# Blits.ai Agentic AI — Enterprise platform for goal-driven AI agents and agentic workflows, with approvals, audit trails and governance. > Source: The Agents Index — https://theagentsindex.com/blits-ai-agentic-ai-for-the-enterprise (structured, researched, re-verified) > Facts last verified: 2026-10-22 Blits.ai Agentic AI is the agentic layer of an enterprise AI platform where goal-driven agents plan, call tools in business systems and verify results. Blits.ai's own team builds, integrates and runs each use case in its cloud, your cloud or on premises. | Fact | Value | | --- | --- | | Website | https://blits.ai/products/agentic-ai | | Price | Price on request | | Free tier | No | No free plan and no trial is published. Every plan's price cell reads 'Contact us' and prices are charged annually in advance; the only no-cost step offered is a demo of the use case before committing. ## What it does **Done-for-you service** · Yes — Blits.ai sells a delivered use case: its team workshops, builds, integrates, deploys and then runs the agent for its whole life. > You buy a working use case, not a license to build one. The Blits.ai team takes it from workshop to production and runs it for as long as it is live. > — https://www.blits.ai/products/agentic-ai · Section 'How we deliver it', intro paragraph above the five numbered steps ## How much it does unattended **Runs autonomously** · Partly — Workflow runs start and finish without a person, but tools that move money or change data wait for approval by default. > A workflow starts from a schedule, a webhook, an incoming email, a new file, an API call or a manual run. It then works until the goal is reached, a stop condition applies or the time limit runs out. > — https://www.blits.ai/how-agents-work · Section 'The agent loop', card 'Runs start without a person' **Multi-agent** · Yes — Agents delegate sub tasks to specialist agents and multi agent workflows run on shared context and shared goals. > Run multi agent workflows on shared context and shared goals. > — https://www.blits.ai/products/agentic-ai · Workflow control center panel, 'Orchestrate at scale' card **Agent permissions** · Yes — A per tool execution policy decides automatic, approval-first or forbidden, with read-only tools running alone and money-moving tools gated. > Tools that move money or change data require approval by default, and read only tools run automatically. Each tool can also be set to automatic or disabled. > — https://www.blits.ai/security-and-governance · Section 'Human oversight and guardrails', card 'Approval per tool' ## Which models it runs on **Claude** · Yes — Anthropic Claude is named among the 300-plus connectable models, and the model can be chosen or switched per agent. > Connect to more than 300 models from OpenAI, Azure OpenAI, Anthropic Claude, Google Gemini and Gemma, Meta Llama, Mistral, Microsoft Phi, Alibaba Qwen, DeepSeek, GPT-OSS, Humain, Fanar and more. > — https://www.blits.ai/features · Section 'Models and providers', feature '300+ LLMs' **GPT** · Yes — OpenAI and Azure OpenAI models are both listed as providers you can connect and select per agent. > Connect to more than 300 models from OpenAI, Azure OpenAI, Anthropic Claude, Google Gemini and Gemma, Meta Llama, Mistral, Microsoft Phi, Alibaba Qwen, DeepSeek, GPT-OSS, Humain, Fanar and more. > — https://www.blits.ai/features · Section 'Models and providers', feature '300+ LLMs' **Open models** · Yes — Open-weight families are supported and can run inside the buyer's own environment, including fully air gapped deployments. > For strict residency, choose models that a cloud provider hosts in the same region, or open weight models that run inside your own environment, including fully air gapped. > — https://www.blits.ai/data-residency · Section 'And the AI models?' **Model choice** · Yes — The buyer picks the model per agent and per task and can switch providers without rebuilding the workflow. > Combine Google, OpenAI, Claude, and more on Blits.ai. Switch providers whenever you choose, without rebuilding a single workflow. > — https://www.blits.ai/products/agentic-ai · Section 'Model flexibility', heading 'The best model for every task. No lock in.' ## Where you use it **In your editor** · No — The vendor's complete 190-feature inventory lists fourteen channels covering chat, voice, Teams and APIs, and no IDE or code-editor extension among them. > A complete, searchable overview of the Blits.ai platform. From AI agents and channels to payments, guardrails, and enterprise governance, all in one layer. > — https://www.blits.ai/features · Page intro above 'Showing 190 of 190 features', read against the Channels section **On the command line** · No — No terminal or command-line tool appears in the complete feature inventory; programmatic use is via REST and WebSocket APIs instead. > A complete, searchable overview of the Blits.ai platform. From AI agents and channels to payments, guardrails, and enterprise governance, all in one layer. > — https://www.blits.ai/features · Page intro above 'Showing 190 of 190 features', read against the APIs and extensibility section **In your pipeline** · Partly — Release gates promote an assistant only when a chosen test suite passes, but this runs inside the platform, not a buyer's CI on pull requests. > Promote an assistant to production only when a selected test suite, including latency checks, passes on the new version. > — https://www.blits.ai/features · Section 'Reliability and recovery', feature 'Release gates' **In a browser** · Yes — The platform is a hosted web application at platform.blits.ai with builders, dashboards and a tenant admin console. > No code and pro code builders, so your teams iterate fast > — https://www.blits.ai/products/agentic-ai · Section 'Flexible by design. Built for any agentic workflow.', first bullet ## Whose machine it runs on **Self-hosted** · Yes — The Kubernetes-native platform runs in the buyer's own cloud account or data centre, including air gapped installs done on site. > Run entirely in your own data center for maximum control and security. Built for organizations with strict data residency