16 профілів, згрупованих за категорією. Фільтруйте або шукайте; розгортайте картку для переваг/слабких місць і джерел.
Enterprise AI search + Work AI assistant/agent platform
Glean sells to mid-market and (mostly) large enterprises — IT/knowledge-management and increasingly Chief-AI-Officer / platform teams — who need a single permission-aware AI layer over all their SaaS tools (Slack, Confluence, Jira, Drive, GitHub, Salesforce, ServiceNow, etc.). It started as enterprise search (index-everything, permission-aware retrieval) and has expanded into "Work AI": a chat assistant (Glean Assistant), embeddable agents inside Slack/Teams/Salesforce/Zendesk, a no-code agent builder + agent library, and an Enterprise Agent Development Lifecycle for governing agents at scale. Buyers are large orgs (contracts typically minimum ~100-250 seats) that already have fragmented SaaS sprawl and want one governed answer layer rather than per-tool search; it is generally too expensive/heavy for small teams and has no self-serve signup.
Ціна Standard (legacy/base search tier): custom/undisclosed · Enterprise (legacy tier, AI/chat bundled): custom/undisclosed · Enterprise Flex (current model, per docs.glean.com): custom/undisclosed per-seat rate + FlexCredits consumption
За що платятьGlean Assistant / chat over Fast ModeThinking Mode / Adaptive Reasoning / premium-model queriesAgent creation, agent library, and no-code Agent BuilderAgent Governance / Enterprise Agent Development Lifecycle (ADLC)Active data & AI governance (continuous scanning/remediation of sensitive data, coverage across 100+ sources, dashboards)
Сильні сторони · слабкі місця
Переваги- Enforces source-system permissions on every query and every agent action, so answers/search results never leak content a given user couldn't already see in the underlying tool — this permission-aware retrieval is Glean's central trust pitch, deeper than most generic wiki-chat tools.
- 275+ pre-built connectors is the largest breadth in the category, reducing integration engineering the buyer would otherwise have to do themselves.
- Goes beyond retrieval into action: a no-code Agent Builder, an Agent Library, and embeds (Glean in Slack/Teams/Salesforce/Zendesk) let non-engineers build workflow automations, not just ask questions — positioned explicitly against 'traditional search tools that just retrieve documents.'
- Explicit Agent Development Lifecycle (build/govern/measure) and 'active data & AI governance' scanning/remediation across 100+ sources targets large, regulated enterprises worried about AI sprawl and data exposure — a governance layer most lighter competitors and generic chat-over-docs tools don't offer.
Слабкі місця- Pricing is completely opaque — no published price on glean.com; buyers must go through sales, and third-party trackers (Vendr, gosearch.ai, TrustRadius) all note pricing 'is not publicly disclosed,' making it hard for smaller buyers to self-serve or budget.
- High floor: minimum contracts commonly cited at ~100-250 seats and $50,000-$225,000+/year, pricing it out of reach for small/mid-size teams — directly opposite LiveCEO's self-serve, custom-domain onboarding model.
- Index-heavy architecture duplicates a centralized copy of enterprise data (including PII) into Glean's own index, which reviewers flag as expanding the compliance/audit surface for regulated industries — a real IT/legal objection.
Enterprise AI copilot / in-app AI assistant embedded in a productivity suite
Bought by IT/procurement at organizations already standardized on Microsoft 365 (E3/E5/E7 or Business Standard/Premium) who want AI woven directly into Word, Excel, PowerPoint, Outlook, Teams and SharePoint, grounded in each user's own Graph-permissioned mail/files/chats, plus a low-code agent-building layer (Copilot Studio) for IT/ops teams to automate workflows. It is less a destination "ask anything about the company" chat product and more an AI layer stapled onto the existing Office estate — the buyer is typically the same person who already owns the M365 tenant, and the pitch is incremental productivity inside familiar apps rather than a new company-wide knowledge system.
Ціна Copilot Chat (free): $0 · Microsoft 365 Copilot Business (standalone add-on, up to 300 users): $18/user/mo annual · Business Standard + Copilot (bundle): $23.50/user/mo annual
За що платятьWork-data-grounded Copilot Chat (Graph grounding across mail/files/Teams/SharePoint)In-app AI (drafting/summarizing/analysis inside Word, Excel, PowerPoint, Outlook, Teams)Copilot Studio custom/autonomous agentsSharePoint Advanced Management (oversharing prevention, site lifecycle, restricted access/discoverability policies)Advanced Microsoft Purview (automatic sensitivity labeling, DLP extended to Teams/endpoints, Insider Risk Management, Communication Compliance, Adaptive Protection, 10-yr Advanced Audit)
Сильні сторони · слабкі місця
Переваги- Deepest possible in-app integration: AI lives natively inside Word/Excel/PowerPoint/Outlook/Teams/SharePoint rather than requiring a separate destination app.
- Permission-aware by construction — reuses each user's existing Microsoft Graph/SharePoint ACLs automatically, no separate permission model to build (though this also means it inherits any pre-existing oversharing problems).
- Largest prebuilt connector catalog among suite incumbents (100+), with a dual synced/federated architecture that lets sensitive live systems (e.g. Salesforce) be queried without ever indexing/copying their data.
- Only major competitor with a dedicated enterprise-wide AI-agent governance control plane (Agent 365, GA as part of E7 in 2026) for agent identity, Conditional Access, lifecycle, and security posture management as agent sprawl becomes a real enterprise risk.
Слабкі місця- Pricing is deliberately layered and hard to total: the advertised $30/user add-on requires a separate qualifying base license, so real all-in cost is commonly cited as $66-$87/user/month at enterprise tier (or $34-$43/user/month at business tier) — a stark contrast to a single transparent per-seat price.
- Free 'Copilot Chat' is a bait-and-switch relative to what buyers actually want: it explicitly cannot see org data (web-only) until the paid add-on is purchased.
- Accuracy/trust sentiment has been trending negative: third-party tracking (Recon Analytics) shows Copilot's accuracy NPS falling from -3.5 (mid-2025) to -24.1 (Sep 2025), only partially recovering to -19.8 by early 2026 — cited via secondary sources, treat as directional not authoritative.
