Public beta · Platform live

Create a self-learning
AI portal for your company.
Live today. Tailored tomorrow.

MAEQ gets to know how your company works, from its goals and playbooks to its numbers, and puts that understanding behind every answer, view and automation. You choose the model and set the rules. MAEQ brings everything around them.

Join the beta See how it fits together
Any LLMManaged agentsEnterprise context layerHuman in the loop
Central agents
deployed by your organisation
Personal agents
built by your people
Human in the loop
a person approves the steps that matter
Runs you can trust

Both kinds of autonomous agents pass through the same approval gate. Nothing acts on its own without a person who can see it, pause it, or wave it through.

Agentic control tower

Your numbers and your knowledge, in one view.

Your teams stop hunting through dashboards and folders to explain a number. Every department steers on two things at once: the governed numbers in your data warehouse and the context in your documents and applications. MAEQ brings both into your portal as an agentic control tower, shaped by the data you connect, whether that is commercial, supply chain, finance or marketing. Performance views refresh on schedule straight from Databricks or Snowflake. Ask why a number moved and agents combine the warehouse figure with what your documents and business applications say, with the calculation shown. Ask for a chart and it is drawn right in the conversation.

✓

A control tower for every department: commercial, supply chain, finance or marketing views, each built from the data you connect.

✓

Performance views from Databricks Genie and Snowflake Cortex Analyst, refreshed on a schedule and one click away for everyone in the portal.

✓

The warehouse computes every figure. MAEQ never re-calculates your numbers, and shows the SQL behind each answer.

✓

Numbers, documents and applications in one answer, so what happened arrives together with why.

✓

Charts on request in the conversation, and any approved query can become a shared view your team opens every morning.

Weekly commercial review refreshed 06:00 daily
CommercialSupply chainFinanceMarketing
Net revenue
€8.6M
▲ 3.1% vs plan
Gross margin
41.8%
▼ 0.9 pts
Price realisation
97.2%
▼ discounting
Why did margin drop, and chart it by channel?
Margin fell mostly in Wholesale, where two promotions overlapped in week 37. Your pricing policy caps stacked promotions at 15%, and both were approved before that rule took effect.
Gross margin by channel, last 4 weeks
Retail44.1%
Online43.0%
Wholesale36.4%
■ computed in Snowflake · context from 2 SharePoint documents
SQL How this was computed: shown with every answer
The Wire

Your whole company, as a live map.

Sources to portals on one map: every knowledge base, agent and orchestrator you configured, and how a question finds its way through them. MAEQ flags what is broken, like an idle source or an agent no portal can reach, before your users notice. This is the picture you get on day one, and it stays true as you grow.

maeq.com/dashboard

Every line is something you configured. Nothing here is a diagram we drew for you.

Runs on your stack

Bring your own everything.

MAEQ connects to the models, data and tools you already use, organised by what they do. Hover to pause and look closer.

Language models
OpenAILLM provider
Anthropic ClaudeLLM provider
Google GeminiLLM provider
MistralEU-hosted LLM
Azure OpenAILLM on your tenant
Azure AI FoundryModels on your tenant
Structured data
Databricks GenieGoverned data warehouse
Snowflake Cortex AnalystGoverned data warehouse
Connected sources
SharePointDocument sync
Google DriveDocument sync
DropboxDocument sync
Amazon S3 bucketsRead in place
Azure Blob containersRead in place
Tools via MCP
1,000+ appsThrough Composio
SlackDirect connector
Microsoft TeamsDirect connector
JiraDirect connector
GitHubDirect connector
Azure DevOpsDirect connector
GmailDirect connector
Your MCP serversCustom tools
AI clients
ClaudeUses your context via MCP
CursorUses your context via MCP
Coding workspaces
E2BAgent sandbox
DaytonaAgent sandbox
BlaxelAgent sandbox
Document reading
Azure Document IntelligenceOCR and layout
Mistral OCREU-hosted OCR
Vector storage
MAEQ Vector CloudEU-hosted, managed
QdrantVector database
ChromaVector database
pgvectorPostgres vectors
Object storage
Amazon S3Object storage
MinIOS3 compatible
Cloudflare R2S3 compatible
Azure BlobObject storage
Guardrails
Lakera GuardPrompt safety
AWS Bedrock GuardrailsContent and PII guardrails
Observability
LangfuseTracing export
Identity & SSO
Microsoft Entra IDSingle sign-on
OktaSingle sign-on
Any OIDC providerSingle sign-on

Four layers. One platform.

Everything below works so the layer above can be trusted. Your context grounds your agents, your agents power your portal, and governance holds around all of it.

The surfacePortals
Branded AI portals, performance views and autonomous automations. The part your company actually uses.
The intelligenceManaged agents
Agents that route, remember and act in your tools, versioned and run for you, with human oversight built in.
The foundationContext layer
A living understanding of your company: its goals, playbooks, processes and language, learned from your documents, data and conversations, and kept current.
The controlGovernance
Policies, residency, cost tracking and your own model endpoints, enforced across every agent and portal.

