AI Readiness Audit
Find where AI could make work better. Test the possibilities with your team.
30 minutes with James · bring the problem as it is
- Who builds the prototypes
- A build engineer pairs with your engineers to develop and test the two or three agreed prototypes in your tenancy. The fee is based on five days of partner input, twenty days from a senior operator and nine engineer days. Published day rates are £1,560 for James Rooney and £1,250 for each senior practitioner; the £44,000 is a fixed fee for the agreed deliverables.
- What you keep
- The audit ends with working prototypes, a board readout and a costed plan. Code lives in your repositories from day one. Any further work is a separate decision and agreement.
- Oversight
- James Rooney provides partner oversight on every engagement, and leads the audits personally. Every person on your engagement is senior, with no pyramid of juniors behind them.
On this page
The AI Readiness Audit gives you four weeks to explore where AI could help across your product, internal processes, operations and software delivery. We work with your team to test two or three promising workflows as working prototypes against your real data, inside your own tenancy. The fee is £44,000 fixed, excluding VAT.
Bring the work your people know best, including the exceptions that make it interesting. Their examples and feedback shape what we test. You leave with the prototypes and a costed plan in three-month increments, capped at twelve months, ready for your board to decide what to fund. The code lives in your repositories; you own the deliverables on payment, whether or not you continue with us.
Our specialty-insurance case study follows discovery in month one and a two-week proof of concept in month two, showing how understanding the workflow shaped the build. See how discovery informed an insurance build→
Bring a workflow you would like to improve. On the call, we can explore what would be useful to test.
Talk it throughWhat you get
In scope for the £44k fixed price.
- Two or three working prototypes, built with your engineers against real data in your tenancy, with measured accuracy, an assessment of confidence in the results and an estimated run cost for each
- A heat map of opportunities across product, internal processes, operations and software delivery, ranked by expected return and feasibility
- A record of data, platform, risk and regulatory constraints, including what would need to change
- A view of how AI could improve software delivery, using your current delivery data as the baseline
- A practical assessment of the skills, time and support your people need to use the proposed workflows
- Changes to roles, decisions, incentives and working practices that would help adoption
- A costed plan in three-month increments, capped at twelve months, with dependencies and review points
- An investment case your finance team can examine, with assumptions and confidence attached to the figures
- Prototype code and written findings you own on payment of the applicable fees, with production work scoped separately
Tell us on the call which decisions these findings need to support. We will agree the audit scope before work starts.
Talk it throughThe evidence behind the plan
Seven sections connect what we learn with the decisions ahead. Your team can follow each recommendation back to the workflow, test or assumption behind it.
- 01
Heat map by domain and workflow
A view of the candidates examined across product, internal processes, operations and software delivery. Each is ranked by expected return and feasibility, with the supporting evidence, so you can compare where to invest.
- 02
Constraint register
The data, platform, risk and regulatory limits affecting each candidate. It distinguishes limits to work within from changes you could make, with costs where they can be estimated.
- 03
Decision inventory
Which decisions an agent could take on and which stay with a person. Each recommendation considers the consequences of a wrong decision and how readily it could be reversed.
- 04
People and adoption
What colleagues find useful, where they need more support and how roles, incentives or working practices could help. Their experience of the prototypes informs the recommendations for each function.
- 05
A costed plan
A sequence in three-month increments, capped at twelve months, with build and estimated running costs, dependencies and points to review what you have learned.
- 06
The investment case
Expected value in currency, with the assumptions and confidence behind each figure. Your finance team can use its own numbers to explore how the case changes.
- 07
What to leave out
Candidates to defer or stop, and the reason for each. This gives your team a record to return to if the data, costs or business priorities change.
You receive the written readout and a session with your leadership to see the prototypes, examine the findings and discuss the next step.
Bring the decision your board needs to make. We can discuss on the call what evidence would help.
Talk it throughHow the engagement runs, week by week
Four weeks from the first conversations to prototypes, findings and a costed plan.
- 01Week 1
Explore the work together. Your team walks us through real tasks across product, internal processes, operations and software delivery: the routine steps, the handoffs and the cases that need judgement. We review delivery data, map AI tools already in use and agree the access and risk constraints. Those conversations refine the questions each agreed prototype needs to answer.
- 02Weeks 2–3
Build something your team can try. Our engineer pairs with yours on two or three workflows, agreed in writing before the audit starts, using real data in your tenancy. Your colleagues bring typical and difficult cases to the tests. We measure accuracy, assess confidence in the results and estimate run cost per unit of work. Their feedback helps us understand what works, what needs attention and what to investigate next.
