Skip to content
    episcope logo
    episcope

    Every use of AI is a deal. Most of the terms go unread.

    episcope is an AI research and consulting studio working on the cognitive new deal: how AI augments your capabilities while preserving judgment, attention and control.

    In short: AI, on your own terms.

    Request a conversationGet the notes by email
    Professional panel discussion with blue stage lighting

    WORK

    Two ways to work together.

    A keynote opens the conversation in a room. Consulting takes it into your organisation. The opposite also works.

    Photo slot: portrait of Margaux

    Margaux Pelen in a burgundy blazer, leaning on a railing and looking at the camera

    Advisory and hands-on consulting

    For organisations already working with AI and the academies and consultancies that train them.

    Once AI is part of the work, the questions get harder. What should people hand over to AI? What should they keep? We start with a conversation and shape the work around your question. A workshop is often a good first step, yet some leaders prefer confidential sparring. Academies and consultancies can also build a module or a programme with me.

    "The workshop was very well prepared and insightful. The whole team felt both aligned and excited about what AI could do with and for us."
    CEO, Fintech startup

    Photo slot: Margaux on stage

    Margaux Pelen on stage at a Converteo all-team event, a screen behind her reading Please, Welcome on Stage

    Keynote

    For people who programme conferences and leaders who run in-house events.

    A keynote on the cognitive new deal: what AI gives and what it takes. Your audience leaves with one practice to try on Monday, not just findings or abstract research. Also available as a panel seat or a moderated conversation.

    Recent stages include several all-in company meetings, industry meetups or tech conference stages. I also co-chair GenerationAI.

    "Thank you Margaux, such a precisely tuned and rich talk!"
    Consulting CEO (500 people)

    Thinking alongside machines: what it gives, and what it costs

    AI adoption is accelerating and the conversation about its cognitive impact is lagging behind. Most AI discourse sits at productivity gains, tool comparisons or existential risk, and misses the middle ground where practitioners actually live: the daily experience of thinking alongside machines.

    The gains are real and measurable: AI lets us go faster and further with small teams. Alongside them, intensive use is changing how people think, learn and collaborate, and there is still no dedicated space for leaders and experienced practitioners to put words on that and work it out together. As agents move from writing the work to running it, the question shifts again, from what the output says to what a person is still able to check.

    The figures

    25%

    faster on tasks within AI’s reach, at quality judged higher

    Dell’Acqua et al., BCG and Harvard, 2023

    40%

    higher human-rated performance on those same tasks

    Dell’Acqua et al., BCG and Harvard, 2023

    39%

    more major errors among workers experiencing Brain fry

    HBR and BCG, March 2026

    17%

    drop in skill formation with AI assistance, paid first by whoever is still learning the job

    Shen and Tamkin, Anthropic, January 2026

    What I keep seeing

    Three patterns

    The gain is real and it relocates the work

    AI measurably improves output, speed and how the work feels. Aimed at extending people rather than replacing them, it creates more value than automation does.

    AugmentationAutomation

    Every AI delegation carries a cognitive tradeoff

    AI makes delegation frictionless. You prompt, it produces, you forward. Each handover also costs something at the moment it happens. Not better, or worse. Different.

    Cognitive taxCognitive debtBrain fryWorkslop

    The default is a decision somebody else made

    Nobody sits down and decides to hand over their judgment. The interface makes one path frictionless, that path gets taken and a default is a decision made by whoever designed the tool.

    Staying above the lineThe Joy of Not Automating

     GATHER

    Curated, intimate events

    Formats that keep the thinking close to those on the front lines of AI research, adoption and perspectives.

    Tandem dinners gathering
    Tandem dinners

    Private dinners carefully composed between practitioners, researchers and leaders on AI and a particular sub-topic. Each dinner centers on one theme and core questions. Guests are chosen for the tension they create in the room.

    Co-curated with Marine Buclon.

    DiscoverRead
    A burning piano on stage at a One Last Prompt gathering
    ONE LAST PROMPT

    The lab for AI practitioners. A show-and-reflect format: 5-minute demos, honest breakdowns, the things that are working and the things that are challenging your ways of thinking and working. 25 to 30 guests each time.

    DiscoverRSVP

     THINK

    The research is public

    The glossary

    The vocabulary of the augmentation AI makes possible, and the tradeoffs that come with it.

    • 26 terms
    • 33 sources
    • English and French

    Each entry says what the term means in plain words, what is at stake, where the word came from and what the research actually found.

    The repository

    The reading behind the glossary and the figures, in one public base.

    • Public, on Airtable
    • No sign-up

    Every paper and article read for this work, with the note that went with it. Added to as the reading goes.

    Get the notes by email

    The latest reflections on AI adoption and cognition, on Substack.

    ABOUT

    Who runs episcope and where it is coming from.

    episcope is an AI research and consulting studio led by Margaux Pelen, with a method built between fieldwork, research and strategic clarity.

    Who runs episcope?

    Margaux Pelen

    Margaux Pelen

    Berlin / Paris / San Francisco

    LinkedIn

    Margaux Pelen runs episcope, an AI research and consulting studio based in Paris and Berlin, with regular travel to California. Her work is on how AI augments people’s capabilities while preserving judgment, attention and control.

    She has spent fifteen years at the intersection of technology, work and learning. She studied at HEC Paris and Télécom Paris, founded her first startup during the first wave of French Tech and won a full scholarship to Singularity University at NASA Research Park (there, she met AI).

    She tests her framing in the room rather than only on the page, through the Tandem dinners in Paris, co-hosted with Marine Buclon, and One Last Prompt in Berlin. Both are run as applied research and sensemaking rather than marketing.

    The questions she brings to every stage: where does AI expand our thinking, what are the tradeoffs?

    What is episcope?

    episcope ['ɛpɪskoʊp]

    from Greek epi- "upon" + skopein "to view"

    1. an optical device that uses reflection to project images of opaque objects onto a screen.

    2. an AI research and consulting studio that acts like a reflective lens. An invitation to reframe strategy, projects and ideas with clarity and creativity when it comes to the impacts of AI on our minds.

    Start a conversation.

    If you are already past the basics and looking for a thinking partner who has been living in this territory, let’s talk.

    Request a conversationWrite me

    Or ping me on LinkedIn