You saw the surface at the top of this page. Here is what holds it up, built from the ground up.

The control

Set the rules once. They hold everywhere.

AI adoption usually stalls on three questions: is our data safe, are the rules followed, and what will it cost? MAEQ answers all three before anyone has to ask. Rules are set once and hold on every agent, portal and tool call. Models run on endpoints you approve and never train on your data: MAEQ calls them through their business APIs, not the public chat apps. Every call is logged with its steps and its cost, so security, legal and finance can say yes with confidence.

✓

Policies set once, enforced everywhere: PII redaction, a prompt injection shield, quotas, rate limits and IP allowlists, with no per project rework.

✓

Your data is never used to train models, and MAEQ trains nothing on your conversations.

✓

Bring your own LLM endpoints and keys, so you keep ownership of cost and infrastructure.

✓

Cost and traces for every call, with the agent behind each one and full trace export to Langfuse.

✓

Roles and single sign-on, with admin and portal sign-in through Microsoft Entra ID, Okta or any OIDC provider.

Guardrails applies to all agents
Prompt injection shieldblocks untrusted instructions in retrieved content
Rate limit120 requests per user, per hour
On
Data residencyEU only, model calls and storage
EU
The context layer

A living understanding of how your company works.

General AI knows the world. MAEQ gets to know your company. From the day you sign up it reads who you are, then learns from everything you connect: your goals and priorities, your playbooks and processes, the words your teams use and the standards they hold themselves to. Every conversation teaches it something, and every change at the source keeps it current. So answers don't just cite the right document. They sound like your company and follow your way of working.

✓

Knows your company from day one. MAEQ reads your website and annual report at sign-up, so it starts out knowing who you are and what you do.

✓

Brings order to the mess. Drop in a folder or an archive, or point it at your SharePoint, and MAEQ removes duplicates, groups documents by topic and proposes tidy knowledge bases for you to approve.

✓

Learns your playbooks. Share a framework or playbook once and MAEQ learns its structure, so agents can check new work against it and point out what is missing.

✓

Remembers what your teams teach it. Policies, processes and terminology that come up in conversation become shared knowledge, and you can see and edit all of it.

✓

Stays true as you change. SharePoint, Google Drive, Dropbox, Amazon S3 and Azure Blob sync on schedule and follow every edit and deletion, duplicates merge into one golden record you can restore, and agents refresh what they know overnight.

✓

Grounded, and you can check it. Every answer shows its sources, and built-in evaluation measures whether answers find the right documents and stay true to them.

✓

Your company's context, wherever people work. Claude and Cursor can draw on it too, through keys that see nothing until you grant access.

✓

Beyond documents Beta ground answers on specialist formats too, including DICOM radiology, whole-slide pathology, and NIfTI scans, with your own imaging models in the loop.

✓

Full control over residency and access: storage in the EU, the US or Asia, or in buckets you own.

What MAEQ has learned kept current
Company profileread from your website and annual report
Learned
Sales playbook12 sections, used to check new proposals
Active
From conversations“Invoices above €10,000 need a second approver”
Remembered
SharePointFinance and Operations sites, only changes fetched
Synced
Managed agents

Agents that learn how you actually work.

Your people ask one question and get the answer your best expert would give. MAEQ sends each question to the agents who know the subject, combines what they find, and remembers what matters to your business and to each person. Agents do the follow-through too, in the tools your teams already use, with a person approving anything that changes something. And you never maintain the plumbing: agents are versioned, refresh what they know as your knowledge changes, and when model providers change their models, MAEQ spots it and proposes the move.

✓

The right expert for every question. An orchestrator routes each question to the specialist agents that know the subject and combines their answers.

✓

Gets to know each person: their role, language and how they like answers, all visible and editable.

✓

Does the work, not just the talking. 1,000+ apps through MCP, from Slack, Jira and Microsoft Teams to GitHub and Gmail, with approval before a tool changes anything.

✓

Hands back finished work, not just text: agents write and run code and deliver Word, Excel, PowerPoint, PDF and ten more formats.

✓

Grows with your company. Drop in your documents and MAEQ suggests the agents they call for.

✓

Human in the loop on every decision that should not run unsupervised.

Waiting for you human in the loop
New agent proposedRates Specialist, from 128 documents you dropped in
Review
Tool action proposedcreate a Jira ticket for the delayed supplier order
Approve
Automation readyWeekly supplier risk digest, drafted from your description
Draft
Portals

A portal your company actually uses.

Give every employee one place to get help that already knows your company: your brand, your sign-in, the model you trust, and answers grounded in how you work. Run it on OpenAI, Anthropic, a model on your own Azure tenant, or several at once. Inside it your people chat with the agents your organisation deploys and the agents they build for themselves, and those agents can also run on a schedule on their own. Interactive when someone wants to ask. Autonomous when work should simply happen. A person approves the steps that matter, every time.

