I keep coming back to a pattern that doesn't yet have a clean name. The usage charts are healthy. Logins look fine if you squint, because agents are now generating most of the activity. Calls per week, actions per seat, traces per workflow: all green. Then the renewal conversation arrives, and the customer says something polite but soft. The champion who used to walk you through the value story sounds less certain. Procurement asks for outcome numbers your dashboard wasn't built to produce.

That gap, between rising activity and softening renewal narrative, is the quiet risk hiding inside agentic SaaS. Brian Gallagher named it hollow usage: more adoption, less revenue, because agents don't defend you. The phrase landed for me because it points at something most product roadmaps haven't priced in. The user we built for is being replaced by an agent we didn't underwrite.

The advocate was the moat

For two decades, SaaS renewal economics rested on a quiet assumption: somewhere inside the customer's organization there was a human who experienced the product, attached their judgment to it, and translated what it did into language a CFO would sign. That human was the moat. Not the data, not the workflow, not the integrations. The advocate. They built the deck, told the story, and absorbed the awkward parts of the conversation.

When an agent takes over those workflows, the advocate doesn't vanish at once, but their proximity to the product fades. They stop seeing what it does. They lose the small daily texture they used to translate upward. And the renewal review gets thinner each cycle. You can watch this happen and still post your best usage numbers ever.

What Adobe and WPP are actually building

The first companies treating this as a product problem are starting to show what the fix looks like. When Adobe announced general availability of CX Enterprise Coworker, the framing was unusually specific. Anjul Bhambhri said most organizations are "struggling to translate AI adoption into measurable business results," and the product was built for exactly that translation.

Read past the marketing language and there is a design choice underneath. Orchestration agents that coordinate analytics, content, journey, and operations are also expected to surface outcomes against defined goals as part of the workflow itself. Not as a quarterly business review export. As a structural emission.

The WPP partnership is the part to study. Adobe and WPP have been knitting paid media and owned experience into a single operating layer for client growth, with measurement designed into the joint output, not bolted on after. The reason to read that partnership as a design pattern rather than a press release is the shape of what it produces. The workflow generates artifacts that a marketing leader can use, inside their own company, to justify spend without needing the vendor's narrative to do the lifting.

Renewal-grade artifacts

That is the category worth naming. Renewal-grade artifacts. The product emits them as a byproduct of working. They are not invoices or usage logs. They are documents that travel through the customer's organization and arrive at the people who control the budget already loaded with the answer to what did we get for this.

A few concrete shapes:

  • Outcome summaries that pair each automated journey with what it changed in conversion, retention, or basket value, attributed cleanly enough to enter a finance review without rework.
  • Attribution reports that name the workflow, the agent action, and the revenue or pipeline downstream, in a format the customer's BI team will accept without rebuilding the join.
  • Decision traces that show, in business language, the choices the system made and the lift they produced against the baseline that existed before the product was bought.
  • Account-level retention or expansion deltas tied to the segments the product actually touched, exportable as evidence inside the customer's own QBRs with their stakeholders.

The common feature is that the customer's internal advocate, even an attenuated one, never has to reconstruct the story. The product has already made the argument in a form that can be quoted in front of a CFO.

Where bargaining power actually moves

What this changes is where bargaining power lives. For a long time, value capture in B2B software was anchored to interface ownership. Your screen was the surface where the customer's team spent their hours, and that time built loyalty, fluency, and political support inside the account. Agents end that arrangement. The interface is no longer where the value is experienced. The output is. The relationship moves from interface ownership to evidence ownership, and the vendor who owns the cleanest evidence trail owns the next conversation.

This is the internal cousin of a pattern showing up at the edges of recent AI work. The Evidence Contract piece looked at the external version: probabilistic systems meeting deterministic institutional gates, where logs and traces are what the receiving system accepts. Renewal is the same shape pointed inward. The customer's finance team is the institutional gate. They accept evidence, not enthusiasm. The vendor that has shaped its product to produce that evidence by default has done the design work that pays in year three.

What this means for product and revenue leaders

For product leaders, this is a roadmap constraint, not a customer success project. The question to put on the next planning page is plain: what artifact does our product emit, without anyone asking, that a buyer can take to their own internal review and use to win the budget defense? If the answer is "the CSM builds a slide," the product hasn't done its job yet. If the answer is "the workflow itself produces a document the finance team will accept," the moat is forming.

For revenue leaders, the implication is that renewal defense stops being a relationship sport. The classic playbook of executive sponsor visits, QBR storytelling, and ROI estimates assembled from logs was already softening. Hollow usage breaks it. What replaces it is a renewal motion built around artifacts the product has been emitting all year, presented to a buyer who has been reading them, with the vendor in a confirming rather than persuading posture. That is a different muscle, and the teams who build it first will price differently than the ones who keep selling stories.

Renewal is the ultimate version of the Day-90 Test: what survives after the launch energy fades. It also sits on the same fault line as integration debt, where AI output fails not because the model was wrong but because the next state in the customer's system wasn't designed to accept it. Renewal-grade artifacts are the decision-layer assets the buyer needs to keep the system inside their organization. Treat them as part of the workflow, not part of the sales motion.

The next twelve months

My read is that the next twelve months of agentic SaaS will sort companies on a single line. Some will keep optimizing for usage metrics and discover, late, that the human who used to defend their renewal has been quietly replaced and never told them. Others will redesign their product around the evidence it emits, and find their renewal conversation has gotten shorter and easier because the argument was already on the buyer's desk before the meeting started.

When agents do the work, attention is no longer the moat. The moat is what your product writes down, on its own, that a customer can hand to their CFO without picking up the phone.

Sources and further reading