On this page
Put Coworker to work on your stack.
Connect Salesforce, Slack, Jira and run your first agent in minutes.
Get started freeFree for 14 days
Enterprise AI
Kore.ai Pricing in 2026: Sessions, Seats, and the Numbers Kore Won't Publish
Kore.ai pricing in 2026: the official session and seat billing mechanics, the widely reported $50-150 tiers and $0.20 sessions, and real enterprise deal sizes.
Kore.ai pricing is a study in confirmed mechanics and unpublished numbers. Officially, via Kore.ai's own billing documentation: the platform sells in Essential, Advanced, and Enterprise tiers, bills conversational AI per 15-minute session, and bills contact-center products per agent seat. Unofficially, because kore.ai/pricing itself returns a 404 (verified July 17, 2026): the widely reported figures are roughly $50 per month for Essential, $150 for Advanced, $0.20 per 15-minute session pay-as-you-go, and $300,000+ per year for typical enterprise contracts, consistent across six or more independent review sites but never confirmed by Kore.ai.
Here is what is official, what is reported, and the session math that actually decides your bill.
Kore.ai pricing at a glance
| Item | Figure | Status |
|---|---|---|
| Essential tier | ~$50/mo (annual) | Reported; tier name official |
| Advanced tier | ~$150/mo (annual) | Reported; tier name official |
| Enterprise tier | Custom quote | Official (no number anywhere) |
| Pay-as-you-go session | ~$0.20 per 15-min session | Reported (consistent across sources) |
| New-account credits | $500, 90-day validity | Reported |
| Standard support add-on | ~$1,000/mo | Reported |
| Typical enterprise deal | $300,000+/yr | Reported (3+ sources agree) |
| G2 perceived cost | 5 of 5 dollar signs (highest bucket) | G2 data |
The honest label matters here: those dollar figures are unusually consistent across independent sources (Cloudtalk, Vida, and others), which makes them reasonable planning numbers, but no Kore.ai-owned page publishes any of them. Treat every one as "widely reported," not list price.
The session billing mechanics (official, and where budgets go wrong)
Kore.ai's docs confirm the core mechanic: Automation AI (including Search AI) bills per 15-minute conversation session. The official example is the trap in miniature: a 31-minute conversation is three billable sessions (0-15, 16-30, 31-end). A support conversation that idles while a user hunts for an order number keeps consuming sessions.
Two more official mechanics with budget consequences:
- Multi-bot routing splits the meter. Sessions are tracked per app; when a parent bot routes a query to a linked child bot, usage bills against the child. Teams running multi-bot architectures routinely underestimate total spend because no single dashboard line shows it.
- Contact Center AI and Agent AI bill per seat, with two models: named seats (every agent counts, shifts irrelevant) or concurrent seats (peak simultaneous logins). A 3-shift operation of 150 agents is 150 named seats but might be 50 concurrent, a 3x swing available in negotiation.
Voice (ASR/TTS via the Voice Gateway) bills separately on top and scales with call minutes.
Quick worked example at the reported rates: a support bot handling 5,000 conversations a month, averaging 20 minutes each, generates 10,000 billable sessions (two per conversation), which is about $2,000 a month at the reported $0.20 rate before voice or seats. Trim average handling to under 15 minutes and the same volume costs half. On this platform, conversation design is cost engineering.
Coworker
Put Coworker to work on your actual stack
Connect Salesforce, Slack, Jira and run your first agent in minutes.
What changed in 2026: the Artemis relaunch
On May 21, 2026, Kore.ai relaunched its XO Platform as the Agent Platform, Artemis edition: an "AI-native" rebuild with a proprietary Agent Blueprint Language, launched Azure-first, covered independently by VentureBeat as a direct challenge to Microsoft and Salesforce. Two buyer-relevant facts: no new pricing model shipped with Artemis (the session/seat/tier structure predates and survives it), and the scale ceiling is real: VentureBeat cites a bank running Kore.ai's AI for Work across 135,000 employees and contractors.
