Start nowالعربية
Step three · The intelligent enterprise

Senaiy Lab — the platform for building AI agents

Turn your existing software… into smarter services.

Build your agent on your knowledge and software; the rest is ready: integration, approvals, measurement, audit, in-country hosting. Top up and pay only for what runs — no seats, subscription or minimum.

For your organization, or for your clients if you're an implementation partner.

Start nowGet help with a use case
  • Sign up yourself
  • See the full pricing inside the platform
  • Build your first agent before you pay anything.
  1. Ask
  2. Ground
  3. Govern
  4. Act
  5. Account

“Follow up on invoices more than 30 days overdue.”

Reads from ERPNext via MCP and finds 3 invoices
Tool allowedProcessed inside the KingdomBudget
Three follow-up messages readyApprove — Collections manager
The cost is recorded in the log

Models inside the Kingdom by default · every step logged

Nobody loses in the demo.

Building a working agent is now easy; the hard questions start after it:

  • Who approves before anything reaches the customer?Governance — approvals, permissions and an audit log.
  • Where is the data processed, and how do we prove it to the auditor?Data residency — your data in your region, with the proof logged.
  • What's the spending cap, and who stops it?Cost control — limits you set and consumption you follow.
  • A month later, how do we know what it did and why?Observability — a full history and a clear reason for every step.
  • Who runs it when the model changes?Operations — the work continues even if the model changes.

Not later add-ons, but the foundation Senaiy Lab is built on.

Two paths, one platform.

For organizations

Add agents to your software and workflows, one process first; the decision stays with your team.

  1. 1The agent prepares
  2. 2The employee approves
  3. 3You measure, then expand

For implementation partners

Offer your clients an AI service on the ERP systems you implement — SAP, Oracle, Odoo, ERPNext — without restarting the relationship.

  1. 1Connect each client within what its environment allows
  2. 2Tailor the agents to its workflow
  3. 3Expand from client to client

A separate space for each client.

  • Separate setupKnowledge, integrations, tools and policies per client.
  • Separate usageCost per client, agent and run, and a budget per client.
  • Clearer billingYour fees per client, with usage reviewed before invoicing.

The client relationship stays with you. Partnership: referral · joint implementation · resale and implementation. Every partner gets a trial environment with three agents, half-day training, and our specialist on the first implementation.

Don't start from a blank page.

Small agent templates: open one, connect it to your software, adjust it.

  • “the Collector”

    Follows up overdue invoices with Arabic or English messages, sent after approval.

    Start with it

  • “the Helper”

    Employees ask it in their language inside the ERP; it reads and prepares the action within their permissions.

    Start with it

  • “the Reader”

    Reads contracts, IDs and forms, Arabic too, and moves them into your software.

    Start with it

  • “the Reference”

    Answers employees from policies and guides, and cites the source.

    Start with it

  • “the Matcher”

    Matches purchases to receipts and invoices, and flags gaps before payment.

    Start with it

  • “the Reporter”

    A weekly email to the manager from your software: what changed and what needs a decision.

    Start with it

Every agent suggests and prepares, and acts only with permission.

You decide… not us.

Models stay in the Kingdom by default, global ones only with your permission. Your data trains no model.

  • Knowledge classificationYour knowledge in zones: public · internal · sensitive · restricted.
  • Who sees whatYou set each class's model and processing location; the default is deny.
  • Retrieval logEvery retrieval logged by agent, zone, model and time, exportable for audit.
  • Technical separation, not a promiseA local-only account's key can't reach a cloud model at all.

From request to action… five governed steps.

  1. 1AskStarts from an employee, a schedule or one of your systems.
  2. 2GroundContext from your approved sources, citing what was used.
    Knowledge approval
  3. 3GovernPermissions, policies and processing location are checked before any action.
    Approval when needed
  4. 4Act or ask permissionRoutine runs; sensitive waits for a named approver who sees exactly what will happen, then runs once.
  5. 5AccountThe run and its cost go into a chained log that can't be silently edited, ready for audit.

Every run governed before, traced after, and a kill switch that stops the agents at once when needed.

AI inside the work, not beside it: every action authorized, every cost capped, every step logged.

Add a layer of intelligence to existing software.

  1. 1We examine the software and its integration points
  2. 2If there's MCP, we use it
  3. 3Otherwise API or HTTP, verified and connected via the governed path.
  • and more

Compatibility is confirmed by installed version, integration point and permissions, not by product name alone.

Everything you need to build, connect, control and own.

  • 01

    Build

    Agents and knowledge, without touching code.