Enterprise AI search + agentic assistant platform
Gemini Enterprise (rebranded/relaunched from Agentspace in October 2025) is Google Cloud's agentic AI platform for large organizations: a single multimodal assistant that provides permission-aware enterprise search, chat, and generative answers grounded across Google Workspace (Gmail, Drive, Docs, Meet) and 100+ third-party connectors (Slack, Jira, Confluence, SharePoint, ServiceNow, Salesforce, GitHub, Notion, HubSpot, Zendesk, etc.), plus pre-built Google agents (Deep Research, NotebookLM Enterprise, Idea Generation, Data Insights) and a no-code/full-code agent builder (Agent Designer/Agentspace agent gallery). Buyers are IT/security leadership at large enterprises already standardized on Google Workspace/Cloud who want a governed, SSO-integrated 'front door' to company knowledge plus a platform to build and host custom agents, rather than a lightweight chat-over-docs tool. It competes directly with Glean, Microsoft Copilot, and Moveworks; it is weakest for organizations not already on Google Workspace/GCP.
Ціна Business: $21/user/month · Standard: $30/user/month · Plus: From $50/user/month
За що платятьFull third-party connector ecosystem (100+ prebuilt: Slack, Jira, Confluence, SharePoint, ServiceNow, Salesforce, GitHub, Notion, HubSpot, Zendesk, Asana, Box, Dropbox, Monday, Linear, Zoho suite, PagerDuty, etc.)Permission-aware enterprise search and generative answers (results and answers respect each user's underlying source-system ACLs)No-code and full-code custom agent building + Agent Marketplace/gallery + basic agent governancePre-built 'Made by Google' agents: Deep Research, Data Insights (structured/analytical querying), NotebookLM Enterprise create & publish, Idea GenerationEnterprise-grade security & compliance controls (VPC Service Controls, CMEK, SSO via Google identity, HIPAA/compliance support)
Сильні сторони · слабкі місця
Переваги- Deep native integration with Google Workspace and Google Cloud (Gmail, Drive, Docs, Meet, BigQuery, Vertex AI) — turnkey for orgs already on that stack, with SSO built on existing Google identity
- Agent-platform-first architecture: positions itself as an agent platform that happens to include search (vs. Glean, which is search-first and adding agents), so it doubles as a build environment for custom task-executing agents with a marketplace/gallery
- Pre-built 'Made by Google' agents (Deep Research, Data Insights, NotebookLM Enterprise, Idea Generation) give out-of-the-box analytical and research capability without custom development
- Backed by Google's frontier model access (priority Gemini model availability on some tiers) and massive connector library (100+) at enterprise scale
Слабкі місця- Pricing and tiering are confusing and inconsistently documented even by Google itself — the official edition-comparison table shows the higher-priced Plus tier missing features that lower tiers (Business/Standard) have (e.g., priority model access, NotebookLM creation, Code Assist), an unusual tier design that creates buyer confusion
- Real-world cost is unpredictable: seat fee is separate from consumption-based Agent Compute/Storage billing, and reviewers report effective all-in cost 'gets more expensive the more you rely on it,' with custom/high-volume agent usage adding significant unbudgeted spend
- Primarily optimized for organizations already standardized on Google Workspace/GCP; third-party system bridging (Salesforce, Jira, custom apps) is reported as shallower than dedicated enterprise-search players like Glean
Enterprise AI search / permission-aware RAG knowledge assistant
Amazon Q Business is AWS's fully-managed, permission-aware conversational assistant that indexes an enterprise's internal content (SharePoint, S3, Salesforce, Confluence, Jira, Slack, ServiceNow, Google Drive, etc.) and answers questions grounded in that content, respecting each source system's native ACLs. It's bought primarily by enterprises already standardized on AWS (using IAM Identity Center for SSO) who want a "co-pilot for company knowledge" without managing their own RAG/vector infra, plus admins who want Q Apps (no-code mini-app generator) and tie-ins to QuickSight BI. CRITICAL CONTEXT: AWS is sunsetting Q Business for new customers — the product closes to new sign-ups on July 31, 2026, existing customers keep bug/security fixes only (no new features), and AWS is actively pushing everyone to migrate to a successor product called "Amazon Quick" (an AI workspace assistant unifying Slack/Teams/Outlook/CRM/DB/docs with agentic Flows, Automate, Research and Spaces). This makes Q Business effectively a legacy/sunset product as of today (2026-07-16), which is itself a major competitive data point.
Ціна Lite (user subscription): $3/user/month · Pro (user subscription): $20/user/month · Starter Index (capacity, separate from user subscription): $0.140/hour per index unit
За що платятьQ Apps (no-code app generator + API access)Custom & built-in action plugins (e.g. Jira, Salesforce, ServiceNow, meeting/PTO lookups) and third-party natural-language actionsQuickSight Reader Pro / Q in QuickSight data insightsLonger, richer answers + image processing + Slack/Outlook/Word/Teams surface integrationEnterprise Index (multi-AZ)
Сильні сторони · слабкі місця
Переваги- Deep native AWS trust/compliance fabric: PrivateLink VPC endpoints, FIPS-compliant endpoints, CloudTrail audit logging, IAM Identity Center-based SSO — appeals to regulated/AWS-committed enterprises out of the box.
- ACL-aware indexing that crawls and enforces each source system's native permissions at query time (not just at ingest), so answers are provably restricted to what the asking user could already see.
- Q Apps: turns a chat thread into a shareable no-code mini-app/workflow, a lightweight app-builder layer most competitors don't bundle.
- Tight coupling to QuickSight for structured/BI data Q&A alongside document Q&A.
Слабкі місця- Being sunset: closed to new customers after July 31, 2026; existing customers get only bug/security fixes, no new features — AWS itself is pushing migration to a different product (Amazon Quick), a major red flag for any prospect evaluating it today.
- Migration path to the successor product has real gaps: Q Apps, guardrails, and Actions are explicitly excluded from the "bring your own index" migration path and must be manually rebuilt; non-IDC identity setups lose per-user/per-group granular access and require custom scripting to rebuild.
