No models, no tools — just three habits you can start today: feed "you" to the AI.
How structured thinking gives your AI superpowers ft. Carola Pescio Canale (Atlassian) | Config 2026 · Carola Pescio Canale
18 min total·Actually worth watching closely: ~4 min·2 must-watch clips
- 0:12 – 1:15Listen
The problem isn't the model, it's the context you never gave it
She opens by taking the blame off the model: AI answers disappoint not because the engine is weak but because it doesn't know you at all. In her words — intelligence is the engine, context is the fuel.
Only you can build your personal context; stop waiting for a stronger model to rescue you.
It's the speaker standing and making an argument; the screen holds only her and one line of text, so listening is enough.▶ Jump to 0:12Speaker · Carola Pescio Canale - 1:15 – 4:20Listen
Habit one: ten minutes a day to empty your brain
Use any recorder that gives you a transcript and talk your head empty with no structure at all — a three-minute ramble or a twenty-minute rant both count, ten minutes a day at most.
It doesn't have to be coherent; the act of externalizing compounds by itself.
This is a habit and a mindset threshold, with nothing to look at — you won't miss anything listening while you do something else.▶ Jump to 1:15Speaker · Carola Pescio Canale - 4:20 – 7:14Listen
Turn the rambling into a thinking document
Tidying up the ramble takes no technical skill: write four bullet points and hand them to the AI to build itself — take the transcript, digest it, make sense of it, turn it into a retrievable document. The crucial step is having it gather what's worth keeping into one thinking document.
With that document in place, every later brainstorm already knows where you stand and how you think.
The four bullets are spoken aloud and short enough that you can just write them down — no screenshot needed.▶ Jump to 4:20Speaker · Carola Pescio Canale - 7:15 – 10:00Listen
Habit two: teach the AI your tone
It sounds like AI because what it writes is everyone's average. Fixing that doesn't require a writing résumé: your sent mail, the last thirty days of chat, and everyday conversations recorded with permission are all material already sitting there.
A handful of samples is enough for the AI to learn your voice — the prerequisite for letting it write on your behalf.
It's argument and examples end to end, about where to find the material; hearing it clearly is all you need.▶ Jump to 7:15Speaker · Carola Pescio Canale - 10:00 – 13:10Watch
Habit three: your phone camera as a prompting machine
A photo plus a three-word prompt keeps building your personal context. She walks through the selling example: photograph an old object, say only "garden statue, sell", and the AI identifies it, compares prices and estimates value — a few months of snapping as you walk around builds a whole for-sale inventory. Sketch by hand, photograph it and have the AI build straight from that: often the least-effort workflow there is.
A photo is worth 100 prompts — and you don't need to know what you'll use it for when you take it.
All the persuasion here lives on screen: the photo going in, the comparable price and valuation coming out. Audio alone sounds like bragging.▶ Jump to 10:00Speaker · Carola Pescio Canale - 13:10 – 14:17Watch
Merge the paper world with the screen
Photos and voice connect the physical world to the digital one: a task handwritten on paper is photographed once and merged into the online tracker. She also calls out the common failing of AI daily digests — things already resolved offline still show up marked urgent.
Until the two worlds connect, neither your task list nor the AI's suggestions count as truly personal.
The moment handwritten notes merge into the online list happens on screen; you need the before-and-after to see what "merge" actually means.▶ Jump to 13:10Speaker · Carola Pescio Canale - 14:17 – 18:04Listen
From a personal graph to a team graph
The close widens the frame: the personal graph built from opinions, decisions, tone and photos is only a starting point, and what you externalize both feeds the AI and can be shown to people. Put project decisions into tooling the whole team uses, and the AI can go from a personal assistant to a collaborator.
Context multiplies at the team level, not on your own hard drive.
This last stretch is a summing-up that scales the three habits upward, with no new demo — just listen for the conclusion.▶ Jump to 14:17Speaker · Carola Pescio Canale