Skip to content
AI & Automation

AI & Automation

Too much of the work is still manual. Not all of it should be automated. I help find the parts that should be, then build them.

Who this is for
Companies and organizations with real repetitive work: operations, support, admin, reporting. Often teams already using AI tools inconsistently, with not much to show for it yet.

The reason people call

This sounds familiar if…

  • People spend hours moving information between systems.
  • The same questions get answered by hand, over and over.
  • Your team uses AI tools inconsistently, and nobody can point at the gain.
  • You want agents but you're not sure what's actually worth building.
  • Someone built an automation last year and now nobody quite trusts it.
  • Leadership wants AI on the roadmap and nobody wants to own the maintenance.

One or two of these is normal. Four or five usually means the cause sits underneath all of them.


How this works

How I help

I start with the work, not the technology. Which workflows are expensive, repetitive, and stable enough to be worth automating. Then which of those are reliable enough to trust once they're running unattended.

Then I build them, or lead the people building them. I do both halves on purpose. The plan is more honest when the person writing it has to build the thing afterwards.

Anything that has to be right every single time is a bad first candidate. Start where being occasionally wrong is survivable, and where you can tell that it happened.

AI & Automation Opportunity Audit

Start with an Opportunity Audit

I go through the work your people actually do, week by week, and find where a machine would genuinely do it better. Some of what I find won't be AI at all. That's usually a good sign.

Where I look

  • Operations and the handoffs between teams
  • Support and the questions that repeat
  • Sales and admin work
  • Engineering workflows
  • Reporting, and who assembles it by hand
  • Internal knowledge nobody can find
  • Decisions people repeat every week
  • Data entry, and data moving between systems

What you get back

  • Usually around ten opportunities, written down with the hours each one actually costs today.
  • Each scored on business impact, build complexity, how reliable it needs to be, security and privacy, running cost, and who maintains it when it breaks.
  • The three I'd actually do, with the case for each and what it would take.
  • The ones I'd leave alone, and why.
How long
Scoped per audit
Language
English or Hebrew
Where
Tel Aviv, or remote

Depends on how many workflows there are and how deep we go into them. A first conversation is usually enough to size it.


After that

Where this usually goes.

Design the ones you pick, build them, wire them into the systems you already run, then measure whether they did what we said they would. Then either hand them over properly, or stay on to maintain them.

Strategy and implementation from the same person. It's the main reason the estimates hold up.


Not ready to talk

Take the letter instead.

One short idea a week about exactly this, plus two or three things worth reading. Read the last few first if you would rather see what arrives before deciding.



Contact

Tell me what's going on.

A few sentences is plenty. What's happening, roughly how big the team or organization is, and the thing you'd fix first if you had the time.