Senior AI Product Manager · Lisbon

What matters is how far we can go.

20+ years across product, operations and consulting. Technical enough to build what I design, from the business case to the code in production.

How I got here ↓ See the work
20+ yrsof professional experience
10+products I built running in production
48people led in a single team
11,000+people served by products I built

The computer is the bicycle for the mind.

Steve Jobs

That's how I see software: a tool people use to move humanity forward.

My thoughts about AI

Technology without context solves nothing.

What changes a business is not the most advanced tool on the market. It is knowing where, when, how and why to apply it.

Technology is a means, not an end.

When nobody can say which problem the AI is solving, it usually ends up as one more tool nobody uses.

Understand the problem, and know what to build.

That combination is rarer than it sounds. Years in sales, delivery and operations showed me where work really breaks, and most of what's on this page started because I saw someone losing hours on something the software could have done for them.

The technology is finally ready.

AI, automation, data: the tools are here. What was missing was someone who knows how to apply them with context.

The window is open now.

Companies applying AI with judgment today are building a lead that will be hard to catch later. That window won't stay open forever.

Where I come from

I come from operations.

Not from a trend. From years of operating, founding companies and consulting. That's the base of everything I build.

2012 → now
Scroll sideways to see the whole timeline →
How I work

How my products come to life.

If I had asked people what they wanted, they would have said faster horses.

Henry Ford
  1. 01

    Find the real need

    People often don't know exactly what they want. Figuring that out is my job, and every solution starts there.

  2. 02

    Validate before building

    I test the idea first: I talk to the people who live the situation, look at the market and try it in practice.

  3. 03

    Security from day one

    Security isn't the last step, it's a premise. Data protection and privacy from the first day.

  4. 04

    Applied intelligence

    AI comes in as a tool to improve processes and deliver features that make a real difference. Always with context, never technology for its own sake.

  5. 05

    Keep evolving

    Nothing I ship stays still. I listen to the people using it, look at real data and keep improving. Launch is only the beginning.

The work

Innovative products that make people's lives easier.

When market experience meets technical depth, you get solutions nobody had built before.

In production

Customers get answers in seconds, at any hour.

1,800+customer questions in 90 days
24 smedian first reply
    Open case →
    In production

    An AI platform for 3,500 people.

    14companies
    3,500people in scope
      Open case →
      simulated
      A
      Assistantfile tools connected
      In production

      Every purchase, explained.

      93%catalogued by the AI on its own
      0.4 sto refilter everything
        Open case →
        Personal build

        An agent built from scratch, on local models.

        19tools
        0cloud calls by default
          Open case →
          Live

          A decision panel for people in the field.

            Open case →
            Native apps

            On the phone and on the Mac.

            iOS

            Services in one app

              iOSWeb admin

              A members-only community

                macOS

                What the charger really delivers

                  macOS

                  Time zones in the menu bar

                    Operations
                    In production

                    A live portal, moved in one planned window.

                    91 minplanned window
                    0real divergences
                      Open case →
                      BeforeProvider's servers
                      AfterOwn cloud environment
                      Cutover window12:25 → 13:56
                      1. Maintenance page
                      2. Data checked
                      3. Real payment
                      4. Reopened
                      Creating with AI
                      Running 24/7

                      A designer that lives in my Telegram.

                      3bots, each in its own session
                      2 minto bring a dead session back
                        Open case →
                        Make a sticker of a pastel de nata surfing
                        Two options. A: faster, about €0.04. B: sharper, about €0.12. Which one?AB
                        B
                        sticker.png · example
                        Also built
                        Certificate orders with AI extractionReads an old certificate to fill the order. Stripe checkout, GOV.BR signature, tracked delivery.
                        Marketing dashboardFour sources, down to cost per sale.
                        Voice agent for outbound callsPhone calls with a native audio model.
                        Autopilot for a retired WhatsApp numberPoints people to the new number.

                        If this makes sense, pull up a chair.

                        Happy to talk about AI products, how they get decided, and anything on this page.

                        hello@mfgomes.com
                        LinkedIn ↗ Portuguese and English · Based in Lisbon