- Historically weak at parsing tables and images in source documents, and at handling complex/multi-step queries, per user reviews (partially addressed only from Dec 2024 onward).
AI-augmented workspace / wiki with enterprise knowledge search
Notion sells to teams that already live in Notion as their docs/wiki/PM tool (startups through large enterprises) and is upselling them into AI: 'Notion Agent', 'AI Meeting Notes', and 'Enterprise Search' that reaches into Slack, Drive, GitHub, Jira, Teams, SharePoint and other connected apps. The buyer is typically an existing Notion workspace admin or IT/security team evaluating whether to consolidate several point-solutions (search, chatbot, meeting notes, writing assistant — pitched by Notion at ~$150/user/month combined) into one Notion Business/Enterprise seat. It competes more as 'replace your wiki + add AI' than as a dedicated cross-company answer engine; its knowledge graph is still Notion-page-centric with connectors bolted on for reach into a handful of other tools, versus a platform purpose-built to auto-build a company-wide wiki from 8+ heterogeneous sources.
Ціна Free: $0/member/mo · Plus: $10/member/mo billed annually · Business: $20/member/mo billed annually
За що платятьFull/production Notion AI ("Notion Agent") + AI Meeting NotesEnterprise Search (cross-app Q&A over Slack, Drive, GitHub, Jira, Teams, SharePoint, OneDrive, PDFs)Third-party AI connectors (Slack, Google Drive, GitHub, Jira, Teams, SharePoint/OneDrive; Linear/Gmail/Salesforce/Zendesk/Box 'coming soon')SAML SSOSCIM provisioning, audit log, advanced security controls, SIEM/DLP integration, EU/US data residency, zero data retention with LLM providers
Сильні сторони · слабкі місця
Переваги- Single tool that is simultaneously the wiki/docs authoring surface AND the AI search layer — no separate 'ingest a company wiki' step for content natively authored in Notion, since Enterprise Search treats first-party Notion pages as first-class alongside connected apps.
- Aggressive bundling/pricing pitch: positions Business plan AI+search as replacing ~$150/user/mo of point solutions (dedicated search tool + chatbot + meeting transcription + writing assistant), an explicit consolidation argument.
- Source citations on every Enterprise Search answer, with stated permission-awareness that filters results to what the querying user can access in each connected app.
- Built-in agent framework (Custom Agents, launched Feb 2026) for workflow automation beyond Q&A, metered by usage credits rather than bundled flatly.
Слабкі місця- Notion AI answer accuracy is inconsistent per third-party testing (~80% for single-source lookups but dropping to ~60% for questions requiring synthesis across multiple pages), with reported hallucinations referencing non-existent content.
- Connector breadth is narrow relative to a purpose-built 'company brain' (roughly 7-8 sources live, several still 'coming soon' — e.g., Gmail and Linear not yet live, and no native Figma, Miro, or Confluence connector mentioned in vendor material).
- Slack connector only indexes public channels, missing private-channel context that often holds the most decision-relevant discussion.
Enterprise AI knowledge management / AI-powered knowledge base + enterprise AI search
Guru positions itself as "The Governed Knowledge Layer for Enterprise AI" — a platform that structures, verifies, and continuously maintains a company's internal knowledge so that both humans and AI tools (ChatGPT, Claude, Copilot, internal agents) get trustworthy, cited, permission-aware answers. Buyers are mid-market to enterprise IT/knowledge-ops, support/CS, and RevOps teams who already have sprawling knowledge across Confluence, Slack, Salesforce, Zendesk, Drive, etc. and need a governance/verification layer plus an AI search/chat surface on top, delivered via a Chrome extension, Slack app, Teams app, web app, and MCP server for third-party AI tools. It is less a "wiki" and more a knowledge-quality/governance product wrapped around AI answers — the SME verification workflow (assign an owner, cadence, auto-archive stale content) is Guru's signature differentiator versus a generic wiki-chat.
Ціна Free / legacy Starter: $0 for small teams; older third-party trackers cite ~$10/user/mo for a bottom paid tier, but this no longer appears on Guru's own current pricing page · Self-Serve (current primary paid plan): $25/seat/month billed annually, or $30/seat/month billed monthly · Enterprise / Expert: Custom, usage-based
За що платятьKnowledge Agents (AI chat, AI research agent, MCP server for third-party AI tools)SME Verification Workflow (assign verifiers, set review cadence, Slack/email reminders, one-click re-verify, auto-archive stale cards, Internal Trust Score)SSO / SCIM + enterprise identity providers (Okta, Azure AD, OneLogin, Ping Identity, Delinea)Compliance & governance package: SOC 2 Type II, HIPAA-ready, GxP support, DLP masking, centralized audit logs, encryption at rest/in transit100+ native data connectors (Salesforce, Zendesk, Confluence, Jira, Google Drive, Dropbox, HubSpot, Freshdesk, plus 20+ HRIS platforms) + Zapier/Workato/Prismatic + custom API
Сильні сторони · слабкі місця
Переваги- SME verification workflow with owner assignment, review cadence, reminders, and a visible 'Verified' badge / Internal Trust Score — a purpose-built governance layer most generic wiki-chat tools lack
- Delivers AI answers inside the tools people already work in (Slack, Chrome, Teams, ChatGPT, Claude, Cursor) via native apps and an MCP server, rather than forcing users into a separate destination
- Automated knowledge-quality maintenance: conflict detection, duplicate merging, auto-archiving of unverified/unused content, claimed to lift 'verified' coverage from ~60% to 100%
- Strong enterprise compliance posture (SOC 2 Type II, HIPAA-ready, GxP, DLP masking) marketed explicitly at regulated/large orgs
Слабкі місця- No published Enterprise pricing — sales-led, usage-based, and opaque, making it hard for buyers to budget or self-serve past 10-ish seats
- Hard seat-based cost floor: every reader (not just editors) needs a paid seat, and Self-Serve has a 10-seat minimum (~$250-300/mo entry), which is expensive for small teams and scales linearly rather than by usage on the lower tier
- Guru's flagship AI capability (Knowledge Agents/chat/MCP) is entirely walled off from Self-Serve customers — smaller teams get only a 'base layer' of AI
Enterprise AI teamwork/knowledge assistant
Rovo is Atlassian's generative-AI layer (Search, Chat, Agents, and Rovo Dev for coding) built on top of the "Teamwork Graph" — Atlassian's permission-aware knowledge graph spanning 150B+ objects across Jira, Confluence, JSM, and 40-90+ connected third-party SaaS tools. It is bought by IT/platform admins and engineering leaders who are already Atlassian Cloud customers (Jira/Confluence/JSM/Teamwork Collection on Standard, Premium or Enterprise) and want AI answers, research, and task automation grounded in their existing work-management data, without adopting a separate standalone "company brain" product. Atlassian's core sales motion since April 2025 has been to bundle Rovo into existing Cloud subscriptions (no separate SKU) rather than sell it standalone, explicitly because customers struggled to build a business case for a separately-priced AI add-on.
