We take AI from first idea to production
DILR.ai is a London AI consulting firm. We use DATS, a five-stage system, to find where AI pays, rank it, decide who owns it, then build and run it with your team. There are three fixed engagements, and most clients start with a 4 to 6 week Placement Diagnostic that ends in a ranked roadmap.
The five stages of DATS, our AI consulting system
- Based inLondon, Companies House 16842656
- Smallest engagement4 to 6 week diagnostic
- Who does the workSenior practitioners only
- What we deliverProduction code, not decks
- Design workOne fixed fee per scope
- Delivery workRetainer plus an outcome fee
Which AI consulting engagement do I need?
Most clients start with the Placement Diagnostic
Pick the sentence closest to where you are today. If AI is not the answer, we will tell you on the first call.
We would suggest
Placement Diagnostic
Start with the diagnostic. It maps your processes, scores every idea and ends with three ranked placements and a list of what not to do.
- Length
- 4 to 6 weeks
- Team
- 2 senior practitioners
- Fee
- Fixed fee, fixed scope
- DATS stages
- 01 and 02
We would suggest
Placement Diagnostic
Bring the list. The diagnostic becomes a pruning exercise, which is usually more useful than writing yet another list.
- Length
- 4 to 6 weeks
- Team
- 2 senior practitioners
- Fee
- Fixed fee, fixed scope
- DATS stages
- 01 and 02
We would suggest
Operating Model
Design the operating model next. Every system gets a named owner, approval gates, evaluation rules and a way to turn it off.
- Length
- 6 to 10 weeks
- Team
- 3 senior practitioners
- Fee
- Fixed fee, fixed scope
- DATS stages
- 03
We would suggest
Execution Office
An Execution Office embeds a senior team with yours, ships the first system, keeps evaluations running and hands it over.
- Length
- 12 months or more
- Team
- 3 to 6 senior people, embedded
- Fee
- Retainer plus an outcome fee
- DATS stages
- 04 and 05
We would suggest
Dilr Voice, set up yourself
You probably do not need consulting. Dilr Voice is an AI receptionist you can set up yourself, from $0.14 a minute plus a monthly fee per phone number.
What does each AI consulting engagement include?
Three engagements with fixed scopes and named deliverables
Each one answers a single question and leaves you with documents or code your team keeps.

- DATS stage 01 and 02
Placement Diagnostic
Where should we use AI, and where should we not?
- Length
- 4 to 6 weeks
- Team
- 2 senior practitioners
- Fee
- Fixed fee, fixed scope
You keep
- Placement map
- Ranked roadmap of three placements
- Feasibility scorecard
- A list of what not to do
- DATS stage 03
Operating Model
Who owns each AI system, and who can change or stop it?
- Length
- 6 to 10 weeks
- Team
- 3 senior practitioners
- Fee
- Fixed fee, fixed scope
You keep
- Governance charter
- RACI matrix
- Lifecycle and evaluation rules
- Review cadence
- Org design
- Policy pack
- DATS stage 04 and 05
Execution Office
How do we get this into production and keep it working?
- Length
- 12 months or more
- Team
- 3 to 6 senior people, embedded
- Fee
- Retainer plus an outcome fee
You keep
- Production code in your repositories
- Evaluation harness
- Runbooks
- An in-house on-call rotation
How long does AI consulting take?
A diagnostic takes 4 to 6 weeks and delivery runs for a year or more
The design work is short on purpose. The long part is running AI in production until your own team can carry it.
- Placement Diagnostic4 to 6 weeks
- Operating Model6 to 10 weeks
- Execution Office12 months or more
How much does AI consulting cost?
Design work has a fixed fee and delivery is paid partly on results
We do not sell hours. The diagnostic and the operating model are each one fixed fee for a published scope. The Execution Office adds an outcome fee that depends on a measure we agree before the build.

Agree
Agree the measure first
Whichever outcome you picked, its baseline and review date are fixed in writing before the build starts. No moving goalposts, in either direction.
