Intras,AI
Book a scan

How we build, govern and run your AI agents We build it with you. Then we run it for you.

One engagement model for every agent: a short scan, a written design you approve, a fixed-fee build, then a monthly run with reporting against a baseline.

Book an Opportunity Scan See pricing
engagement path Every agent
  1. 01
    Opportunity ScanThree workflows, a first business case
    You choose
  2. 02
    Assess and designBaseline measures, written design
    You approve
  3. 03
    Build and hypercareTested with the people who own the work
    You accept
  4. 04
    RunOperated, tuned and reported monthly
    You review
Nothing moves to the next stage without your decision

In short

How Intras AI engagements and pricing work.

Every agent starts with a 30-minute Opportunity Scan, then a written design you approve, a build with hypercare, and a monthly run reported against your baseline.

Pricing is a fixed fee per agent for the build, billed on signature and at written design approval, plus a monthly subscription per agent to run it. Hosting and model costs run in your own accounts with no markup, and you own the source code.

The engagement model

Four stages, each with a clear decision on your side.

Every agent follows the same path, whether it handles payables, hiring or service requests. You see what happens at each stage, what you get from it and what you decide before the next one begins.

Stage 01

Opportunity Scan

A 30-minute executive conversation about where your team's time goes across finance, operations, HR, revenue and IT. We identify three workflows where an agent can return real capacity.

No commitment
What happens
We talk through the work your leaders want off their teams' plates, the systems involved and what it costs you today.
What you get
A short written summary of the three workflows, with a first business case for the strongest one.
Who's involved on your side
An executive who knows where time and money leak, such as a COO, CFO, HR or revenue leader.
Your decision
Whether to take one workflow into assessment and design.
Stage 02

Assess and design

We map the workflow end to end, the systems it touches and who decides what. Then we write down exactly how the agent will work before anything is built.

What happens
Working sessions with the people who own the work. We baseline the measures the agent will be judged on: hours returned, cycle time, risk and leakage, and cost against the alternative.
What you get
A written design covering scope, hosting, model choice, approval rules and the plan and dates for the build.
Who's involved on your side
Your executive sponsor, the process owner, and the people who will approve the agent's actions.
Your decision
You approve the design in writing. Nothing is built until you do.
Stage 03

Build and hypercare

Our engineers build against the approved design, in your environment, and test it with the people who own the work until it behaves the way the design says it should.

What happens
Build, testing with your team on real cases, go-live, then a hypercare period of close watching and adjustment.
What you get
A working agent in your accounts, with source code and documentation in your Git and approval rules in place from day one.
Who's involved on your side
The process owner and the people who do the work today, to test and give feedback, plus access to the systems in scope.
Your decision
Go-live, and acceptance once hypercare shows the agent is working as designed.
Stage 04

Run

We operate the agent so your team doesn't have to. It keeps working, keeps improving, and reports to you every month in terms your leadership already uses.

What happens
Monitoring, tuning, policy and model updates, and new skills added within the approved catalog.
What you get
A monthly report on what the agent did, where people stepped in, and the measures against your baseline.
Who's involved on your side
The process owner and named approvers, who keep approving the decisions that need a person.
Your decision
Each month: keep going, adjust scope, or extend the agent to the next workflow.

What we need from you

A small team on your side, with clear owners.

  • 01An executive sponsor

    Owns the business case and the decision to go ahead.

  • 02A process owner

    Knows how the work really gets done today and what good looks like.

  • 03Named approvers

    A person for each decision point, so the agent always knows who to ask.

  • 04Access to the systems in scope

    Granted through your own controls, limited to what the design calls for.

Pricing

Fixed-fee build. Monthly run. No hourly meter.

You know what an agent costs before work begins, and you never pay for our time by the hour.

Build

One fixed fee per agent

Covers assessment, design, build and hypercare, for one agreed scope.

  • Billed at two milestones: on signature, and at written design approval
  • Scope and fee agreed in writing before work begins
  • Source code and documentation delivered to your Git

Run

Monthly subscription per agent

Covers operation, monitoring, tuning and updates within the approved catalog.

  • Monthly report against your baseline measures
  • Once a measured baseline exists, you can move to usage-based or outcome-based pricing
  • New skills within the approved catalog are part of the run

Hosting and model costs

The agent runs in your own cloud and model accounts. You see those costs directly, and we don't mark them up.

How we size it

The fee for each agent is tied to its business case: hours returned, cycle time, risk and leakage, and cost against the alternative. How we build the business case

Governance and ownership

Agents execute. People approve. Everything is logged.

The rules are written before the agent exists, and they don't change without your approval. What we build belongs to you.

What is a governed AI agent?

  • Decision rights up front. The design names who approves what, before anything is built.
  • Catalog-bounded actions. The agent can only take actions in its approved catalog. Anything new goes through you first.
  • People approve what matters. Changes to access, credentials and money always follow a person's approval.
  • A full audit trail. Every action, approval and exception is logged and reviewable.
  • You own what we build. Source code and documentation live in your Git, and the agent runs in your cloud subscriptions and model accounts.
  • Provider-neutral. We work across the leading model providers and local models, and choose in the design what fits your policies and costs.
  • Your context is your IP. The knowledge your agents build up forms a corporate brain that you control, not us.

Questions leaders ask

The straight answers.

How long until we see value?

It depends on the workflow and the systems involved. The written design sets the plan and dates before you commit to a build, and the monthly report shows the measures against your baseline from the first month of the run.

How much does it cost to build and run an AI agent?

Each agent has a fixed build fee, agreed in writing before work begins, and a monthly subscription to run it. The fee is sized against the agent's business case, and hosting and model costs are billed to your own accounts with no markup.

Do you lock us into one AI model or platform?

No. We work across the leading model providers and local models, and choose in the design what fits your policies and costs. The agent runs in your accounts, so you can change course later.

What if we want to stop?

You can. You own the source code and documentation, and the agent runs in your own cloud and model accounts, so nothing is locked inside our systems.

Can we start small?

Yes. Most engagements start with one workflow. Once it's running and measured, you decide whether to extend it or add the next one.

Who sees our data?

Your data stays in your environment and your model accounts, under your policies. We access only the systems in scope, through the access you grant.

Do you work through partners?

Yes. MSPs and consultants can deliver our agents to their clients under co-branded, white-label or referral models. See the partner program

Start with a scan

Every agent starts with a 30-minute conversation.

Tell us where your team's time goes. We'll come back with three workflows worth automating and a first business case, with no commitment to build.

Request an Opportunity Scan Serving clients? See the partner program