Theodoros Ampas
Join the newsletter

Understand first

Find out where AI is actually justified before you build anything.

Workflow Discovery maps how sales, marketing, operations, customer service, handoffs and decisions work in your company today. You leave with a ranked list of bottlenecks and repeated work, and a clear answer on which ones deserve an AI layer.

Start a project

Every engagement starts here. If the answer is that AI is not the fix, you will hear it in the report.

By Last updated

In one sentence

Workflow Discovery & AI Opportunity Mapping is a short, paid engagement in which Theodoros Ampas documents how a company’s work actually happens, measures where time and context are lost, and ranks the workflows where an embedded AI system would pay for itself.

Most AI projects fail at the first question, not the last one.

The usual starting point is a tool. Someone sees a demo, buys a licence, and then looks for a workflow to put it on. Six months later the licence is still paid and the workflow is still manual.

The teams I work with do not lack ideas. They lack a shared, written picture of how a lead, an order or a support request actually moves between people, inboxes and systems, and where it stalls.

Discovery produces that picture first. Only then does the question of AI make sense, and usually the honest answer is “yes, but only in two or three places”.

What discovery looks at.

How the work flows

Who touches a lead, a request or a document, in what order, in which tool, and where it waits for someone to notice it.

Where time and context leak

Repeated copy-paste, re-keyed data, duplicate follow-ups, decisions made without the information that already exists elsewhere.

Where AI is justified

For each candidate workflow: the baseline today, the smallest useful system, the risk of getting it wrong, and the number that would prove it worked.

Two weeks, three steps.

  1. Days 1 to 3

    Observe

    Short sessions with the people who do the work, not only the people who manage it. Screen recordings, sample records, the actual inbox.

  2. Days 4 to 8

    Map and measure

    A written map of each workflow with time per step, handoffs, tools and failure points. A baseline the team agrees is fair.

  3. Days 9 to 10

    Rank and recommend

    A ranked list of opportunities with expected return, risk and effort. A build recommendation you can use with me or with anyone else.

A map, not a sales deck.

  • The report is yours whether or not we build together. It is written so another engineer could act on it.
  • Every opportunity carries a number that would prove it worked, agreed before any build starts.
  • If a workflow should stay manual, the report says so and explains why.
  • No customer data leaves the company’s accounts during discovery.

What you leave with.

  • A written map of the workflows that matter, with time, tools and handoffs per step.
  • A baseline for each candidate workflow that the team accepts as fair.
  • A ranked list of AI opportunities with return, risk and effort.
  • A build recommendation, or a clear recommendation not to build.

This is for you if

  • A capable person is spending hours a week on work that should be a system, and nobody has written down exactly what that work is.
  • You have tried an AI tool and it did not stick, and you want to understand why before trying again.
  • You are about to commission a build and want an independent view of what it should be.

Questions founders ask

How long does discovery take?

Usually two weeks of calendar time, with around ten to fifteen hours of your team’s involvement spread across short sessions.

Is discovery paid?

Yes. It is a fixed-fee engagement. The fee is credited against a build if we continue together.

What do you need access to?

Read-only access to the tools where the work happens, sample records, and time with the people who do the work. No changes are made to any system during discovery.

What if the answer is that AI will not help?

Then the report says so, ranks the non-AI fixes it found instead, and the engagement ends there.

Start with the map.

Describe the workflow that is costing the most time and I will tell you on the first call whether discovery is the right first step.

Start a project