How it works
Follow the work before choosing the fix.
Elekevu starts with one recurring operational problem. We trace how the work actually moves, identify where the evidence points, and decide whether a change is justified before choosing technology.
One problem, followed properly
Choose one repeated symptom that can be observed.
Trace people, records, workarounds and handoffs.
Separate assumptions from what the process actually shows.
Choose the smallest justified next step.
1. Start with one problem you can point to
The useful starting point is not a wish to automate. It is a repeated operating problem: an order entered twice, stock checked again because the record is not trusted, work waiting on one approval, or a sale completed without a clear view of the margin.
The Fit Check is the first triage. It establishes whether the problem is specific enough, important enough and observable enough to justify a deeper look.
2. Follow what really happens
Process diagrams and system descriptions are useful, but they are not the whole operation. The diagnostic follows how work moves through people, conversations, records and systems, including workarounds that have quietly become part of the routine.
- Where does the work first enter the process?
- Which information is copied, checked or reconciled again?
- Where does work wait, and what is it waiting for?
- Which record is trusted when two sources disagree?
- What happens when the usual person is not available?
3. Test the evidence, not the assumption
A recurring symptom can have several causes. Slow approvals may be a capacity problem, a poor delegation rule, missing information, or a genuine control that is doing useful work. Repeated stock checks may come from delayed updates, undefined adjustments or several competing records.
The point of the diagnostic is to separate what looks plausible from what the process actually shows. That can mean reviewing a small sample of transactions, following a handoff end to end, timing a queue, comparing records, or asking the people doing the work where the exceptions really happen.
- A small sample of real cases rather than a hypothetical process.
- The records, messages or handoffs used to move those cases forward.
- Where time, re-entry, checking or ambiguity actually accumulates.
- Which exception rules and informal workarounds keep the process moving.
4. Choose the smallest justified next step
The useful response may be a clearer rule, a different owner, a process change, a simple tool, an integration, automation, AI, or a decision not to change anything yet. The size of the intervention should follow the evidence, not the ambition of the technology.
If the Fit Check points to something worth investigating, the next stage is a paid, fixed-scope diagnostic around one operational problem. It is deliberately bounded: enough depth to test the problem and a plausible response without turning the engagement into an open-ended transformation programme.
What you should leave with
The output is a decision you can act on, supported by the evidence gathered during the diagnostic. Depending on the problem, that may include a map of the actual workflow, the main sources of friction, the evidence behind them, and a small pilot or recommendation for the next step.
- A clear statement of the operational problem and what is causing it.
- Evidence that distinguishes the main issue from symptoms around it.
- A bounded recommendation or pilot, with assumptions made explicit.
- A GO, STOP or NOT YET decision on whether further investment is justified.
GO, STOP or NOT YET
GO means there is enough evidence to proceed with a defined next step. STOP means further spending is not justified by what we found. NOT YET means an important uncertainty still needs evidence before a larger commitment makes sense.
The technology, if any, comes after that decision.