Dr Abdelrahman
Elashrafi

Leadership & strategy

How I build a venture

I start with the people living with the problem, then test whether a team can deliver something they will actually use.

From observation to scale

  1. Observe

    Watch the work where it happens. Identify a problem that comes up repeatedly.

  2. Understand

    Ask users, buyers and operators what the problem costs them.

  3. Design

    Agree on a first offer, who it serves and who owns the decision.

  4. Build

    Recruit complementary skills and deliver a small, usable version.

  5. Validate

    Look for actual commitments, repeat use and workable delivery costs.

  6. Improve

    Find the bottleneck, change one thing and check what changed for the customer.

  7. Scale

    Expand when demand, cash and team capacity can support reliable delivery.

Decisions along the way

Choosing a team

I look for people who understand different sides of the same problem. Give each person an outcome to own and a clear way to flag a decision they cannot make alone.

Choosing a partner

A shared interest is a starting point. The next discussion is about responsibilities, delivery capacity, customer access and how both sides will make decisions.

Choosing what to pause

I compare the next commitment with the evidence so far. A promising idea can wait when the team cannot serve it well or customers have yet to commit.

How I would test an AI workflow

For an AI pilot that routes equipment enquiries, I would look at the time until a request reaches the right engineer. I’d count the requests someone has to reassign, then ask the engineer whether the information is enough to respond. That gives the team a reason to change the process or continue the pilot.

Read the note on measuring an automation pilot

Partnerships

Have a problem worth solving?

Discuss a partnership