Chama Inteligente

Intelligent Flame

Software shaped for your businessLisbon

Members

A coach for working with AI.

You meet with Elliot regularly and bring what you are working on. He keeps you up to date, sorts what matters to you from the noise, and shows you what your AI can do this month that it could not do last month.

A line loops upward in a widening coil. Each loop is a session with Elliot, and each one is wider and higher than the one before. The top loop is drawn in ember, and a faint line leaves it, heading up and out.

The shortest path up the learning curve

Experience matters more than the tool.

AI platforms add new models, settings, and features every few months, and they catch up with each other quickly. What decides your results is experience: knowing what a model will do with a request, how much thinking to ask of it, and how to check what comes back.

Each new capability is worth more to someone who mastered the last one, and they get it before it becomes easy enough for everyone. Someone who only chases each new tool starts over every time.

Since 2023, nearly all of Elliot’s work has gone into learning what the most capable AI systems can do. Coaching passes that experience on, so you spend less time working it out alone.

Two lines over time. Chasing the tools gives a short burst with every new capability, then starts over. Growing with AI compounds past it and keeps climbing.WHAT YOU CAN DOTIMEeach tick: a new capability arriveseach tick: a new capability arrivesCHASINGGROWING

How coaching works

One to one with individuals and small teams.

Because the sessions are regular, Elliot gets to know how you work and what you are trying to do, and each conversation builds on the last.

Bring the questions, decisions, and ideas in front of you. Or bring yourself, and Elliot will show you what is new since you last talked, or something older you have not tried yet.

You learn which model and settings suit each job, how to check the work, which news matters to you, and which features have been sitting in your tools unused.

One line, drawn left to right. It arrives tangled in loops: unsure what is possible. It turns ember and circles the flame once, a session. It leaves straight, ending in an arrow: with a direction.A SESSIONYOU ARRIVEunsure what is possibleYOU LEAVEwith a direction

Team training

Each new model can do more, and some old habits start to cost you.

When a new model drops, the team that learns what it can do first has it before its competitors do. A team that keeps the habits it formed on the last model runs into problems it did not have before.

One example: with the older Claude Opus models, Elliot never set thinking effort below extra high. With the current ones, high effort costs more and drifts from what you asked, and low or medium effort gives better answers. A team still on its old settings pays more for worse work.

Using AI the same way for months is a risk. Elliot trains engineering teams, and any team that works with AI agents, to pick up each new capability early and to review their habits every time a new model drops.

Four new models arrive one after another. After each one, a trained team, drawn in ember, climbs a step. A team that keeps the same habits falls into a pit at each new model, marked with a cross, and climbs out a little lower. The gap between the two grows with every model.TIMEa new model arrivesTHE GAPA TRAINED TEAMSAME HABITS

Building it yourselves

Prefer to build your own software?

We also build alongside individuals and small teams, and teach them to develop software the AI-native way.

Whichever suits you, we would be glad to help. See the software we build for businesses, or start a conversation and tell us what you have in mind.

Be AI-native.

Get in touch

Want to talk it through?

Start a conversation and ask Elliot to get in touch. See the software we build for businesses.