- 03Week 4
Turn the findings into a decision. We explore which roles and decisions would change, what governance is needed and how people would adopt the workflow. We then present the prototypes and a costed plan to your leadership, with the assumptions, risks and evidence available to examine. Your board can decide what to fund, what to revisit and where to stop.
- Week 1
Explore the work together. Your team walks us through real tasks across product, internal processes, operations and software delivery: the routine steps, the handoffs and the cases that need judgement. We review delivery data, map AI tools already in use and agree the access and risk constraints. Those conversations refine the questions each agreed prototype needs to answer.
- Weeks 2–3
Build something your team can try. Our engineer pairs with yours on two or three workflows, agreed in writing before the audit starts, using real data in your tenancy. Your colleagues bring typical and difficult cases to the tests. We measure accuracy, assess confidence in the results and estimate run cost per unit of work. Their feedback helps us understand what works, what needs attention and what to investigate next.
- Week 4
Turn the findings into a decision. We explore which roles and decisions would change, what governance is needed and how people would adopt the workflow. We then present the prototypes and a costed plan to your leadership, with the assumptions, risks and evidence available to examine. Your board can decide what to fund, what to revisit and where to stop.
Have a date in mind? We can discuss availability and what your team would need to prepare on the call.
Talk it throughWould an audit help?
Start with the decision you need to make. These are some of the situations an audit can help clarify.
A useful starting point for
- Boards deciding which AI opportunities deserve investment
- Engineering teams exploring how AI could improve software delivery
- Product teams choosing where AI could help their customers
- Teams trying to understand a stalled Copilot or Gemini rollout
- Businesses mapping informal AI use and the work it supports
- Organisations reviewing overlapping tools bought by different functions
- Leadership teams bringing different views of AI readiness into one plan
- Buyers seeking operational AI due diligence before an acquisition or investment
- Sponsors testing the scope and assumptions of a proposed programme
Consider another route for
- Teams with a current assessment and one well-defined workflow to test: consider a proof of concept
- Teams choosing licences against an already agreed specification: a procurement exercise may fit better
- Organisations seeking formal legal, regulatory or certification assurance: use the relevant specialist
- Sponsors whose investment decision is settled and who cannot act on different findings
You do not need to have the scope worked out. Tell us where you are on the call and we will help you find a starting point.
Talk it throughPrefer to talk it through? Ask us on a discovery call →
You can also bring your questions to a 30-minute call with James.
Talk it throughWhich domains and workflows to fund first, with the evidence behind the order.
£44k·4 weeks·Fixed price
- The first three priorities, their expected return and the confidence in each estimate
- Build and running-cost estimates in a plan of three-month increments, capped at twelve months
- The decisions people retain, the proposed changes to their work and the support they need
- The constraints to resolve before further automation, with dependencies and review points
- What to defer or stop, with a reason for each recommendation
Tell us on the call what your board needs to decide, and we can explore how the audit would support it.
Talk it throughFind the support that fits
If you already know what to build, or need help designing or leading the work, explore these engagements.
Agentic Proof of Concept
Pick the workflow. Two to four weeks later, look at a working thing.
Fixed price · £20k–£55k2–4 weeks03Agentic Design Team
A pair who design the AI-native operating model, and the agentic systems to run it at scale.
£35k–£55k / month2–4 months04Agentic Build Team
A team of three who build and ship it, under partner oversight.
£70k–£85k / month6–12 months05Programme & Delivery Management
We will govern the programme whether or not we are building any of it.
£18k–£35k / monthProgramme durationWe can help you compare these options on the call, including whether your team has what it needs to take the next step itself.
Talk it throughBring the workflow everyone works around.
A 30-minute discovery call with James Rooney. Tell us what you would like to improve, even if the idea is still taking shape. We will explore whether an audit would help and outline a possible scope for you to consider, with no obligation.
Calendar not loading? Open it on cal.com or email hello@tenhaw.com.
Can't see the calendar? Open it in a new tab.
Useful groundwork for a Head of AI
If you are appointing a Head of AI, the audit can help define the work they will lead. If they are already in post, we can work with them to test priorities and shape the investment case.
Support for a Head of AI
Relevant roles: Head of AI, AI strategy lead, Chief AI Officer
The four weeks establish which opportunities deserve attention, what your data and platform can support, and what your people need to make progress. A new AI leader can use that evidence to shape their priorities, team and budget. The audit can also help you write the role specification while you recruit a permanent leader.