    • Agent builder
    • Knowledge engine
    • Graph Workflow
    • Agent-to-agent delegation
    • Text, voice and image
  • 02

    Connect

    Intelligence inside the software where the work happens.

    • MCP and integrations
    • API and SDK
    • A gateway inside your network
    • Your brand (White label)
  • 03

    Control

    Every action authorized, every cost under a cap.

    • Tool limits
    • Human in the loop
    • Consumption control
    • Activity log
    • Input and output screening
    • Kill switch
  • 04

    Own

    Your intelligence close to your data and your users.

    • Digital sovereignty
    • Hosting in your country
    • An agent whose definition you can take with you

Top up. Build. Pay for what runs.

  1. 01Open your accountOn your own, no sales or contract, with a welcome balance.
  2. 02Top up your walletAny amount; balance never expires, and bigger top-ups get a discount.
  3. 03Connect and buildConnect via MCP, upload your documents, and build your first agent.
  4. 04Pay for what runsEvery run metered live, under a cap you set in advance.
Pricing inside the platform

After signing in you see full pricing — inference, knowledge prep, storage and indexing, tool execution — in your currency and country's tax, and start right away.

Sign in to see pricing

Caps: per organization · client · agent · run. At a cap, execution stops and you're alerted — no surprise invoice.

No seats, annual commitment or procurement cycle. Balance, not subscription: if you stop, you lose nothing; your knowledge stays indexed.

Ready agents, or build your own?

Senaiy agents — add them to your package

What it is
Ready agents you add to your package anytime, working inside it
For whom
The process owner: a service or operations manager
How you pay
An add-on to your package subscription
Looking for ready agents to add to your package? See ServeFast

Senaiy Lab — this platform

What it is
A platform for building your own agents on your knowledge and software
For whom
A technical team, an implementation partner, or the product development team with you
How you pay
Only for what runs

No development team? We build with you.

The product development team builds your agents on Senaiy Lab, inside your space.

  • The account is yoursThe space, agents, knowledge and wallet are yours from day one.
  • With your team, not instead of itYour team works beside ours, so the capability stays after we leave.
  • Start with it, then do without itIt builds with you and leaves when your team takes over; the platform stays the same.
Request a team

Find your starting point.

Start with one workflow, or with one client.

If you're an organizationBring a process owner and a workflow you want to improve.
If you're an implementation partnerBring one client's environment and the installed software version.
The sessionA free 30-minute diagnostic where we define the use case, confirm the integration path, and draw the first trial's boundaries.
The trialA limited trial that ends in a clear decision:
  1. Discovery
  2. Integration check
  3. Trial
  4. Scale decision
Start nowGet help with a use case

Media center

LLM: المخ اللي بيفهم النصوص

Frequently asked questions

How do I build an AI agent for my company on Senaiy Lab?

Senaiy Lab is a platform for building AI agents on your knowledge and your software, with everything underneath ready: integration, approvals, measurement and auditing. You open your account yourself, connect through MCP, upload your documents, and build your first agent before paying anything, without touching code. The platform has ready agent templates to start from, and if you don't have a development team, the product development team builds your agents inside your space.

Does Senaiy Lab connect with ERPNext, SAP and Odoo?

Yes, with SAP, Oracle, Odoo, ERPNext and others. We examine the software and its integration points; if MCP is available we use it, otherwise API or HTTP, which we verify and connect through the governed path. Compatibility is confirmed according to the installed version, the integration point and the permissions.

Are the data and models hosted inside Saudi Arabia?

The models run inside the Kingdom by default, and global models only if you allow them. For each classification of your knowledge, you set the permitted model and where it's processed, and the default is to deny. Your data is not used to train any model.

How do approvals and auditing work for what the agent does?

Permissions, policies and the place of processing are checked before any action. Routine actions run; sensitive ones wait for a named approver who sees exactly what will happen, and run only once. Every run and its cost go into a chained log that can't be silently edited and can be exported for audit, and a kill switch stops the agents' activity immediately when needed.

How much does Senaiy Lab cost?

You top up your balance, build, and pay only for what runs: no seats, no subscription, no minimum and no annual commitment. You set spending caps, and when a cap is reached, execution stops and you get an alert, not a surprise invoice. The full pricing is shown inside the platform after you sign in.

Can implementation partners offer AI agents to their clients?

Yes. You offer an AI service to your clients on top of the ERP systems you implement, with a separate space for each client, covering its setup, usage and billing, and the client relationship stays with you. The partnership models are referral, joint implementation, and resale and implementation. Every partner gets a sandbox with three agents, a half-day training, and a SolvFast specialist on the first implementation.