Ціна Core Rovo (Search, Chat, Agents) — bundled into Jira / Confluence / JSM on Standard: $0 additional · Core Rovo — Premium tier: $0 additional · Core Rovo — Enterprise tier: $0 additional
За що платятьHigher monthly Rovo credit allowance (Chat/Agents/Deep Research usage)Rovo Dev (autonomous coding agent / CLI)Rovo Studio custom agent governance (approvals, versioning, audit logs, restrict agent creation to selected groups/admins-only)Synced (admin-configured) connectors to third-party SaaS for full indexed searchAudit logging of Rovo admin/user actions and MCP server connection controls (allow-listing tools/domains, OAuth 2.1 policy)
Сильні сторони · слабкі місця
Переваги- Teamwork Graph: a 150B+ object/relationship knowledge graph purpose-built from years of Atlassian work-management data (issues, projects, docs, code, design assets), which Atlassian argues gives Rovo deeper organizational context than a generic RAG-over-documents wiki-chat.
- Deep native integration with Jira/Confluence/JSM workflows — Agents can read and act on live work-tracking objects (issues, tickets, boards), not just answer questions about static documents.
- Rovo Dev: a dedicated coding agent/CLI (separate from chat/search) for PR review and code generation tied directly into the dev toolchain (Bitbucket, GitHub, GitLab, Azure DevOps).
- Zero-friction adoption for existing Atlassian customers — Rovo is auto-enabled on paid Cloud sites with no new purchase decision, which lowers the bar to trial versus buying a standalone AI-brain product.
Слабкі місця- Pricing model is still unstable/in flux — Atlassian reversed from a standalone $20-24/user/month SKU (Oct 2024) to bundling for free into subscriptions (Apr 2025) within six months, and overage billing for credits above the included pool is 'not currently' charged but explicitly a future possibility with only 90 days' notice promised — hard for buyers to model long-term cost.
- Locked to the Atlassian ecosystem: to get real value (and to even evaluate it, since Confluence Free excludes Rovo) you must already be a paying Jira/Confluence/JSM customer, unlike a standalone cross-tool assistant.
- Credit-metered usage (10 credits/Chat or Agent request, 100 credits/Deep Research) creates a hard usage ceiling per plan tier that can throttle heavy users mid-month.
Enterprise AI knowledge assistant / workplace search
Dashworks positioned itself as a Slack-first (also web/browser-extension) AI assistant that answers workplace questions by querying connected apps (Slack, Google Drive, Confluence, Jira, Salesforce, etc.) live via API rather than building a persistent index, then synthesizing an LLM answer with citations while respecting each source's native permissions. Buyers were mid-market to enterprise IT/ops and RevOps teams wanting fast, low-setup Q&A over scattered tools without a full data-lake/ETL project. In April 2025 HubSpot acquired Dashworks to fold its search/reasoning tech into Breeze (Copilot/Agents); third-party sources report the standalone product stopped new signups and shut down as an independent product around July 16, 2025, so it is likely no longer being actively sold — the live marketing/pricing site appears to be frozen legacy content.
Ціна Team: $12/seat/mo billed monthly, $10/seat/mo billed annually · Business: $15/seat/mo billed monthly, $12/seat/mo billed annually · Enterprise: custom/undisclosed
За що платятьCustom bots + LLM selection + org-wide integration accounts + AI customizationSSO + SCIM directory syncAPI accessAnalytics/insightsCustom data retention policies
Сильні сторони · слабкі місця
Переваги- Live, non-indexed federated search architecture — queries source apps in real time per question and inherits native ACLs, avoiding a stale/duplicated copy of sensitive data (vendor/third-party claim, not independently verified).
- Broadest out-of-the-box HRIS and design-tool connector coverage among comparable assistants (Workday, BambooHR, Gusto, HiBob, Paylocity, Darwinbox; Figma, Miro, Mural, Canva, Lucidchart) — niche categories many competitors skip.
- Multiple access surfaces from day one: native Slackbot, web app, and browser extension, minimizing change management for end users.
- Low entry price point and no-minimum Team tier ($10-12/seat) made it accessible to small teams, unlike enterprise-search incumbents (e.g. Glean) that typically require larger contracts.
Слабкі місця- Product's future is uncertain/likely discontinued as a standalone offering: acquired by HubSpot in April 2025 and, per secondary sources, stopped new signups and shut down independently around July 16, 2025 — its search/reasoning tech is being absorbed into HubSpot Breeze rather than sold as its own product, so current pricing/feature claims may reflect frozen legacy content rather than an active roadmap.
- G2 comparison data (pre-shutdown) rated Dashworks below Glean on federated search (9.0 vs 9.5) and ease of use (9.2 vs 9.5), suggesting search quality/UX was a relative weak point versus the category leader.
- No self-serve custom domains, alerting/dashboards, analytical SQL lane, or voice interface were surfaced in any source reviewed — its scope is Q&A/search over documents/apps, not the broader operational-intelligence layer (alerts, dashboards, SQL analytics) that LiveCEO offers.
