Who this is for
Support, CX, and operations leaders who have been asked to do something with AI and want an answer they can defend. You have workflows that run hundreds or thousands of times a month, a team that works them, and no dedicated AI group. It also fits founders who do not have an AI lead and do not want to hire one yet.
What happens in the assessment
I start with the work, not the technology. Over two to three weeks I sit with your team, pull volume and cost data from your ticketing, CRM, and back-office systems, and map how each workflow actually flows, including the handoffs nobody wrote down.
Every candidate workflow is scored on volume, cost, cycle time, error rate, and risk, then sorted with the Eliminate, Automate, Too Late framework: remove work that should not exist, automate what is routine and well defined, and keep people on cases where judgment or exposure is high.
- Workflow map with volume, cost, cycle time, and failure points
- Scored opportunity list, ranked by payback and risk
- Build, buy, or wait call for each opportunity, with the tradeoffs written down
- A sequenced 90-day plan with owners, metrics, and a budget range
- A one-page readout you can hand to the CEO or the board
Workflow inventory and scoring
I define each workflow by its trigger, inputs, output, handoffs, exceptions, and owner. The inventory records monthly volume, active handling time, cycle time, loaded labor cost, errors and rework, systems, and data sensitivity. Unknown inputs remain visible rather than becoming optimistic assumptions.
Value, feasibility, and exposure are considered separately. A large labor burden does not overrule a sensitive decision or an unreliable data source. The shortlist distinguishes work to eliminate, bounded automation to test, and judgment to retain. For a copyable starting point, use the workflow-scoring worksheet.
Human-review gates and measurement
For each shortlisted opportunity, I identify who approves consequential actions, what evidence they need, and when exceptions return to people. The plan includes baseline measures, pilot acceptance criteria, stop conditions, and a rollback owner. Released capacity is reported separately from realized cash savings, so the readout does not turn saved minutes into a payroll claim.
The 90-day sequence gives each next decision an owner and an outcome metric. Budget ranges and build, buy, or wait tradeoffs sit alongside the operational evidence. Return to the Session Player Studio overview for the broader approach.
What you will not get
No generic AI strategy deck. No list of forty use cases with no way to choose between them. No vendor pitch disguised as advice. The output is a short list and a plan, priced as a fixed fee so there is no incentive to stretch it.
What happens next
Most clients take the top opportunity straight into a Pilot to Production engagement, where I build and roll out the first workflow and measure it against the old way. Some run the plan with their own team. Either path is fine. The assessment is designed to stand on its own.