Skip to content
Intura
enid
Enterprise Small AI Models: Custom Training and Integration / Custom SaaS
Build

Software built around the models, for the people who will use it.

Some problems need more than a model behind a service; they need a screen. A review queue for the cases a model flags, a dashboard over what it decided, a browser extension that checks a message before someone replies, a workflow app that connects a model to the systems your team already uses. We design and build that product on top of Intura's models, deploy it, and either hand it to your team or keep running it for you.

The problem

The problem it solves

A model's decision is only useful once it reaches the person or the system that acts on it. Off-the-shelf tools rarely fit how your team works, internal IT has a queue, and a general-purpose AI product sends your data somewhere you did not choose. The result is a good model that nobody uses. We build the missing piece: the software that puts the model's output in front of the right people, in the shape their work already has.

Who it is for

Who it is for

  • Operations and risk teams who need a review queue or a dashboard over model decisions

  • Support and sales teams who work in a browser and want a check where they type

  • Companies that want an internal AI tool without sending their data to a third-party product

  • Teams on a workflow tool such as n8n who need a small app around it

What we do

What we do

  1. 1

    Discovery

    We sit with the people who will use it and map the job: what they look at, what they decide, what they need from the model and what they do next.

  2. 2

    Design

    Screens and flows are drafted and reviewed with those users before code is written, so the first build is the right one more often.

  3. 3

    Build

    The product is built on Intura's models, running on a CPU on infrastructure you choose, with sign-in, roles and the connections it needs into your systems.

  4. 4

    Deploy

    Released to your environment, or to infrastructure we operate for you, with monitoring and backups set up from the start.

  5. 5

    Hand over or run

    Your team receives the code and the runbook, or we keep running and improving it under a support agreement. You can start with us running it and take it over later.

What you get

What you take delivery of

  • A working product, deployed where you chose
  • The source code, in a repository you own
  • Design files, and documentation for users and administrators
  • A runbook for deploying, updating and recovering it
  • Either a handover to your team or an agreement for us to run it

How it runs

How an engagement runs

A typical build starts small: one workflow, used by one team, in production early, then extended from what that team asks for. The proposal after the first call states the first release and the timeline for your case.

  1. 01

    First call

    What the tool is for, and who will use it.

  2. 02

    Scoping

    A written proposal: the first release, what it connects to, where it runs and who operates it afterwards.

  3. 03

    Build in releases

    Working software early and often, reviewed by the people who will use it.

  4. 04

    Handover or operation

    Code and runbook to your team, or a support agreement with us.

FAQ

Common questions

Who owns the code?

You do. The source is delivered to a repository you own, whether or not we keep running the product for you.

Can it run inside our network?

Yes. The models run on a CPU, so the product can be deployed inside your VPC or on-premise like any other internal application, with no call to an outside AI provider.

Do you build browser extensions and n8n workflows?

Yes. Both are common shapes for this work: an extension where your team already types, and a workflow that connects a model to the systems your team already uses.

What happens after launch?

Either your team takes it over with the code and the runbook, or we keep running it under a support agreement with terms written for how you use it.

The models it is built on

Each catalogue model does one job and runs on a CPU. A custom product can use them as they are, or a custom model trained for it.

Browse the models

The other Enterprise services

Enquire

Tell us what you need

Describe the problem, the team and where it has to run. An engineer reads every enquiry and replies by email.

You will get an engineer on the first call, not a qualification script.