Enterprise AI agent / "AI teammate" platform
Dust is a Paris-based (Series B, ~$40M raised) platform for building and running custom multi-agent "AI teammates" that search, reason over, and take action across a company's connected tools and data — not just answer questions. Buyers are mid-market to enterprise teams (support, sales, ops, engineering) who want no-code custom agents built on top of company data (Slack, Notion, Google Drive, Confluence, GitHub, Salesforce, Zendesk, Snowflake/BigQuery, etc.) rather than a single generic chat. It explicitly markets itself against Glean as "agents that act, not just search," and cites high internal adoption rates (70-90% team usage cited at reference customer Vanta) as proof of stickiness. Admins/IT buy it for governed, permission-aware rollout across the org; individual teams build specialized agents for their own workflows (support triage, CRM updates, etc.).
Ціна Free: $0/user/mo · Pro: $30/user/mo billed monthly, or $24/user/mo billed annually · Max: $150/user/mo billed monthly, or $120/user/mo billed annually
За що платятьNumber of connected data sourcesSCIM automated user/group provisioningAudit logs & custom data retentionSingle-tenant deploymentUsage volume (credits/month)
Сильні сторони · слабкі місця
Переваги- Agent-first, action-taking framing: agents are pitched as multi-step task executors (draft replies, update CRM records, triage tickets, run workflows) not just a search/answer box — a step beyond typical wiki-chat Q&A.
- No-code custom agent builder available even on lower/free tiers, so individual teams can spin up narrow, domain-specific agents rather than relying on one generic assistant.
- Model-agnostic: 20+ frontier models (GPT, Claude, Gemini, Mistral, DeepSeek) selectable per agent, avoiding single-vendor lock-in.
- Fine-grained permission model via Spaces (open vs restricted data containers) plus Member/Builder/Admin roles, so agents can only see/act on data the invoking user is entitled to.
Слабкі місця- Business-tier connector cap of just 'up to 3' data sources is restrictive for companies wanting broad coverage (Slack+Drive+Jira+Confluence+GitHub+Gmail simultaneously, as LiveCEO offers) — forces an upsell to custom-quoted Enterprise.
- Credits reset monthly with no rollover, and 'Advanced' agent actions cost 3x a basic action — usage-based metering that heavy users report burns through Pro's 8,000 credits quickly, and pricing/packaging changed materially as recently as June 2026 (flat unlimited-message €29/seat plan replaced by this tiered credit system), signaling packaging instability for buyers evaluating long-term cost.
- GitHub connector only ingests issues/discussions/top-level PR comments — no source code or inline PR review comments — a materially thinner GitHub ingestion than a code-aware wiki.
Enterprise AI assistant / agent platform
Sana AI (product name "Sana" / "Sana Agents") is sold to mid-market and enterprise teams (ops, HR, sales, tech companies) who want a single AI layer that finds company knowledge across 100+ connected apps, joins/summarizes meetings, and lets non-technical users build no-code multi-step agents/workflows that both answer questions and take actions in connected systems (CRM updates, RFP drafting, dashboards). It competes for the same buyer as LiveCEO (a company-wide knowledge assistant) but leans harder into agent/workflow automation and, since the Workday acquisition, into HR/finance-adjacent enterprise workflows. Its sister product Sana Learn targets L&D/HR buyers for AI-native LMS/authoring, sold at a minimum of 300 seats.
Ціна Free (Sana / Sana Agents): $0/user/mo · Team (Sana / Sana Agents): $30/user/mo · Enterprise (Sana / Sana Agents): custom/undisclosed
За що платятьUnlimited queries / meeting recordings & >5 workspace membersBroader connector set (Asana, Gmail, Outlook, Zendesk, etc.) and higher per-integration document ceiling (10,000 vs 1,000 docs)Model selection (OpenAI + Claude)SAML SSO, SCIM (user provisioning), domain verificationUnlimited workspace members and unlimited documents per integration
Сильні сторони · слабкі місця
Переваги- Agentic four-part capability model (Find / Act / Build / Automate) — goes beyond Q&A retrieval into taking actions in connected systems (CRM updates, contract drafting, RFP responses) and building live dashboards, not just answering questions.
- No-code multi-step workflow/agent builder aimed at non-technical business users (HR, finance, ops), not just engineers.
- Native meeting capture across Zoom/Meet/Teams with recap generation baked into the core product (not a bolt-on).
- Model-agnostic: lets orgs choose/switch between OpenAI, Claude, and other LLMs rather than a single fixed model.
Слабкі місця- Pricing/branding is fragmented across at least two distinct products (the AI assistant/'Sana Agents' vs. the 'Sana Learn' LMS) with separate pricing pages, which is confusing for buyers comparing offers.
- True enterprise pricing (the tier with SSO/SCIM/custom integrations/single-tenant/SLA) is entirely custom/undisclosed — no published $/user number, unlike the flat $30/user Team tier.
- Sana Learn side of the business shows real product gaps for a modern SaaS product: no native mobile app (browser-only) and weaker gamification vs. established LMS competitors — suggests the company's engineering focus/maturity varies by product line.
Enterprise copilot / agentic AI assistant for IT & HR employee support automation
Moveworks sells to large enterprises (typically 1,000+ employees, often existing ServiceNow shops) as a company-wide AI assistant that sits in Slack/Teams/ServiceNow/web and both answers employee questions (enterprise search across IT/HR/finance/facilities content) and autonomously resolves tickets/requests by calling back-end systems (Workday, ServiceNow, Salesforce, SharePoint, etc.). Buyers are IT/HR/Employee-Experience leaders trying to cut ticket volume and deflect Tier-1 support. ServiceNow acquired Moveworks for $2.85B (announced Mar 2025, completed Dec 2025); by Feb 2026 ServiceNow folded its conversational-AI/enterprise-search layer into a new offering called "ServiceNow EmployeeWorks" (part of a broader "Autonomous Workforce" push), while continuing to sell Moveworks as a standalone product too. This materially changes the competitive picture: it's no longer an independent point solution but the front-door AI layer of the ServiceNow platform, with ~350+ enterprise customers and ~5M+ employee users, and roughly 250 customers already overlapping with ServiceNow at acquisition time.