Build
Build in the measurement
The placement carries its own measurement, so proving it worked is a query either of us can run, not a slide.
Verify
Bill only after it works
Checked against the agreed baseline. If it did not move, the value invoice does not go out. That risk is ours, deliberately.
What AI systems does DILR build for clients?
The six kinds of AI system we build and run
Each one already runs inside a DILR product, so you can see it working before you hire us.
- RequestCall, form, ticket
- AgentsPlan and hand off
- KnowledgeRetrieve context
- ToolsAct in your systems
- GuardrailsRules, evals, approval
- OutcomeLogged and measured
Autonomous agents
Multi-agent pipelines that plan, hand off between specialists and act on the result.
Runs in Dilr VoiceKnowledge retrieval (RAG)
Your documents chunked, indexed and searched, so answers quote your source instead of guessing.
Voice knowledge basesWorkflow automation
What happens after the conversation: branches, emails, lead records and webhooks.
Voice post-call flowsEvaluation and proof
Work is only done when it is proved: attached evidence, versioned and checked.
Cognibl proof gateGovernance
Ownership, approval points and lifecycle rules designed before the build, not after.
AI operating modelIs DILR the best AI consulting firm in the UK?
We are the right firm when you want working systems rather than a strategy deck
No firm is best for everyone. This is when to hire us, and when a large consultancy is the better buy.
Hire DILR when
- You want the senior people from the first call to be the ones writing the code.
- You want a fixed fee and a fixed scope before any work starts.
- You need one process fixed and measured, not a company-wide transformation programme.
- Your data cannot leave your own infrastructure and the models have to run on your hardware.
- You want to own the code, prompts and evaluations and run them yourself afterwards.
Hire a large consultancy when
- You need hundreds of people working across many countries at the same time.
- Your board or regulator expects a household name on the report.
- The work is mostly organisation redesign or change management beyond AI.
- Your procurement rules only allow suppliers already on an approved framework.
- You need an audit opinion. DILR designs governance but is not an audit firm.
What else do people ask?
Short answers to common questions about AI consulting at DILR
DATS, cost, size, code, confidentiality and how to choose a firm.
- DATS is the five-stage AI consulting system DILR.ai uses on every engagement: Discover and diagnose, Prioritise and place, Operating model, Pilot to production, and Scale and run. Each engagement covers one or more stages, so the work builds on itself instead of starting over. You can start at any stage, and most clients start with the Placement Diagnostic.
- DILR prices by engagement, not by the hour. The Placement Diagnostic and the Operating Model are each one fixed fee for a scope agreed before work starts. The Execution Office is a retainer plus an outcome fee tied to a measure agreed up front. Exact figures are given on a free 30-minute call once the scope is clear.
- The smallest engagement is the 4 to 6 week Placement Diagnostic, run by two senior practitioners. DILR does not sell half-day workshops or keynote briefings, because neither produces a decision your team can build against.
- DILR writes code. Execution Office engagements ship production software inside your stack, including pipelines, evaluation harnesses, integrations and guardrails, and the code is yours. DILR also runs two live products, Dilr Voice and Cognibl, which is where the production experience comes from.
- Yes. The method is the same at any size and only the size of the project changes. If the problem is unanswered phone calls, a small business usually does not need consulting at all: an AI receptionist from Dilr Voice starts at $0.14 a minute plus a monthly fee per phone number.
- Ask five questions of any firm. Who writes the code, and are they the people in the pitch? Is the scope and fee fixed before work starts? How will success be measured, and when is that agreed? Who owns the code, prompts and evaluations afterwards? What happens when the consultants leave? DILR publishes its own answers on our approach page.
- Yes. Every conversation is treated as confidential by default, and DILR will sign your NDA before a scoping call. Send it to contact@dilr.ai.
How do I start AI consulting with DILR?
Book a free 30‑minute AI consultation
Bring the process that costs you the most. We will tell you whether AI can help, which engagement fits and what it would take. If the honest answer is no, we will say so.