✓

A branded portal for your company, live in days, not quarters.

✓

Central and personal agents side by side, the ones your organisation rolls out and the ones your people create.

✓

Autonomous runs with human approval, so work happens on a schedule and a person signs off before anything acts.

✓

Answers people can check, with sources and reasoning shown, and conversations shareable by link.

✓

One place for every department and region, under a single set of controls.

My automations running
Monday revenue digestevery Monday, 8am
Scheduled
Send regional summary4 team leads
Awaiting approval
The other way to ship it

Or drop the assistant onto the site you already have.

A portal is one home for your AI, but you do not always need a whole portal. Add the same assistant to any page you already run, your marketing site, your help centre, your product, with a single snippet. Same agents, same data, the same governance and human-in-the-loop approvals as the portal, meeting people right where they already are.

✓

One snippet, any page. Paste it into the site you run today, no rebuild and no new domain.

✓

The same brains as your portal. Your agents, your data, and your rules, in a corner bubble.

✓

Wears your brand. Colours, position, and greeting match the page it lives on.

Built for organisations that run AI on their own terms.

Enterprises

One governed home for AI across departments, regions, and providers, with every team on the model and rules you approve.

Public sector

Sovereign, auditable AI with full policy control and data residency, so you can adopt AI without giving up custody of your data.

Consultancies and SIs

One multi-tenant platform for every client. Each client gets its own isolated tenant, with its own branding and data, delivered by you.

SMEs and scale-ups

A company AI that is live in days, grows new agents as you grow, and never locks you to one vendor.

Questions, answered plainly.

What is MAEQ?

MAEQ is a platform for managed AI agents. It gives your company a branded AI portal on the model you choose, an enterprise knowledge and context layer built from your documents and your warehouse data, and governance you set once. Agents, knowledge and policies are configured in MAEQ, not coded.

How does MAEQ learn how our company works?

From the start it reads your website and annual report, then everything you connect: documents, playbooks and warehouse data. It learns the structure of your frameworks so agents can check work against them, and it remembers the policies, processes and terminology that come up in conversations. You can see and edit everything it has learned, and it stays current as your sources change.

Which AI models can we use?

OpenAI, Anthropic Claude, Google Gemini, Mistral (EU-hosted), and models on your own Azure OpenAI or Azure AI Foundry tenant, side by side if you like. Bring your own keys, or use access through MAEQ billed by usage.

Where is our data stored?

In the region you choose: MAEQ-managed storage in the EU (Frankfurt), the US or Asia, or your own Amazon S3 or Azure Blob storage. Vectors default to MAEQ’s EU-hosted vector cloud, and buckets can be read where they sit without copying them.

Is our data used to train AI models?

No. MAEQ calls models through their business APIs, not the public chat apps, and trains nothing on your conversations or documents.

Can Claude or Cursor use our company knowledge?

Yes. MAEQ is also an MCP server. With a scoped key, Claude Desktop, Cursor and other MCP clients can search the knowledge bases and ask the portals you allow, and nothing else. Access is closed until you grant it and can be revoked at any time.

Which data sources and tools connect?

SharePoint, Google Drive, Dropbox, Amazon S3 and Azure Blob for documents; Databricks Genie and Snowflake Cortex Analyst for governed performance data; and more than 1,000 apps through MCP, including Slack, Jira, Microsoft Teams, GitHub and Gmail.

How is MAEQ different from LangGraph or other agent frameworks?

Frameworks give developers building blocks to code an agent. MAEQ is the finished platform around the agents: portals, curated knowledge, governance, approvals and cost tracking are included and configured, not built. A side-by-side comparison with LangGraph, Copilot Studio, Bedrock AgentCore and others is at maeq.com/compare.

What does it cost to start?

MAEQ is in public beta. You can sign up for free, every plan lets you bring your own model keys, and a guided proof of concept on your own data is free of charge.

You will not be doing this alone.

Our team helps you connect your first data source, shape your agents, and stand up your portal, so you are live fast and confident. Want to see it on your own data before you commit? Start with a proof of concept at no charge.

Book a free proof of concept Sign up and explore the beta yourself Guided, or self-serve. No cost either way.

Your company's self-learning AI portal.
Live today. Tailored tomorrow.

Start with a branded portal on the model you choose. From there, MAEQ keeps learning how your company works, and every answer, view and automation becomes more useful for it. Our team helps you get there.

Join the beta Talk to us
MAEQ The platform that gives every company its own AI, without giving up control. © 2026 MAEQ
Compare Contact Terms of Use Privacy Policy
All third-party product names, logos, and brands shown above are the property of their respective owners and are used here only to identify the services MAEQ can integrate with. Their use does not imply any affiliation with, sponsorship by, or endorsement from those owners.