Independent AI consulting & implementation, now taking on new projectsStart a conversation
How it works

Six stages, each with an output and a decision

You never commit to more than the next stage. Each one ends with something concrete in your hands and an honest recommendation about whether to continue.

At a glance

The six stages

  1. 01

    Assess

    Map where time goes and which tasks suit AI.

    Ranked shortlist
  2. 02

    Design

    Agree the workflow, controls and success measures.

    Written design & quote
  3. 03

    Build

    Build in short increments, connected to your tools.

    Working system in test
  4. 04

    Test on your real cases

    Run your own examples and fix what fails.

    Test results
  5. 05

    Deploy

    Roll out carefully, often with every output reviewed at first.

    Live system
  6. 06

    Hand over

    Train your team and hand over everything they need.

    Docs, runbook, access list
01 / 06

Assess

Understand what your business is trying to achieve, how the work is done today, and which tasks are realistic candidates for AI.

What happens

  • Conversations with the people who do the work
  • Review of tools, data and documents
  • Each task rated for value, feasibility and risk

You provide

  • Time with a few key people
  • Examples of the work (anonymized if needed)
  • A view of your systems

You get

  • A ranked shortlist
  • A recommended first project
  • Known risks and open questions
Opportunity shortlist
TaskValueEffort
Invoice intakeHighLow
Ticket triageHighMedium
Contract reviewMediumHigh
Board reportingLowMedium
02 / 06

Design

Agree exactly what will be built, how it fits the way your team works, and how you'll know whether it's working.

What happens

  • Map the workflow step by step
  • Decide where a person approves
  • Choose tools and models to fit your rules
  • Collect real examples to test against

You provide

  • Feedback on the proposed workflow
  • Agreement on success measures
  • Access to test environments

You get

  • A written design and scope
  • An agreed test set
  • A written quote: scope, price and timeline
Design document
  • Workflow, step by step
  • Where a person approves
  • Systems and access needed
  • Model and provider choice
  • Success measures
  • Signed off by you
03 / 06

Build

Build in short increments, showing working progress regularly rather than disappearing until the end.

What happens

  • Connect to your systems, starting in test environments
  • Regular check-ins to review progress

You provide

  • Quick answers to questions as they come up
  • Feedback from the people who'll use it

You get

  • A working system in a test environment
  • Visible progress at every check-in

Timeline set in the written quote

Build log
Connected to inbox (test)
Field extraction working
PO matching added
Demo to finance team
04 / 06

Test on your real cases

Run the agreed set of real examples, including the awkward ones, and fix what fails before anything is used for real work.

What happens

  • Run the full test set
  • Fix failures and re-run
  • Check behavior on unusual and adversarial inputs

You provide

  • Your experts' judgment on borderline outputs

You get

  • Test results against the agreed criteria
  • A go / no-go recommendation
Test run ยท your examples
CaseExpectedResult
Standard invoiceMatch POpass
Scanned copyMatch POpass
Wrong supplierFlagpass
Two in one PDFSplitfix
05 / 06

Deploy

Put the system into real use carefully, usually with one team first and every output reviewed until you're confident in it.

What happens

  • Staged rollout
  • Close monitoring after launch
  • Adjustments based on real use

You provide

  • An internal owner for the system
  • Feedback from users

You get

  • A live system in your accounts
  • A clear view of how it's performing
Rollout plan
  1. Pilot with one team, every output reviewed
  2. Review results together
  3. Widen use, approvals stay on
  4. Adjust what is reviewed, if you choose
06 / 06

Hand over

Make sure your team can run, troubleshoot and change the system without depending on any one person.

What happens

  • Training for users and the system owner
  • Personal access removed

You provide

  • The people who will look after it

You get

  • The handover pack
  • Optional monthly support
Handover pack
  • How it works (plain language)
  • Runbook: operate & troubleshoot
  • List of every credential used
  • Test set to re-run after changes
  • Training for users and owner
Choosing a first project

What makes a good first AI project

A good first project builds confidence and shows your organization how AI systems behave, without betting anything important on it.

  1. It happens often

    Frequent tasks give enough volume for the time saved to matter, and enough examples to test with.

  2. It's mostly reading, sorting or drafting

    These are the things current AI models handle best.

  3. Mistakes can be caught

    A person can check the output before it affects a customer, a payment or a record.

  4. The data is reachable

    The information needed lives in systems that can be connected to safely.

  5. Someone owns it

    A person on your side cares about the outcome and will use the result.

Measuring results

Success is defined before anything is built

Measures are agreed in the design stage and compared with how the work is done today, so results can be judged honestly.

Quality

Is the output right?

How often the output is correct or acceptable without edits, on the test set and then in real use.

Time

Is it faster?

How long the task takes end to end, and how much human time it needs, compared with today.

Adoption

Is it used?

Whether the people it was built for actually use it, and what they would change.

Next step

Have a process you suspect AI could help with?

Start with a conversation about how the work is done today. You'll leave with an honest view of whether AI is worth pursuing for it.