Sharif Sediqui

<start of post>

Most companies don’t have an AI problem.

They have a prioritization problem.

Every week there is another tool, model or workflow that promises to make the team faster. And honestly, a lot of them are very impressive.

But adding more tools doesn’t automatically create more value.

We often see teams testing ChatGPT, Claude, Gemini, custom GPTs, agents, MCPs and automation platforms all at the same time. The result is usually a lot of experiments, but not a clear way of deciding what should actually move into production.

The business problem you are trying to solve

Who is going to use the solution

How often the process happens

What data is available

Which systems need to be connected

Who owns the result after launch

What governance and security restrictions apply

The last point is especially important.

A proof of concept can look amazing in a demo, but still be completely unrealistic in the real environment. Existing architecture, client data, internal knowledge and compliance requirements all influence what is actually possible.

This is also why I’m not a big fan of the “just automate everything” approach.

Some processes should be automated.

Some should be AI assisted.

And some are better solved by improving the process first.

The right answer depends on the context.

We’ve been helping teams map these opportunities and turn the most realistic ones into working solutions. Not because every company needs more AI, but because the right use case can make a very big difference.

I'm curious how you are currently deciding which AI ideas to implement.

And which ones to leave as experiments :)

<end of post>