This week, enterprise software stocks took a meaningful hit. ServiceNow fell 11.4% after earnings raised agentic risk concerns. The software ETF IGV is down 23% YTD. The proximate cause is the so-called seat-count crisis: AI agents are executing the work that human seat-holders used to do, compressing per-user license counts and triggering a re-rating of SaaS valuations across the board [1].
I want to talk about the second-order effect that almost nobody is focused on yet.
The seat-count story is getting coverage as a pricing model story — as if the main problem is that Salesforce charges $150 per seat and needs to figure out how to charge $2 per conversation instead. That transition is real and disruptive. But it’s a symptom, not the diagnosis. The deeper structural change is this: when AI agents replace human operators, the “buyer” itself changes. And most founders I talk to have not fully internalized what that means for how they build, position, and grow.
Human buyers discover software through a recognizable set of channels: a colleague’s recommendation, a G2 review, a LinkedIn ad, a conference, a case study that Google surfaced when they were researching a problem. Every PLG flywheel, every content marketing strategy, every outbound sales motion is engineered to intersect with these discovery paths.
AI agents discover software differently. They don’t browse. They retrieve. When an enterprise procurement agent is evaluating vendor options in a category, it synthesizes its training corpus, queries configured retrieval systems, and produces a shortlist from the information available in its environment. It doesn’t click ads. It doesn’t attend webinars. It doesn’t follow your founder on X [2].
If your company’s positioning, product information, and category authority don’t exist in the data environments those agents draw from, you are invisible to that discovery layer — regardless of how good your product is and regardless of how well your human-facing marketing performs.
This is the Machine Relations gap that will define which companies survive the agentic transition and which don’t. Machine Relations is what happens when you take your brand positioning seriously as a machine-readable infrastructure problem, not just a human communication problem. The companies that win in an agentic marketplace will be the ones with structured, retrievable, authoritative positioning in every context an AI system might evaluate them.
Here’s the part that I think founders are underweighting. Per-seat pricing wasn’t just a revenue model. It was a strategic lever. It created natural expansion motions, gave customer success teams utilization metrics to act on, and created a growth ceiling tied to headcount — which meant your best customers were your largest customers, and your ARR grew with them.
Outcome-based and usage-based pricing doesn’t have those properties. Outcome-based pricing means your revenue goes down as your agents get more efficient. Usage-based pricing means revenue spikes are unpredictable and hard to forecast against. Neither model creates the same alignment between “customer success” (in the human sense) and “revenue growth” (for you) that per-seat delivered [3].
The commercial architecture of the next generation of enterprise software needs to be rebuilt around this reality. The companies that figure it out first will have a genuine structural advantage — not just a pricing advantage, but a retention and expansion architecture that actually maps to agentic value delivery.
I think the following is likely:
The research and shortlisting layer becomes machine-controlled at most enterprise companies. Procurement agents will pre-qualify and rank vendors before any human is involved in most buying categories. The vendors who appear in that shortlist will have built Machine Relations infrastructure; the ones who don’t will be systematically underrepresented in enterprise pipelines — not because they have worse products, but because they’re invisible to the machine doing the research.
The champion relationship gets abstracted. The human champion who used to carry your renewal through an org change will be partially replaced by agent configurations and workflow dependencies. If your product’s value is embedded in an agent workflow, removal is operationally expensive. If it’s not, removal is frictionless. Your retention strategy in an agentic world is about making your product the dependency inside the agent’s operational mandate, not the preferred tool of a human advocate.
Category ownership becomes the primary moat. When AI systems answer “what’s the best solution for X?”, they answer based on what entities they recognize as authoritative in that category. Building that recognition — through structured publishing, citation authority, and entity clarity — is the long-term competitive advantage. This is what Machine Relations exists to operationalize. It’s not a content marketing strategy. It’s infrastructure [4].
One more piece of this that matters: Futurum Group’s February research shows that CFOs have moved their AI ROI measurement from productivity proxies to direct P&L impact — revenue growth and margin improvement nearly doubled as primary measurement criteria year-over-year [5]. That shift has a direct implication for founders: if your pitch is “save your team 4 hours a week,” you are entering a losing conversation with the enterprise buyer who is now trained to ask “what’s the line item on the P&L?”
The companies that get funded, retained, and expanded in the next 18 months will be the ones that can answer that P&L question with evidence — not efficiency narratives. Building that evidence layer means building for agentic measurement from the ground up: structured outcomes, machine-readable value proof, and pricing that scales with verifiable results rather than human headcount.
The seat is dead. The question is what you build on its foundation.
Run an AI visibility audit to see how your company appears in AI systems today.
Sources
[1] Financial Content Markets — The Seat-Count Crisis: How AI Agents Triggered the 2026 Software Sell-Off
[2] Anthropic Research — Measuring Agent Autonomy
[3] Flexera — From Seats to Consumption: Why SaaS Pricing Has Entered Its Hybrid Era
[4] Cloud Wars — Enterprise AI in 2026: Scaling AI Agents with Autonomy, Orchestration, and Accountability
[5] Futurum Group — Enterprise AI ROI Shifts as Agentic Priorities Surge