The work we cover
An AI strategy is useful when it changes a funding, operating, or delivery decision. We help leaders decide which work to improve, whether AI is appropriate, what must be ready first, and who can deliver and own the result. The assessment is grounded in your priorities and current operation.
We examine candidate uses across customer service, finance, sales administration, internal knowledge, and other recurring workflows. For each, we connect the proposed change to a business outcome and a responsible owner. We distinguish a useful personal assistant from a change to an end-to-end business process, because their integration and operating needs differ.
Readiness is specific to the proposed task. A company may have usable records for invoice preparation while lacking approved sources for customer advice. We inspect source quality, access, system interfaces, staff capacity, relevant controls, and the cost of maintaining a release. The output states what can proceed and what depends on other work.
How we deliver
We start with the commercial decision. A managing director may need to protect margin; a scale-up may need to serve more customers with its existing team; a board may need to choose between several proposed investments. We record the constraint and the evidence needed to choose a response.
Interviews include operators, managers, system owners, and the person responsible for the economics. We follow samples of real work and inspect approved records. Ideas receive an assessment of potential value, feasibility, dependency, risk, and internal capacity. We make uncertainty visible rather than compressing it into an apparently precise score.
For a material technical uncertainty, we can agree a bounded feasibility test. That might examine whether documents contain the required information or whether a connector can retrieve it with the right permissions. Its scope names the question and evidence required; production implementation has a separate acceptance decision.
We compare existing products, changes to existing tools, rules-based automation, custom AI, and leaving the process alone. We then sequence the work, identifying prerequisites and stopping points. The National AI Centre's SME research identifies uncertainty about relevance and how to begin. Our assessment addresses those decisions in the context of the company's own work.
An illustrative engagement
This is an illustrative assessment, not a client outcome claim. An established services company has ideas for a customer chatbot, automated invoice preparation, and an internal policy assistant. Each team considers its own idea the priority.
We compare the current workload and consequences of error. The customer chatbot depends on account access and unresolved service policies. Invoice preparation has a defined document population and an existing approval process. The policy assistant has conflicting versions of several important documents.
The resulting roadmap explains which project has a testable first scope and which needs preparation. It names the owner of the policy corrections, the decision needed on account access, and the acceptance measures for a first finance workflow. The board receives a practical sequence with the reasons for each recommendation and can fund one stage without committing to the entire programme.
How we measure value
We test the business case against the current cost and quality of work. That includes manual effort, delays, rework, software, suppliers, and service consequences. Where reliable records are missing, we define a short observation exercise and state which estimates remain uncertain.
Benefits are separated into released capacity, avoided costs, direct cost reductions, revenue opportunities, and reduced exposure to errors. A manager's estimate of time saved does not become a guaranteed saving. Finance tests whether a proposed benefit can reach the P&L and when the required operating change occurs.
We also assess the cost of the decision itself. A worthwhile readiness engagement leaves management able to choose a first project, reject weak ideas, or fund a specific prerequisite. A long list of possible technologies is insufficient. We agree the questions the assessment must answer before starting.
What your team owns
You receive a prioritised opportunity register, source and access findings, recorded assumptions, material risks, and a delivery sequence. The first-project brief describes scope, exclusions, ownership, acceptance evidence, and the dependencies that affect a schedule. The findings remain useful for an internal team or another supplier.
We walk through the recommendation with business and technical owners. Each unresolved dependency has a named decision or action, preventing the roadmap from becoming an unowned backlog. If you proceed with us, the next engagement has its own scope and acceptance criteria; the assessment does not oblige a subsequent build.
Questions before you start
Do we need a company-wide AI strategy first? A focused assessment can be enough when there is a clear workflow and sponsor. Broader strategy work helps when teams are competing for funding, policies are inconsistent, or several dependencies must be coordinated.
Can you assess tools we already bought? Yes. We examine what the installed tools can do, how staff currently use them, and what access or process changes affect their value. Existing capability is part of the comparison.
What happens if we are not ready? We identify the particular gap, its business effect, and the smallest useful next action. That can be data integration, a clearer policy, process simplification, or a decision to defer the proposed use. We explain the evidence behind that conclusion.