TL;DR
The story this cycle is that the vendor-controlled buying surface has officially collapsed: IDC's 2026 data shows buyers comparing products inside AI tools while vendor websites and salespeople bleed out as evaluation sources, Carbon finds that 94% satisfaction no longer prevents 59% of buyers from re-running comparisons, and Vertice's Ana brings an autonomous AI negotiator to the buy side. Last cycle's theme — proof over promise — now has teeth: legal teams are writing AI clauses that formally challenge unsupported vendor claims, and your next renewal may be negotiated by an agent that has seen every comparable contract.
The vendor-owned funnel is gone
The two channels vendors fully control — the website and the sales team — are no longer where enterprise evaluation happens, and the drop is steep enough to be a structural break, not a trend line.
"vendor website usage as an evaluation source has dropped from 66% to just 37%, as buyers increasingly treat vendor-published content as marketing rather than something to lean on." — IDC blog: Your Buyer Has Moved. Has Your GTM?
IDC's 2026 buyer survey found 68% of buyers using AI always or often to compare products across vendors, with only 33% relying on the vendor salesperson as an active evaluation source [idc-ai-mediated-buyer-journey-2026]. This extends last cycle's G2 finding that chatbots assemble the shortlist before a vendor is in the deal — but IDC adds the sharper detail that buyers trust industry experts and analysts (45%) nearly as much as they trust AI tools (46%) for vendor evaluation, meaning validation runs through third parties, never back to the vendor's own pages [idc-ai-mediated-buyer-journey-2026]. The kicker for founders: only 3.1% of vendor organizations have reached IDC's "optimized AI maturity" stage, so sellers are structurally behind their own buyers [idc-ai-mediated-buyer-journey-2026].
What to watch: Whether the vendor-side maturity gap closes in IDC's next annual data drop, and whether website evaluation usage keeps falling below 37%.
Satisfaction is no longer a moat
The renewal has quietly become a standing auction: buyers are happy with you and comparing you anyway, because the effort of side-by-side comparison has collapsed to a prompt.
"'Happy' doesn't mean settled." — Carbon 2026 B2B Buyer's Report findings
Carbon's study of 189 B2B decision-makers found 94% satisfied with current partners while 59% plan to reconsider at least one within 12 months — and 57% already use AI specifically for comparing providers side by side [satisfied-but-reconsidering-b2b-buyers]. The switch triggers are exactly the things AI makes easiest to verify: better value for cost (59%) and better proof of results or ROI (39%) top the list [satisfied-but-reconsidering-b2b-buyers]. This is the behavioral micro-foundation of last cycle's renewal-cliff dynamic — the difference is that Carbon shows the re-evaluation isn't triggered by dissatisfaction, it's ambient. Vendors whose value story isn't externally evidenced are precisely the ones an AI comparison will disqualify.
What to watch: Whether "re-evaluation rate among satisfied customers" becomes a tracked churn metric that displaces traditional NRR forecasting.
The buy side arms itself: agents now negotiate
Procurement is agentifying in earnest — Gartner projects 90% of B2B buying will be AI-intermediated by 2028, and the first at-scale buy-side negotiation agent just shipped.
"She brings market evidence, probabilistic outcome modelling, and the pattern recognition of hundreds of thousands of expert negotiations to every contract. She knows what good looks like." — Eldar Tuvey, Founder and CEO, Vertice — Vertice launches Ana
Vertice's Ana, launched September 8, 2026, has run 4,000+ live software negotiations at 18% average savings, trained on a dataset of 250,000+ negotiated contracts across 32,000+ SaaS vendors [agentic-procurement-commerce-shift]. The gap it monetizes is the procurement capacity shortfall vendors have long enjoyed: 1 in 7 software contracts auto-renews without oversight, and 3 in 4 in tail spend renew unreviewed [agentic-procurement-commerce-shift]. With customers including ARM, Brex, and Twilio, this is enterprise-grade, not experimentation — and it lands on vendors as price compression, faster cycles, and forced transparency. It also connects upward to the governance floor from last cycle: IDC found 90% of enterprise legal buyers now write AI clauses into guidelines, creating formal triggers to challenge unsupported vendor claims [idc-ai-mediated-buyer-journey-2026] — the private-sector mirror of the GSA clause.
What to watch: Whether benchmark-armed negotiation agents become standard on both sides of the table, and what that does to list-price discipline across SaaS categories.
What surprised us
- Legal, not procurement, may be the real gatekeeper. IDC's 90% legal-AI-clause stat means unsupported claims now trigger formal challenges written into procurement guidelines [idc-ai-mediated-buyer-journey-2026]. Marketing claims are becoming contract exposure.
- The negotiation asymmetry was already priced in. Ana's example wins — like rebutting a 20% license-count increase with usage data for an 11.8% below-prior-year renewal — show vendors were raising prices on unreviewed contracts, and agents are simply auditing the gap [agentic-procurement-commerce-shift].
- Switch drivers are boring on purpose. Value for cost (59%) and proof of ROI (39%) beat innovation (38%) and reputation (35%) — consistent with last cycle's finding that 62% buy for efficiency while only 12% buy for transformation [satisfied-but-reconsidering-b2b-buyers].
- Only 14% of buyers use no AI at all in evaluation [satisfied-but-reconsidering-b2b-buyers]. The non-AI buyer journey is now a rounding error.
Open threads worth a vote
- [watch] IDC AI-mediated buyer journey annual data refresh — steer which maturity and channel-decline stats we track at the next drop.
- [question] Outsell 2026 B2B Buyers Study — help locate the primary publication so we can extract its AI-in-vendor-selection findings.