We hold our clients' campaigns to a simple rule: if a claim cannot be sourced, it does not ship. The same rule applies to us. Below is every market figure we quote, where it comes from, and how current it is. Where a figure derives from licensed analyst research, we identify the category of source here and provide the full licensed citation in proposals and scoping calls, as the licence requires.
LAST REVIEWED · JULY 2026 · QUESTIONS: [email protected]
Enterprise buying committees routinely stall between problem definition and supplier selection — most often at a stakeholder the vendor never engaged.
The buying group has roughly doubled in a decade: around thirteen people inside the account shape the decision, and roughly nine more outside it — analysts, peers and advisors the vendor rarely meets. Forrester's 2026 outlook reports buying groups continuing to expand as organisations scrutinise AI-related purchases.
Committees research, shortlist and largely decide before a seller is invited in. Two-thirds of B2B buyers now say they prefer a rep-free experience for most of the journey.
Post-purchase regret is the norm, not the exception — driven by consensus gaps the selling process never resolved. It is why validation, references and evidence now outweigh promise.
Forrester's 2026 predictions estimate B2B companies will lose more than ten billion dollars through ungoverned use of generative AI — unsourced claims, brand damage and rework. It is the reason no Conversera agent output reaches a buyer without a named human approving it.
Our own claim, derived from the figures above and from buying-group coverage analysis across Conversera engagements: most demand programmes are built to find the person who fills in the form, and that person rarely decides alone.
If you believe any figure here is out of date or mis-stated, tell us — [email protected] — and we will correct it or take it down. That is the standard we sell; it would be strange not to live by it.