On AI for Work specifically (the employee-facing product): its official page contains no pricing at all, and no per-employee rate card surfaced anywhere in this research. It is sold as a custom enterprise module. Anyone quoting you a specific AI for Work per-seat price is guessing.
Is Kore.ai worth it? The honest read
The strengths: genuine enterprise scale (the 135K-employee deployment above), officially documented and flexible billing mechanics, deep contact-center capability, and a platform architecture that analysts place in the same conversation as Microsoft's and Salesforce's agent stacks.
The caveats, consistently cited in reviews: G2's perceived-cost rating sits in the highest bucket, implementation is a real project (G2's average runs about 2 months, but analyses of complex enterprise setups describe considerably longer), and the pricing opacity itself is a cost: you cannot model a Kore.ai bill without a sales cycle, and the session mechanics reward careful conversation design.
How Kore.ai compares in 2026
| Platform | Price | Model |
|---|---|---|
| Kore.ai | Reported $50-150/mo tiers; enterprise $300K+/yr | Sessions + seats, quote-only at scale |
| Aisera | Quote-only; median ~$108K/yr reported | Modules + users + volume (breakdown) |
| Moveworks | Quote-only; reported $50K-130K+/yr | Per employee/year (breakdown) |
| Zendesk AI | Seats + $50 Copilot + per-resolution | Layered (breakdown) |
| Coworker | $0 free, $29.99/user/mo flat, self-serve | Seat |
The pattern across the enterprise agent platforms, documented in our quarterly price index: the more agentic the platform, the more likely the real price is a negotiated contract plus a usage meter.
Where Coworker fits
Kore.ai sells to organizations building governed fleets of customer-facing and contact-center agents at six-figure scale, and it is credible there. Coworker solves the other end of the problem: teams that want AI agents working across their own tools (Slack, Jira, HubSpot, Google Workspace, GitHub, 50+ connectors) today, self-serve, at a flat $29.99 per user per month with no session meters or sales cycle. If you are budgeting a contact-center transformation, shortlist Kore.ai. If you are trying to give your team an AI coworker this week, get started free, or scan the best enterprise AI platforms for the full field.
Frequently asked questions
How much does Kore.ai cost? Kore.ai publishes no prices; its pricing page 404s. Widely reported figures: Essential around $50 per month, Advanced around $150 (annual contracts), pay-as-you-go around $0.20 per 15-minute session, and enterprise contracts from roughly $300,000 per year. Official docs confirm only the tier names and billing mechanics.
How does Kore.ai's session billing work? Automation AI bills per 15-minute conversation session, officially: a 31-minute conversation equals three sessions. Contact Center AI and Agent AI bill per agent seat, in named or concurrent models. Voice capabilities bill separately by usage.
Does Kore.ai have a free trial? Reported: new accounts get $500 in credits valid for 90 days (roughly 2,500 test sessions at the reported rate), with a reported $100 minimum purchase to start on pay-as-you-go. Neither figure appears on an official Kore.ai page.
What is Kore.ai Artemis? The May 2026 relaunch of Kore.ai's platform (formerly XO Platform) as the Agent Platform, Artemis edition: an AI-native rebuild with its own agent-definition language, launched first on Azure. No pricing changes were announced with it.
How much does Kore.ai AI for Work cost? No per-seat or per-employee rate exists publicly. It is sold as a custom enterprise module on the Agent Platform; the largest disclosed deployment runs to 135,000 employees at one bank.
Is Kore.ai expensive? G2 reviewers rate its perceived cost in the highest bracket, and reported enterprise deals start around $300,000 per year. For small teams, the reported $50 to $150 monthly tiers exist, but the platform's center of gravity is large enterprise.
Related reading
Ready to get started?
Put Coworker to work inside your actual stack
Connect Salesforce, Slack, Jira, whatever you use, and run your first agent in minutes.
Free for 14 days