Workflow & process mapping
See how work really flows, including the spreadsheet nobody admits to.
Roadmap: Discovery & consulting core
Why it matters
AI and automation create value at decision points, the moments where a person stops, gathers context and makes a call. You can only find those moments by mapping how work actually happens. Documented processes almost never match reality, and the gap between them is where most deployments go wrong.
What it's made of
- Mapping actor → step → system → data → decision → outcome
- Finding the real bottleneck, not the loudest complaint
- Cataloguing exceptions and escalation paths
- Marking each decision as reversible or irreversible
- Spotting cost asymmetry (a miss vs a false alarm)
- Defining where a human must approve
Map the work, not the org chart
Ask someone to show you the last real case they handled, on their screen, start to finish. Write down every tool they open, every copy-paste, every time they ask a colleague, and every place they wait. That list is your map.
Then lay it out as a chain: who acts, what they do, in which system, using what data, what they decide, and what happens next. Most opportunities show up as obvious repetition or waiting.
Find the decision points
Mark every step where a person uses judgment: approving, classifying, prioritising, choosing a reply. These are the candidates for AI assistance. Steps that just move data around are usually better served by plain integration code.
- Is the decision reversible? Reversible decisions can be automated sooner.
- What does a wrong answer cost? Missing a fraud case might cost 100× a false alarm.
- Who is accountable today? That person needs to stay in the loop at first.
Exceptions are the real spec
Ask "what's the weirdest case you handled last month?" and "when do you escalate?" The answers define the edge cases your system must handle or route to a human. A design that ignores them works in the demo and fails in week two.
Mid-level looks like
Maps the happy path and the main steps accurately, based on interviews and documentation.
Senior looks like
Watches the work directly, finds the shadow spreadsheet, the political constraint and the step whose output nobody actually uses, then redesigns around them.
Common pitfalls
- Designing from the org chart or the SOP instead of watching real work.
- Mapping only the happy path. Exceptions are usually 20% of cases and 80% of the effort.
- Automating a step that nobody owns, so nobody fixes it when it breaks.
Practice drills
Drill 1
Map a refund process end to end for an online retailer. Mark each human decision as reversible or irreversible, and pick the single best step to assist with AI.
A strong answer: A complete chain including exceptions (partial refunds, fraud checks), clear reversibility labels, and a justified choice of a reversible, high-volume decision.
Drill 2
For a ticket-triage problem, write category definitions as tests ("needs someone to look at this specific account") rather than topics. Test them on 10 sample tickets with a partner and measure how often you agree.
A strong answer: Definitions precise enough that two people agree on most items. That agreement rate is the ceiling any model can reach.
Best resources for this skill
Repos, courses, articles and podcasts, ranked by how much they teach.
Read next
The Mom Test
Rob Fitzpatrick
A short, practical guide to asking customers questions that produce facts about their current behaviour instead of polite encouragement. Use it before every discovery session so you scope the real problem rather than the one they pitch.
The Trusted Advisor
David H. Maister, Charles H. Green, Robert M. Galford
Breaks trust down into credibility, reliability, intimacy and self-orientation. FDEs live or die on whether the customer's team lets them into their systems and their problems, and this gives you a model for earning that.
The Secrets of Consulting
Gerald M. Weinberg
Witty, compact rules about how clients actually behave: why they resist the change they paid for, how to price and pace your help, and how to avoid burning out. Directly applicable to multi-month customer embeds.
Flawless Consulting
Peter Block
Walks through each phase of a consulting engagement (contracting, discovery, feedback, implementation) with emphasis on the contracting conversation. Useful for setting expectations with a customer sponsor before writing any code.