If your procurement weighs analyst placement, Kore.ai wins that argument outright and it is not close. They were the only vendor holding both the 2026 Gartner quadrant and the Q2 2026 Forrester Wave when this page was first written; a Q3 Wave for employee services has landed since, and the honest update is that the gap widened rather than closed. A security review that asks for that placement will not accept a smaller vendor’s architecture page as a substitute.
Two things changed enough to be worth saying plainly. The first is that they now make a claim adjacent to ours, “complete audit trails, ready when you need them. Understand why the AI made a decision.” That is not marketing noise, and a comparison page that ignored it would be the kind of page nobody believes.
The second is what the two things are actually for. An audit trail is built for compliance: it proves, afterwards, that a decision was taken and on what basis. A retrieval trace is built for debugging: it shows the passages that were considered before the answer existed, what each scored, and which fell below the threshold and why. When a customer is quoted a wrong price, an audit trail tells you the assistant said it. The trace tells you whether the right document was never retrieved, or was retrieved and misread, two failures with nothing in common except the symptom, and two different fixes.
The difference underneath all of it is scale of engagement. Kore.ai is bought as a programme; this is bought as a script tag. If you have a contact centre, a systems integrator and a year, that programme will fit you better than this will.
We do not reproduce Kore.ai's pricing here. Check it with them directly, a price we cannot date is worse than no price at all.