expansion as a discovery process, because in an agentic product the expansion signal is sitting in the usage telemetry, published continuously by the customer themselves. Where usage depth is growing, the next use case is announcing itself. Where a team adjacent to the deployed workflow keeps touching the product's edges, the expansion path is drawn on the map before anyone sells anything. The rep's job is not to invent demand at renewal time; it is to read the demand the loop is already generating and formalise it.
Datadog is the cleanest public illustration of the mechanics. The motion starts with developers adopting one monitoring product on a free or low-cost tier, then expands across the observability platform, product by product, team by team. The product opens the door; sales widens it. The result of running that loop for a decade: $3.4 billion of revenue in fiscal 2025, up 28%, and 603 customers now generating over $1 million in ARR, up from 462 a year earlier. Those 603 did not sign million-dollar deals. They became million-dollar customers, one discovered use case at a time. That is what expansion looks like as a loop stage rather than a sales event: continuous, telemetry-led, and increasingly detached from any calendar.
Which is the other structural change: the renewal stops being an event. When expansion is continuous and value is verified monthly, the annual renewal negotiation, that adversarial ritual where twelve months of silence gets litigated in one meeting, dissolves into the operating rhythm. Commitment refreshes become administrative because the value case was never allowed to go stale.
The cadence: quarterly expansion planning per account, built from the usage map, owned jointly by the seller and whoever owns realisation. The input is telemetry, not intuition. The output is a named next use case with a time-to-value target, which feeds straight back into the land motion. The loop, looping.
Stage four: Advocate. Engineered, not hoped for.
The line model treated advocacy as a byproduct: a reference call if you were lucky, a logo slide if legal allowed it. In the loop, advocacy is a manufactured output with its own motion, and Season 1's GTM Collapse issue explains why it graduated from nice-to-have to load-bearing: the shortlist now forms inside AI answers, and those answers are built from third-party evidence: reviews, citations, named case studies, the verifiable public record of value delivered. Advocacy is no longer a favour your happiest customer does you. It is the top of your funnel, feeding the machine that decides whether the next buyer ever learns you exist.
So the loop closes on itself. Every verified outcome from the realise stage is a harvestable proof asset: a named metric, a review, a case study, a citation-ready artefact. The motion is to harvest systematically at the moment of realised value, when the customer's enthusiasm is highest and the evidence is freshest, and to place that proof where the machine reads it. Which means the loop's output lands the next customer pre-decided, at a lower acquisition cost, with a shorter land stage, and the wheel turns faster on every revolution. This is the compounding the funnel never had: the line consumed spend to fill its top; the loop generates its own.
The cadence: proof harvesting is an agenda item in every monthly value review. A quarterly advocacy audit asks one question: for our top realised outcomes this quarter, where does the public, machine-readable evidence live, and if the answer is nowhere, why did we let the proof evaporate?
The operating rhythm, assembled
Put the four stages together and the new machine has a drumbeat, and it is nothing like the old one.
The line model ran on an annual calendar with a weekly pipeline meeting bolted on: close, onboard, go quiet, panic at renewal. The loop runs on a stacked rhythm. Daily: telemetry, automated, watching activation and usage depth. Weekly: the value standup and the reshaped pipeline review, both asking versions of the same question, which is how fast is value moving. Monthly: the customer value review, realised value verified and the next tranche unblocked, proof harvested. Quarterly: expansion planning from the usage map, and the advocacy audit. Annually: almost nothing, and that is the point. The model that concentrated its truth-telling into one renewal meeting a year has been replaced by one that tells the truth every day, in small doses, while it can still be acted on.
Notice what this rhythm costs: roughly nothing new. It is the same commercial capacity the line model spends today, reallocated. The hours currently burned on stage-three pipeline theatre, renewal panic and QBR slideware fund the value standups and the realisation discipline. This is not a bigger machine. It is the same machine, pointed at where the revenue actually is.
The spine, and what threatens it
That is the commercial operating model for the agentic era: a loop that lands fast, realises relentlessly, expands by discovery, and advocates by design, on a cadence that compounds. Every issue remaining this season builds a component of it. The instrument that verifies the value. The GTM that wins the machine. The pricing that captures what's proven. The comp that pays for the motion instead of the moment.
But before any of that, there's a harder problem, and I want to be straight about it because I've watched it kill this model in practice more than once. A loop drawn on a slide reverts to a line within two quarters unless the organisation is physically shaped to hold it. The old model doesn't die when you announce the new one. It reassembles itself quietly, through reporting lines, through who owns which number, through the wall between selling and serving that everyone agrees to dissolve and nobody actually dissolves, because structure eats intention for breakfast.
That is Issue 2.5: the org design that stops the line growing back. It's the hard half of this issue, which is exactly why it gets its own.
Build the model the product deserves.
First published in the Agentic Commercial Model newsletter on LinkedIn, August 4, 2026. Read the original on LinkedIn.