Ціна Only tier: custom enterprise contract (no self-serve, no published list price): custom/undisclosed; third-party data pegs it around $15-$45 per employee per year · ServiceNow EmployeeWorks (post-acquisition bundle, GA ~early 2026): custom/undisclosed
За що платятьCompany-wide licensing on full employee headcount (not just power users)Agent Studio / Creator Studio (low/pro-code IDE with 1,000+ pre-built agent templates, Agent Architect guidance)AI Agent Marketplace (pre-built solution packs for IT, HR, finance, facilities; MCP support for connecting agents)Permission-aware enterprise search ('Airlock' real-time permission validation + ReBAC normalization across 100+ content systems, ACL-mirrored File Search)100+ prebuilt connectors/integrations (ServiceNow, Workday 100+ prebuilt agents, Salesforce, SharePoint, Slack/Teams/Google Chat/Webex, Zendesk, Jira, Gong, Zoom, IAM/HRIS/CRM systems)
Сильні сторони · слабкі місця
Переваги- Agentic 'Reasoning Engine' that plans and autonomously executes multi-step actions across back-end systems (not just answering questions) - cited autonomous resolution rates >88% for distributed workflows, going beyond a generic wiki-chat that only retrieves and answers
- Purpose-built permission infrastructure (Airlock real-time validation, ReBAC normalization, ACL-mirrored file search) engineered specifically for enterprise-search-scale permission enforcement across hundreds of systems
- Deep, pre-built two-way integration depth with major systems of record (100+ prebuilt Workday agents, native ServiceNow ticketing actions) rather than generic read-only connectors
- Now backed by ServiceNow's platform and distribution (post-$2.85B acquisition) - bundled into EmployeeWorks/Autonomous Workforce, giving it access to ServiceNow's action/orchestration layer and its large existing customer base (~250 shared customers pre-acquisition)
Слабкі місця- Zero pricing transparency - no public pricing page, no self-serve tier, every deal requires a custom sales-led quote, which several review sites flag as the top reason buyers seek alternatives
- Priced on total employee headcount, not active usage - companies pay for every employee whether they use it or not, making it expensive and inflexible for SMBs or fast-growing teams
- Heavy implementation burden: 8-16 week typical rollout and $50K-$200K+ professional-services cost on top of license fees before go-live
OpenAI ChatGPT Enterprise
General-purpose AI chat assistant with enterprise "company knowledge" retrieval layer
ChatGPT Enterprise is OpenAI's model-first chat assistant repackaged for large organizations: unlimited/high-usage access to frontier models, an expanding library of 1st- and 3rd-party "connectors" (Slack, Google Drive, SharePoint, GitHub, Confluence/Jira via Atlassian Rovo MCP, HubSpot, Zendesk, Asana, Gmail/Calendar, etc.) surfaced through a feature called "Company Knowledge," plus admin/security controls (SSO, SCIM, IP allowlisting, data residency, EKM, audit/compliance logs) needed to pass procurement. Buyers are IT/security-approved enterprise rollouts (150+ seat minimum, annual contract) who already trust ChatGPT for general productivity and want it wired into company data with governance guardrails, rather than a dedicated cross-source knowledge-retrieval product. The smaller ChatGPT Business/Team tier (self-serve, 2-seat minimum) is the entry point for teams not ready for Enterprise procurement; Company Knowledge itself is exposed on Business, Enterprise, and Edu alike, so much of the connector value is not actually Enterprise-exclusive — Enterprise mainly buys governance/compliance/scale, not more knowledge features.
Ціна Free: $0 · Go: ~$8/user/mo · Plus: $20/mo
За що платятьSCIM auto-provisioning / deprovisioning tied to Okta, Entra ID, Google WorkspaceCustom role-based access control (RBAC) across features (Canvas, Agent Mode, Codex, custom GPT creation, memory, web search)IP allowlistingData residency across 10 global regions + Enterprise Key Management (customer-held cloud KMS keys)Compliance Logs Platform — append-only chat/auth logs exportable to SIEM/eDiscovery
Сильні сторони · слабкі місця
Переваги- Runs on OpenAI's frontier models directly (fastest access to newest GPT releases), which is the main draw for orgs already standardized on ChatGPT for general productivity
- Company Knowledge blends live connector retrieval with the general-purpose assistant (coding, writing, analysis, image/voice, Deep Research, Agent Mode) in one product, rather than requiring a separate specialized search tool
- Broadest brand recognition and largest reported enterprise market share (cited ~35% in one comparison), which lowers change-management friction for adoption
- MCP developer mode gives technically sophisticated customers an extensible, standards-based way to wire in proprietary internal tools with write actions
Слабкі місця- Not purpose-built for cross-source enterprise search: an independent benchmark cited by a competitor (Glean) found human graders preferred Glean's answers ~1.9x as often as ChatGPT's on enterprise-context questions
- No standard citations-by-default across all claims — high-stakes use (legal/medical/financial) still needs human review; hallucination risk persists
- Opaque, quote-only Enterprise pricing (no published list price) with a steep 150-seat / annual-prepaid minimum (~$108K/year floor), unlike self-serve competitors — slows/complicates small and mid-market deals
AI answer engine / enterprise search & knowledge assistant
Perplexity started as a consumer/prosumer AI search engine (cited, real-time web answers) and has layered an enterprise tier on top: Enterprise Pro/Max sell to IT and knowledge-worker teams at mid-size to large companies (named customers include Stripe, Zoom, Databricks, Snowflake, HP, Vercel, Replit) who want a single chat that blends live web search with internal company files/apps, with SSO/SCIM, admin controls, and a "your data is never used for training" guarantee. It is bought by IT/security-conscious buyers as a safer, governed version of the consumer product, and increasingly pitched as an agentic "Computer" that can act (browse, generate docs, query databases) inside Slack/Snowflake/Salesforce, not just answer questions.
