Workflow automation / Work with Kevin Manner
When the same task keeps crossing your desk.
Copying information between tools. Turning observations into reports. Chasing the next person in a process. I help investigate those recurring handoffs and build a focused way to handle them.
Choose one process you can describe.
A useful starting point is a task with recognizable inputs, a repeated sequence, and someone responsible for the result. Bring a recent example, the tools involved, and the places where someone has to correct or re-enter information.
We can compare connecting the existing tools, changing the process, and building a small application. The first decision is whether a change is worth making and what it should accomplish.
From my independent work
From observations to reviewable findings.
Input
A synthetic device assessment records that a backup has not been completed and that the disaster recovery plan is unknown, alongside other device observations.
What happens
Explicit rules separate confirmed problems from missing evidence, rank the findings, and group related work so one response is not counted several times.
Observable result
The checked fixture produces 10 response items and 3 evidence gaps, grouped into 7 recommended actions. Completing and verifying the backup appears first; the unknown recovery plan remains an evidence gap.
Human review
The assessor compares the findings with the field notes and verifies unknowns before relying on the report. The tool structures that review; the assessor remains responsible for it.
Decide what to automate—and what to review.
In this example, explicit rules own the findings and their order. An optional AI debrief can help with the narrative, while the assessor checks it against the observations. The decision about missing evidence is visible before any narrative is written.
For your workflow, we would agree where a record starts, which steps can run automatically, how incomplete or failed work is shown, and when a person needs to decide. Those choices define a useful pilot and its acceptance checks.
A first commitment with a useful result.
Start with a fit conversation.
Describe the task, how often it happens, the tools and people involved, and what you need to change.
Investigate if the path is unclear.
A separately scoped paid investigation can give you a workflow map, feasible options, key uncertainties, and a recommendation on whether to build. You keep that result even if no implementation follows.
Pilot one agreed task.
Use representative examples, clear review points, and an accepted result. Straightforward work can move directly to implementation; every project does not need each stage.
Make operation part of the scope.
Agree who owns the result, how failures are recovered, and what documentation is needed. Provider costs, changes, and ongoing support need explicit agreement.