Understand what is worth changing.

The owner kickoff confirms the important constraints, a practical interview plan and relevant records. The proposed assessment combines a staff survey, interviews, financial review and sample work. The depth is set for the business rather than imposed through a universal interview count or a fixed turnaround promise.

Before detailed information is shared, agree scope, fees, timing, confidentiality and permitted access. The request should be proportionate to the question. Where a dependency cannot be made available, that should be reflected in the assessment plan and the confidence of its findings.

A walkthrough follows a real item from trigger to finished output. Who receives it? What happens next? Which approval or input is needed? What causes an exception? Different accounts are compared so that an owner's summary or a single interview does not become the only version of the process.

Your team identifies the people who know the work, makes time for the agreed sessions and helps correct the emerging picture. Recording and AI notes, if used, need consent. The aim is an accurate account of operations, including workarounds that documents may not capture.

Financial inputs may include a prior-period profit and loss statement, receivables ageing and a balance sheet, with job or product margins, cash forecasts and administrative costs where relevant and available. These records are possible inputs to the agreed review. Confidential records should not be emailed before the relevant arrangements are agreed.

Evidence needs context. A sample may omit difficult cases; a staff time estimate may be approximate; an export may not be available in live operation. The readout should state those limitations and identify what further validation would change the recommendation.

The useful finding is the specific gap and what to do about it. For example, unclear process ownership needs a business decision; inaccessible records need an access or integration plan. An attractive overall label should not hide the dependency that prevents a workflow from being used.

Time released is capacity until something changes in spending or productive use. Earlier collection can release working capital once; it is not recurring profit. Any financing-cost effect depends on actual borrowing and terms. These categories should be presented separately rather than accumulated into one impressive annual number.

Estimates identify their inputs and assumptions. Missing information remains visible, rather than being replaced by unsourced averages. Financial review informs prioritisation and a test plan; it does not guarantee a return or extend into specialist advice outside the engagement's scope.

Feasibility includes integration effort, data preparation, operating costs, human approval, exception handling and maintainability. A technically possible workflow can still be a poor priority if it depends on unavailable information or more supervision than the team can provide.

Where included in scope, a bounded demonstration on approved samples can examine a specific uncertainty. Record what it shows and what it does not. A demonstration is not a client result or production acceptance, and there is no promise that every assessment will produce a profitable automation.

The format and deliverables are agreed, using the proposed structure below as a starting point. A focused assessment may cover fewer areas; a wider review may require more detail. The point is useful evidence and a clear recommendation, rather than a slide count.

If you proceed, a separate proposal confirms deliverables, owners, acceptance criteria, fees, timing and support. It identifies which assumptions still need testing and what would cause the plan to be reconsidered. Assessment credits, fixed delivery dates and commercial guarantees are not implied by this process description.

You can begin discovery with a broad description of the business and the work you want to improve. Detailed records, credentials and customer information belong in the agreed assessment arrangements, not in the introductory email.