# AI strategy, build and infrastructure

Canonical: [AI strategy, build and infrastructure](https://humanfacetech.com/services)

## Strategy

Find the useful opportunity before choosing the technology.

We start by tracing how work happens in your organisation: where information moves, where it stalls, which steps depend on someone knowing the answer. That gives us a short list of use cases worth testing, and what each one needs before a model is involved.

### Business & workflow analysis

We map the process as it runs today and mark where it waits or repeats. You get a ranked list with the information each idea needs.

### AI workshops & training

Your people bring their own tasks. We show where models hold up and where they slip, so your team can judge a proposal.

### Use-case exploration

Each candidate comes with its constraints: the data, who reviews the output, what a first test looks like. We sort what can run now.

## Build

Turn an idea into something your team can try, test, and improve.

We build an early version your team can try. We test it together, check where it succeeds or struggles, and use that feedback to decide what to develop next.

### Prototypes & product development

Each prototype tests one assumption, so trying it gives an answer: does the model handle this input, does retrieval find the right document.

### Knowledge & search systems

We build search around your documents, source access, and retrieval checks. Your team can trace answers back to the material they came from.

### Workflow automation

We connect automation to tools your team already uses, with clear review and recovery steps so people can inspect and correct failed actions.

## Infrastructure

Connect the models, data, and systems behind the experience.

The hosting decision follows your constraints. We compare running models hosted or on your own hardware against where the data can sit, the latency your use case tolerates, the running cost at your volume, and who looks after it.

### Model deployment & integration

We compare hosted APIs and self-hosted models on data handling, operational effort, and running costs, then choose a setup your team can maintain.

### AI gateways & usage governance

A gateway sits between your app and the providers, so you can see which model handled which request and track spend per team.

### Architecture & operations

Before production we agree what happens when a provider is slow or down. Handover is a short doc on how it is built and monitored.

## Working with your team

We work alongside your team throughout the project.

Sessions are remote by default. When a workshop or a review is better in person, we travel to you, and travel costs are agreed case by case.

## How we work

### Assess

Understand the people, the workflow, and the constraints. Agree on a problem worth solving.

### Build

Make a focused prototype. Test it with the people who will use it, then develop what proves useful.

### Operate

Plan for integration, visibility, and the improvements that come after launch.

## Start a conversation

For a faster route, [Book a 30-minute session with us](https://calendar.google.com/calendar/u/0/appointments/schedules/AcZssZ3MvmJCrtltllV0bv1h7BMvEUHyLWWS76p5viktzbOT5Jhs-dvfPRvTpFJ3Nt6X8CqHCM1f1q_M). Otherwise use the [contact form](https://humanfacetech.com/contact).