No scripted decision trees. No keyword matching. Kai understands what customers actually mean - and gives them a real answer in seconds.
Multichannel from one inbox
Widget, email, WhatsApp, Instagram, Messenger, Telegram. A customer who starts in chat and follows up by email gets seamless context.
6 channels out of the box
Web widget, email, WhatsApp, Instagram, Messenger, Telegram.
80+ languages, auto-detected
No language-specific bots to maintain. Kai speaks the customer's language.
Live in under 10 minutes
Paste your domain. Point Kai at your help center. Done.
SOC 2 Type 2 + GDPR
Topic restrictions, EU residency, full audit logging.
Smart escalation
Investigated before a human sees it.
If Kai can't resolve, it escalates to Kai Resolve. The agent investigates with MCP tools, source context, ticket history, APIs, and approved read-only data access, then routes the issue to Kai Code, Kai PM, a safe operational action, or a Gleap Inbox Task for a human.
When Kai isn't resolving, it's helping your team. Drafts replies, suggests articles, translates messages, summarises long threads - right inside the inbox.
Airspeed switched from UserSnap to Gleap to support multiple platforms and more than 75,000 users with a stronger feedback and support workflow.
Gaylan Greenawalt
PM, Airspeed
Choose your AI brain
Auto on Team. All AI models on Pro.
Frontier models
OpenAI GPT, Anthropic Claude, Google Gemini, xAI Grok - frontier model families, available for Pro workflows that need manual model choice.
Open source & specialist
Llama, Mistral, Qwen, DeepSeek, and other open-source or specialist models. Pro teams can pick cost-efficient models for long context and multilingual support.
Bring your own LLM
Enterprise plans support custom model endpoints. Full control over cost, latency, and data residency. Bring your own cloud model or run it on your own hardware.
Kai learns from everything you have
Seven content sources, one coherent answer
Snippets
Add reusable answers, policy blocks, and product notes Kai can cite consistently across conversations.
Websites
Point Kai at your domain. It indexes pricing pages, feature pages, FAQs - anything public.
Help articles
Native Gleap help center plus imports from Intercom, Zendesk, and Freshdesk. Cited in every answer.
Files
Upload PDFs, product manuals, onboarding guides, and internal SOPs. Kai indexes and retrieves automatically.
Code repositories
Index connected repositories directly. Kai finds answers in code when docs haven't caught up yet.
Notion
Read your Notion pages and databases. Respects permissions. Stays in sync as docs evolve.
YouTube
Index videos and playlists so Kai can answer from tutorials, webinars, and product walkthroughs.
Team
$149/mo
Kai support with token-based AI usage and unlimited seats.
Kai included
AI usage billed by actual tokens
Auto model choice
Unlimited seats
All support channels (widget, email, WhatsApp, Instagram, Messenger, Telegram)
Custom domains
Priority human support
Pro unlocks all AI models, Kai PM, Kai Code, and Kai Resolve
The same 20 questions account for 80% of typical SaaS ticket volume - password resets, billing, how-tos, basic troubleshooting. Kai reads your snippets, websites, help articles, files, code repositories, Notion workspace, and YouTube content, then answers using your actual product context - not keyword matching. Most teams see 60–85% deflection within two weeks.
How is Kai different from Intercom Fin?
Different pricing models. Fin is priced at roughly $0.99 per resolution. Kai usage is billed by the actual tokens used - on average about $0.04 per message, with a conversation running roughly 4.3 messages - so you pay for the AI work performed instead of a per-resolution success fee. Team uses Auto as the model choice. Pro unlocks all AI models across OpenAI GPT, Anthropic Claude, Google Gemini, xAI Grok, and open-source models. On Pro, unresolved issues can route to Kai Resolve for investigation before involving a human.
How is Kai different from Zendesk's AI agent?
Kai is built into a platform of specialist AI agents that can include investigation and routing (Kai Resolve), self-driving development (Kai Code), and an AI product manager (Kai PM). Zendesk's AI stops at Tier 1 deflection. Kai also captures full session replay, console logs, and network requests on bug-related tickets - Zendesk doesn't offer that natively. Pricing starts at $49/mo for Starter with Kai; Team is $149/mo with unlimited seats and Auto as the model choice; Pro adds all AI models, Kai PM, Kai Code, and Kai Resolve.
What channels does Kai work on?
Six channels out of the box: in-app web widget, mobile (iOS, Android, React Native, Flutter), email, WhatsApp Business, Instagram DMs, Facebook Messenger, and Telegram. All land in one unified inbox - a customer who starts in chat and follows up by email gets a seamless thread. Kai responds in 80+ languages based on detected language.
What happens when Kai can't resolve a ticket?
On Pro, it escalates to Kai Resolve first. Kai Resolve investigates with MCP tools, source code, ticket history, APIs, and approved read-only data access. Confirmed bugs go to Kai Code, feature or product gaps go to Kai PM, safe operational issues can be resolved directly, and judgment calls become Gleap Inbox Tasks for a human with the investigation dossier attached.
Does Kai work with my existing knowledge base?
Yes. Kai imports from Gleap's native help center plus Intercom, Zendesk, and Freshdesk knowledge bases, and also reads snippets, website crawls, uploaded files, Notion workspaces, YouTube playlists, and connected code repositories. Most teams have Kai live in under 10 minutes.
How does Kai handle sensitive or regulated topics?
Configure topic restrictions and escalation rules per subject - medical, legal, financial - so Kai declines to advise on restricted topics and either routes to a human or returns a pre-approved response. Kai is SOC 2 Type II certified and GDPR compliant, with EU and US data residency on Enterprise.
Your customers deserve an instant answer. Give them one.