requirements. > — https://www.blits.ai/deployment-options · Section 'Run on premises', intro paragraph ## What it costs to run **How it meters** · Yes — Plans meter monthly interactions alongside bot and seat counts, with extra interactions and AI engine costs listed as separate lines. > 1k interactions/month 1 bot 1 seat 2 channels 2 AI engines AI engine costs Hosting of bots Self guided onboarding Extra interactions > — https://www.blits.ai/pricing · Section 'Choose your plan', Basic plan feature list **Free tier** · No — All four plans read Contact us, prices are on request and billed annually in advance; the only no-cost route is a sales demo. > Prices are available on request. All prices exclude VAT and are charged annually in advance. > — https://www.blits.ai/pricing · Footnote directly beneath the four plan cards **API access** · Yes — REST and WebSocket APIs start conversations, stream answers and run workflows, with API tokens per agent, bot and workflow. > Start conversations, send messages and stream answers over REST or WebSocket, and run workflows through the API. > — https://www.blits.ai/integration-standards · Section 'Open standards', card 'REST and WebSocket APIs' **MCP server** · Yes — It is both MCP client and MCP server: the platform hosts an API-key protected MCP server per app so other agents can call its tools. > The platform hosts an MCP server per app, protected by an API key, that gives other agents and assistants tools over the data you store in the platform. > — https://www.blits.ai/integration-standards · Section 'Open standards', card 'MCP server' **Bring your own key** · Yes — Buyers set their own key and endpoint per agent for any OpenAI-compatible model, and usage is tagged platform key or customer key. > Set your own API key and endpoint per agent. Usage is tagged as platform key or customer key, so costs stay traceable. > — https://www.blits.ai/integration-standards · Section 'Model providers', card 'Bring your own key' ## Buying it for a team **A company can buy it** · Yes — Every plan is bought through sales on annual invoicing, with multi-tenant hierarchy, subtenants and an admin console for the organisation. > Four plans on one platform. Start with a single bot, scale to unlimited bots, and get pricing built around your volume and requirements. > — https://www.blits.ai/pricing · Page intro beneath the 'Blits.ai pricing' heading **Seat model** · Yes — Seats are capped per plan rather than sold individually: 1 on Basic, 5 on Pro, 15 on Business, unlimited on Enterprise. > 15 seats > — https://www.blits.ai/pricing · Section 'Choose your plan', Business plan feature list, seats line **Pooled budget** · Yes — Each plan carries one monthly interaction allowance for the whole tenant, with extra interactions billed beyond it, so consumption is shared not per seat. > 250k interactions/month > — https://www.blits.ai/pricing · Section 'Choose your plan', Business plan feature list, first line **Admin controls** · Yes — A tool execution policy sets per tool whether an agent may run it, must ask first, or may not use it at all. > Set per tool whether an agent may run it automatically, needs approval first, or may not use it at all. > — https://www.blits.ai/features · Section 'Agentic workflows and tasks', feature 'Tool execution policy' **Audit log** · Yes — User actions and sign ins are logged and exportable as CSV, and every workflow run stores its tool calls and returns. > Changes and sign ins by platform users are logged, and tenant admins can export the audit log as CSV. > — https://www.blits.ai/security-and-governance · Section 'Audit and monitoring', card 'Audit log of user actions' **Single sign-on** · Partly — Microsoft Entra and Google SSO are supported for admin login; the vendor names no SAML or generic OIDC provider support. > Single sign on Microsoft (Azure AD / Entra) and Google SSO. > — https://www.blits.ai/features · Section 'Security and compliance', feature 'Single sign on' ## What happens to your code **Opt out of training** · Yes — The privacy policy rules out training on customer data by default and says model providers are contracted on terms that forbid it too. > Blits.ai does not use customer data or conversation content to train, fine tune or improve any artificial intelligence model, whether our own or that of a third party > — https://www.blits.ai/privacy-policy · Section 5 'Artificial intelligence and conversation content', clause 5.1 **Data residency** · Yes — Six named managed regions each hold a tenant's data in place, and any other country is served by private cloud or on premises. > Blits.ai runs managed production environments in the European Union, Qatar, Singapore, the United Arab Emirates, the United States and Australia. > — https://www.blits.ai/data-residency · Page intro paragraph beneath the heading **Getting out** · Yes — Assistants export as one encrypted package with BPMN flows, and conversations, training data and consent records export as CSV. > An assistant exports as one package with its flows, intents and training data, functions, and the agents, workflows and model settings it uses. Flows are modelled in BPMN, an open process standard. > — https://www.blits.ai/data-portability · Section 'What you can export', card 'Complete assistants' **Certifications** · Yes — Four independent certifications are claimed for the platform and managed service, with GDPR and EU AI Act alignment alongside them. > ISO 27001, ISO 42001, SOC 2 Type II and PCI DSS Level 1. > — https://www.blits.ai/security-and-governance · Section 'Certifications and frameworks', card 'Certified' ## What changed - 2026-10-22 — Capabilities · Free tier · Price · Terms - 2026-10-13 — Capabilities · Free tier · Price · Terms ## Keeping it current These facts are re-checked against Blits.ai Agentic AI's own pages and updated when they change. If you represent Blits.ai Agentic AI, sign in and update the listing from your dashboard at https://theagentsindex.com/dashboard. The operations this site accepts from an agent are listed at https://theagentsindex.com/.well-known/agents.json — that catalog needs no key to read, and each entry states its own auth.