The audit is a four-week engagement with agreed deliverables and a fixed fee. You work with senior practitioners, screened to BS7858 standard before client access and contracted to written confidentiality and data-handling terms. They are accountable to James Rooney as well as to you. There is no introduction fee or permanent-placement conversion clause. Either party can end a fixed-price engagement on written notice, with fees due for work performed and committed costs incurred up to termination.
How the associate pool is selected and screened is set out on our team page, and the day rates behind every figure here are on the rate card.
If you are planning an AI leadership hire, we can discuss on the call how the audit would help define the role.
Talk it throughAI Readiness Audit: your questions
What is an AI readiness audit?
An AI readiness audit helps an organisation decide where AI could create useful, measurable value and what would need to change to achieve it. It examines the work, the available data, the risks and the support people need.
Tenhaw's audit covers product, internal processes, operations and software delivery. Over four weeks, we build two or three working prototypes with your team and turn the findings into a sequenced, costed plan. The fixed fee is £44,000, excluding VAT.
What does the audit actually look at?
Four connected areas: the product your customers use, the internal processes that run the business, operations and software delivery. We follow real workflows with the people who know them, review the data and examine the tools already in use.
That might mean exploring how a customer finds an answer, how a team gathers information for a decision, or how engineers test a change. We compare opportunities within and across the four areas so your board can decide where to invest first.
How much does an AI readiness assessment cost in the UK?
Tenhaw's AI Readiness Audit costs £44,000 fixed for an agreed four-week scope, excluding VAT. We agree the business areas, access and two or three prototype workflows before work starts. Organisation size shapes where we focus the four weeks; it does not create a different price band.
Pre-agreed expenses are charged at cost, and model and platform use sits on your own accounts. The fee covers the agreed deliverables. If you request a scope change, we agree it and any fee or timeline changes in writing before the additional work starts. When comparing assessments, check the deliverables, engineering time and client involvement as well as the fee.
Should we start with the audit or a proof of concept?
Choose the audit when you need to decide where to invest across the organisation. It explores four areas, tests two or three candidates and gives your board a costed plan.
A proof of concept fits when you already have one workflow in mind and want to find out whether the approach works. It costs £20,000–£55,000 fixed over two to four weeks, excluding VAT. Your team can use either engagement to learn before committing to a production build.
Who from Tenhaw actually does the work?
James Rooney leads every audit personally, supported by a senior operator and a build engineer. The operator explores workflows with your teams; the engineer pairs with your engineers to build and test the prototypes.
The fee is based on five partner days, twenty operator days and nine engineer days across the four weeks. Associates come through work James knows at first hand or referrals from people he has delivered with. Everyone completes BS7858-standard screening before client access, with written confidentiality and data-handling terms. Your Statement of Work names the people and their responsibilities.
Should we hire a Head of AI or run an audit first?
It depends on the decision you need to make. If you need permanent leadership, the hire matters. An audit can help define that role by establishing the opportunities, constraints and investment priorities they would inherit.
If your Head of AI is already in post, we can work with them over the four weeks to test their priorities and develop the plan with your teams. The audit provides evidence and delivery support while responsibility for the permanent role remains with your organisation.
What is included in the £44,000 audit price?
The agreed four weeks of work: two or three working prototypes, a heat map across the four domains, a constraint register, a decision inventory, an assessment of people and adoption, and an investment case. You also receive estimated running costs and a plan in three-month increments, capped at twelve months, presented to your leadership.
Code is built in your repositories, and you own the deliverables on payment of the applicable fees. VAT, pre-agreed expenses at cost and model or platform use on your accounts are separate. The fee is fixed for the agreed scope; client-requested changes are agreed in writing before additional work begins.
Can the audit sort out the AI tools we have already bought?
It can help you decide what to keep, improve, consolidate or retire. We map the tools people use, including informal use, to the workflows they support. Your colleagues can explain why they chose them, what helps and where they still have to work around a limitation.
We examine overlap and running costs alongside the opportunities for new work. The plan records the recommended changes and their reasons. Tenhaw does not sell software licences.
What does an AI readiness audit deliver in four weeks?
Two or three working prototypes and the evidence to decide what to do next. Each prototype runs against your real data in your tenancy, with measured accuracy, an assessment of confidence in the results and an estimated run cost per unit of work.
The written readout covers opportunities, constraints, decisions, people and adoption, the investment case and candidates to leave out. A costed plan sets out the next steps in three-month increments, capped at twelve months. Preparing a prototype for production is a separate scope.
How do you assess whether a data estate is ready for agentic AI?
We trace the data a real workflow needs: where it lives, who can access it, how complete it is and whether an answer can be checked against its source. Your teams help us find the exceptions and dependencies.