Ціна Free: $0 · Pro (individual): $20/user/mo · Max (individual): $200/user/mo
За що платятьInternal Knowledge Search (file upload + org-wide indexing across PDF/Word/PPT/Excel, searched jointly with live web)App Connectors (400+ prebuilt: Google Docs, Notion, Slack native via Carbon acquisition, plus Salesforce, HubSpot, Box, Vercel, Jira, GitHub, Linear, etc.)SSO/SAML 2.0 + SCIM provisioning (Okta, Azure AD, Google Workspace, OneLogin), MFA, short-lived session credentials, domain-restricted signupSecurity Hub (centralized admin command center), role-based permissions, model enable/disable toggles, developer-API toggle, org-wide Space visibility controlsSOC 2 Type II, GDPR/HIPAA compliance posture, 'data never used to train models', EU data-center availability, HIPAA BAA (Max only), expanded/exportable audit logs
Сильні сторони · слабкі місця
Переваги- Live web search fused with internal knowledge in a single query — Perplexity's core strength is real-time, cited web answers, which a pure internal-wiki tool does not natively provide.
- Snowflake Data Map gives a warehouse-aware NL-to-SQL analytical lane out of the box, directly comparable to LiveCEO's SQL lane, but tied specifically to Snowflake rather than warehouse-agnostic.
- Broad prebuilt connector catalog (400+) via Carbon acquisition plus open MCP extensibility, so IT can plug in tools without waiting on Perplexity's roadmap.
- Agentic 'Computer' layer (browser agent, document creation/editing, multi-model routing across 20+ models) pushes beyond Q&A into task execution.
Слабкі місця- No auto-built company wiki: Perplexity relies on manual/ad-hoc file uploads and live connector queries rather than continuously synthesizing a structured, browsable knowledge base the way LiveCEO's auto-built wiki does — it's a search/retrieval layer, not a curated knowledge artifact.
- Citation and accuracy issues reported broadly: reviewers describe source hallucinations, citations that don't actually support the claim made, misattributed quotes, and shallow reasoning on complex multi-step tasks — a real risk for an 'answers any question about the company' use case.
- Team collaboration and enterprise workflow depth criticized as limited compared to purpose-built enterprise-knowledge tools; some reviewers call enterprise pricing expensive relative to capability.
Agentic AI knowledge/research platform for finance and legal
Hebbia sells "Matrix," a multi-agent AI platform built specifically for the rigor of finance and legal work — investment banks, asset managers, PE/hedge funds, law firms, consultancies, and government (e.g., US Air Force). Buyers are enterprise procurement/IT plus line-of-business heads (research, deal teams, legal ops) at large, well-capitalized organizations — 50+ Fortune 500 clients including Goldman Sachs, McKinsey, and Bridgewater, with penetration into ~33% of top global asset managers by AUM. It is not a general company wiki-chat; it's pitched as an "AI associate/analyst" that automates document-heavy workflows (due diligence, M&A analysis, earnings-call review, credit agreement review, investment memo drafting) end-to-end, with an "infinite effective context window" claim (processing full documents rather than excerpts) and heavy emphasis on cited, auditable outputs. Sold exclusively via sales-led enterprise contracts — no self-serve tier, no public rate card.
Ціна Self-serve / free: none · Lite seat (reported/estimated): ~$3,000–$3,500/user/year · Professional seat (reported/estimated): ~$10,000/user/year
За що платятьMatrix multi-document/multi-company agentic analysis (parallel agent orchestration across o1/o3-mini/GPT-4o-class models, 'infinite effective context window')Skills & Agents — automated, proactively-triggered agents that run firm processes continuously and update a self-improving indexDraft — generation of branded, client-ready spreadsheets/slides/reports directly from analysis (incl. FlashDocs acquisition, June 2025)Premium financial/legal data connectors (S&P Capital IQ, FactSet, PitchBook, Preqin, Third Bridge, Snowflake, DealCloud, Salesforce, SharePoint, Box, public filings)API & MCP connector for embedding Matrix's document intelligence into internal tools
Сильні сторони · слабкі місця
Переваги- Purpose-built for finance/legal document rigor rather than general company Q&A: full-document reasoning (not chunked RAG) with an 'infinite effective context window' claim, reported at 92% accuracy vs 68% for standard RAG on complex financial/legal benchmarks (per Hebbia/OpenAI-sourced claims).
- Multi-agent orchestration across multiple frontier models in parallel per query (o1/o3-mini/GPT-4o-class), routing sub-tasks to the best model rather than a single model/prompt.
- Deliverable-generation (Draft/FlashDocs) that outputs client-ready branded slides/spreadsheets/reports, not just chat answers — addresses the 'last mile' most Q&A tools ignore.
- Standing/proactive agents (Skills) that continuously monitor and trigger workflows rather than only answering on-demand queries.
Слабкі місця- Pricing is entirely opaque and sales-gated — no self-serve tier, no published rate card, long (1+ year) minimum contracts; multiple trackers explicitly call this out as a barrier to small teams and transparent budget planning.
- Very high cost floor (reported low-to-mid five figures per seat/year, six-figure-plus typical contracts) puts it out of reach for SMB/mid-market and makes it a poor fit outside finance/legal verticals where the ROI case is hardest to build.
- Narrow vertical focus (finance/legal document analysis) — not designed as a general cross-functional company brain covering Slack/Jira/GitHub/engineering or broad org-wide knowledge the way a horizontal wiki-chat tool would be.
Enterprise AI search + chat / "AI knowledge assistant"
Onyx positions itself as "the application layer for LLMs" — an open-source (MIT-core) internal search + chat assistant that connects to 50+ company data sources (Slack, Google Drive, Confluence, Jira, GitHub, Salesforce, etc.), does permission-aware RAG/agentic retrieval, and can be running in ~30 minutes. Its go-to-market leans on open source as the wedge: engineering-led buyers (platform/IT teams) self-host free to avoid a sales cycle and evaluate before committing; once usage grows, the same org converts to Cloud or buys an Enterprise license for SSO, permission sync, RBAC, analytics and support. Backed by $10M seed (Khosla Ventures, First Round, YC); customers cited include Netflix, Ramp, and Thales Group. Sits squarely in the same "one chat over your whole company's data" category as LiveCEO, but reaches the buyer via a free/open-source self-host motion rather than self-serve SaaS onboarding.
Ціна Community Edition (self-hosted, free): $0 · Business (Cloud, managed): $20/user/month · Enterprise (Cloud or self-hosted with EE license): Custom/undisclosed
За що платятьSSO (OIDC/SAML) + SCIM user/group provisioningPermission sync connectors (inherit source ACLs from Slack/Drive/Confluence/GitHub etc. into search results)RBAC + group-based permissions on connectors, document sets, and agentsUsage analytics & query history/auditWhite-labeling / custom branding
Сильні сторони · слабкі місця
Переваги- Open-source (MIT) core with a real self-host path — lets engineering-led buyers try/deploy for free before any sales conversation, unlike SaaS-only competitors including LiveCEO's self-serve-but-hosted model.
- Broad, source-agnostic connector catalog (50+) with an explicit 'Lite Mode' (<1GB memory) for lightweight/local deployments alongside the full standard stack.
- Agentic/deep-research mode plus code execution, image generation, and 'Artifacts' output — pitched as a general LLM application layer, not just search, and reportedly benchmarks well on deep-research leaderboards.
- MCP and Actions support for connecting arbitrary external tools/APIs into agent workflows, giving more extensibility than a fixed-connector product.
Слабкі місця- Core trust/governance features — permission sync, SSO, RBAC, analytics — are NOT in the free/open-source tier; a security-conscious company-wide rollout effectively requires paying for Cloud or an EE license, undercutting the 'free and open' pitch for the exact features that matter most for broad deployment.
- Self-hosted Enterprise Edition pricing is fully undisclosed/custom — no self-serve enterprise checkout, meaning larger deployments still funnel into a traditional sales/contact-us cycle despite the open-source framing.
- Reviewers note limited polished end-user UI out of the box (more of a toolkit than a finished product), formal support/SLA guarantees are still maturing for an early-stage OSS vendor, and maintaining live connectors requires ongoing operational upkeep.
Ніша: Slack AI · Tettra · Slite · Bloomfire · GoSearch · Stack Internal…
Fragmented knowledge-management-AI niche
8-product survey, one line each. (1) Slack AI: AI layered onto the Slack inbox itself — search/summarize/recap the conversations you're already having; bought by existing Slack Business+/Enterprise+ customers, not a standalone KB tool. (2) Tettra + Kai: lightweight internal wiki for ops/HR/support teams, with Kai as a Slack-native bot that answers from the wiki and mines Slack threads into docs; bought by small-mid teams who find Confluence too heavy. (3) Slite Ask: docs/wiki tool for startups with an AI 'Ask' layer that answers from docs and (on paid tiers) connected apps; bought by fast-moving teams wanting Notion-like docs plus a chat-search box. (4) Bloomfire AI: enterprise knowledge-engagement platform (heavy in insurance, financial services, customer support) with AI-enhanced search/authoring on top of a curated, community-style knowledge base; bought by orgs needing structured, governed, searchable institutional knowledge at scale. (5) Qatalog: AI 'work hub'/enterprise search across ops tools, recently acquired by ClickUp and folded into ClickUp's ecosystem; bought by ops/IT leaders wanting a single search bar across HR/IT/business systems. (6) GoSearch: agentic enterprise search + AI agents platform with 100+ connectors (incl. personal/private connector search) and a genuinely free tier; bought by IT/knowledge leaders as a lower-cost, transparent-pricing alternative to Glean. (7) Capacity: CX/support automation platform with an AI knowledge base at its core, usage-based agent pricing; bought by contact-center and support orgs to power AI agents/chatbots and agent-assist, not general company-wide knowledge chat. (8) Stack Overflow OverflowAI / Stack Internal: internal Q&A knowledge base (rebranded 'Stack Internal') with AI-enhanced search/auto-answer and, at Enterprise, an MCP server to ground coding agents; bought by engineering orgs wanting a Stack-Overflow-style trusted-answer culture internally.
Ціна Slack AI — Free/Pro: included, basic tier · Slack AI — Business+/Enterprise+: Business+ ~$15/user/mo; Enterprise+ custom · Tettra — Scaling: $8/user/mo annual
За що платятьSlackbot 'Advanced AI' + Enterprise Search across connected apps/databasesKai AI bot (Slack-native Q&A + auto-mining Slack threads into wiki answers)Ask AI unlimited queries + cross-app search (Slack, Jira, Drive, GitHub, etc.)Company-wide AI-enhanced knowledge search/authoring at scale, premium support, advanced integrationsUnlimited AI agent workflows, private connectors, advanced LLMs
Сильні сторони · слабкі місця
Переваги- All eight are single-surface add-ons (chat inbox, one wiki, one community board, one helpdesk) rather than a unified cross-source wiki — none ingests Slack+Drive+Jira+Confluence+Figma+Miro+GitHub+Gmail+meetings into one grounded knowledge base the way LiveCEO does.
- Slack AI's only real differentiator is zero-friction distribution — it's already inside a tool everyone has open — but it can only reason over Slack's own content plus, at the top tier, some connected-app search; it is not a company-wide knowledge layer.
- Tettra/Kai and Slite/Ask both win on ease of adoption for small teams (cheap, fast setup, Slack-native) but cap out on scale — no permission-aware answers across a dozen enterprise systems, no analytical/SQL lane, no alerts/dashboards.
- GoSearch is the most credible 'enterprise search' comparator: 100+ connectors, real SSO/SCIM/audit/BYO-LLM at Enterprise, and a genuinely free tier for individual trial — closest architecturally to LiveCEO's ingestion breadth, but it stays a search/agent layer over existing tools rather than an auto-built company wiki, and has no voice, Telegram bot, or self-serve custom-domain onboarding story.
Слабкі місця- Slack AI: locked to Slack's own content plus limited connected search; mandatory all-seats purchase model when it was an add-on; no wiki, no SQL analytics, no cross-tool permission-aware answers.
- Tettra + Kai: small-team scale ceiling (250 users before Professional/Enterprise); AI limited to what's in the Tettra wiki + Slack history, no Drive/Jira/GitHub/meetings ingestion.
- Slite Ask: Basic tier's 30-questions/month cap makes the AI feel like a demo, not a daily tool; full value requires jumping to $20/user/mo Pro.