The prototypes then test those conditions in practice. We examine retrieval quality, missing context, permissions and links back to the source, and record what would need to improve before relying on the workflow. Findings apply to the data and workflows examined, with wider questions recorded for follow-up.
How do you work out where AI will not deliver value?
We compare the possible benefit with the data, risk, effort and running costs involved. Some candidates have too little volume to justify a build. Others need better information, a different process or a decision that should remain with a person.
Your team's examples help us understand which constraint matters. We record candidates to defer or stop with a reason, so you can revisit them if circumstances change. A simpler process improvement may be the useful outcome.
Which stakeholders need to be involved in an AI readiness audit?
The people doing the work and engineers who can pair on the prototypes are central. Their examples, judgement and feedback help us choose and test the candidates. We also need an executive sponsor empowered to decide within the agreed scope, the platform or data owner, and someone from risk to agree the classification and controls before prototyping.
Finance helps examine the cost baseline and investment assumptions. We agree the people, access and time needed during scoping so each team can plan its contribution.
How do you baseline current operating costs so AI savings can be measured later?
We record volume, elapsed time, people involved and rework for the workflow as it runs today, then work with your team to express that baseline in currency. Finance can check the assumptions and the value assigned to people's time.
Prototype results and estimated model and platform costs help us explore potential improvements against that baseline. Time released is capacity; it becomes a cash saving only if expenditure changes. The audit estimates the opportunity, while realised savings need to be measured after adoption in live work.
What security constraints are examined before an AI build starts?
We establish where the data may go, who may access it and which actions the prototype is allowed to take. That includes identity and access controls, scoped permissions, approved locations for model inference, data-handling requirements and the evidence your risk team needs.
The prototypes run in your tenancy on your contracts, within the controls agreed with your platform and risk owners. Those colleagues help us identify what can be tested and what needs another decision before work proceeds.
Can an audit tell us whether our engineering team can build AI-native?
It can show how your engineers work with the approach on the selected workflows, and where further support would help. We review how software is delivered today and pair with your engineers while building the prototypes.
That gives us practical evidence about requirements, implementation, testing and review. Their feedback helps identify useful changes to tools and working practices. Four weeks provides a view of the work examined, which you can use to plan further learning and delivery.
How does this compare with an AI strategy engagement from a large consultancy?
Compare the scope you need with the people and deliverables in each proposal. Tenhaw's audit is a four-week engagement at £44,000 excluding VAT, with James Rooney, a senior operator and a build engineer contributing an agreed allocation of time. It includes prototypes built with your engineers and a costed investment plan.
A larger firm may suit a broader international programme or a need for several specialist disciplines. Ask each supplier what it will test, what your team receives, who does the work and what involvement it needs from you. Our pricing page provides the rate card and published framework comparisons for context.
How do you make sure an audit ends in engineering work rather than slides?
Prototyping is part of the agreed scope, with an engineer allocated to build it. During weeks two and three, our engineer pairs with yours on two or three workflows using your real data and repositories.
The tests and your team's feedback inform the plan presented in week four. You can inspect the code and results as they develop, and you own the deliverables on payment even if the audit is the only engagement you buy.
What is the return on £44,000 spent before committing to a build?
The audit helps you make a better investment decision. It can identify a promising workflow, expose a costly dependency or show why a proposed build should stop. The value of that decision depends on the work and spend under consideration; we do not promise a financial return from the audit.
You receive a cost baseline, prototype measurements, running-cost estimates and a plan with assumptions your finance team can examine. Together, those give you a way to judge the next investment before making it.
Can the audit be used as AI due diligence before an acquisition?
Yes, as an operational assessment scoped to the target or business unit, at the same £44,000 fixed fee over four weeks, excluding VAT. We examine the work, systems, evidence of value and dependencies within the agreed scope.
Access to the people doing the work, engineers and relevant data is needed for the prototypes. We check whether that involvement is practical within your transaction timetable before you commit. Legal, financial and regulatory due diligence remain separate specialist work.
Is an agent readiness audit the same as an AI readiness audit?
Tenhaw's service was originally called an Agent Readiness Audit. It now covers AI opportunities more broadly across product, internal processes, operations and software delivery. An agent is one possible approach to a workflow.
The agent-specific work remains in the decision inventory: what a system could act on, what a person should decide, and how consequences and reversibility shape that choice. The engagement runs four weeks at £44,000 fixed, excluding VAT, and tests two or three candidates as working prototypes.
Can you audit a programme our current AI supplier is already building?
Yes. We can examine the workflows, assumptions and proposed investment alongside your team and supplier. We agree access, responsibilities and scheduling before starting, including how to avoid disrupting ongoing work.
The audit tests two or three candidate workflows and compares the findings with expected value, feasibility and running costs. Your board receives a costed plan and recommendations about what to continue, change or stop. Tenhaw does not sell licences, and you can take the findings forward with your existing supplier.
What happens in the weeks after the audit readout?
The engagement ends after the agreed four weeks. Nothing renews automatically. You have the prototype code in your repositories, the written findings and a costed plan with review points. You own the deliverables on payment of the applicable fees.
Your team can take the plan forward, use another supplier or agree further work with Tenhaw. That could mean a focused proof of concept, operating-model design, a production build or programme support. Each is a separate scope, and a decision to stop is also a valid next step.
How many hours a week does the audit ask of our teams?
There is no single weekly allowance: it depends on the workflows, access arrangements and number of engineers pairing with us. We agree the expected involvement before you commit so your teams can make room for it.
Week one needs colleagues to walk through their work and help us understand the data. Weeks two and three need engineering time for pairing and colleagues to review prototype results. Week four needs the relevant leaders, with finance and risk input, to examine the findings and decide what follows.
Is the readout a document or a session with our board?
Both. You receive a written readout in seven sections and a session with your leadership in week four. We show the prototypes, explain the findings and work through the investment choices with you.
Bring the people who will own the decisions, alongside the finance and risk colleagues whose input the plan needs. They can examine the assumptions, ask about difficult cases and agree which questions still need answering.
Is the heat map a score out of five, or is there evidence behind it?
Each placement names its evidence. We rank the workflows examined by expected return and feasibility, using the workflow observations, available data, constraints and prototype results.
The two or three prototyped candidates have measured accuracy and an assessment of confidence in the results. Other candidates may have less direct evidence, which is made clear. Your team can see why a workflow sits where it does and which assumption or test could change its position.
Is four weeks long enough for an organisation our size?
Four weeks can answer a useful, bounded set of questions in a large organisation. Before starting, we agree which business areas and workflows to examine across the four domains and which two or three candidates to prototype.
It is not an exhaustive assessment of every operating company at once. We use the scoping conversation to establish whether the questions that matter can be answered within the time available. If they cannot, we say so before you commit and discuss a more focused starting point.
What if our board challenges the numbers in the investment case?
That is a useful part of the readout. Each value is stated in currency with the assumptions, confidence and arithmetic available for review. Your finance team can change the inputs and see how the case changes.
James Rooney leads the audit personally and can explain how the findings support the recommendation. The prototype results are there to examine too. A challenged assumption may change the order of investment or identify the next thing to test.
If the audit says do not proceed, do we still pay the £44,000?
Yes. The fee pays for the agreed assessment, prototypes and findings, including a recommendation to stop where the evidence supports it. You retain the written findings and own the deliverables on payment, whether or not you proceed.
A stop recommendation may apply to one workflow while others remain worth pursuing. The readout explains the reasons and the conditions that would justify revisiting the decision.
Does the costed plan go stale if we sit on it for six months?
Some assumptions may need refreshing. The plan marks review points within its three-month increments, capped at twelve months, so you can check what has changed before funding the next stage.
Model pricing, platform costs, data access, business volumes and team capacity can all affect the case. Your team can revisit those inputs alongside the baseline and constraints recorded during the audit. The age of the plan matters less than whether its assumptions still fit the work.
Do the prototypes cost us anything to run after the audit ends?
Model inference, compute and other platform use are charged on your own accounts during and after the audit. Those costs are separate from the £44,000 fee. You control the prototypes in your tenancy and can switch them off when you choose.
The audit includes an estimated run cost per unit of work for each candidate. Costs at production volume depend on the workload and the additional controls, monitoring and support needed for live use, so the prototype estimate is a starting point for that budget.
Can we put the audit prototypes straight into production?
No. They are built to test the selected workflows against your real data and provide evidence for an investment decision. A production system needs a separately agreed scope for hardening, assurance and operation.
Your engineers can build on the code in your repositories. The costed plan identifies the build work, dependencies and estimated costs involved in taking it forward, including the volumes and error tolerance the workflow needs to support.
What if our data access is not ready when week two starts?
We check access dependencies during scoping and address identity, permissions and risk classification early in week one. Your platform and data owners help us establish what can be reached and when, because the prototypes use your real data in your tenancy.
If access is delayed, we flag the effect on delivery in writing and work with your sponsor to agree how to proceed. We do not assume the original end date is achievable regardless of the delay. The fee remains fixed for the agreed scope; any requested scope change is agreed in writing before additional work starts.