Hi there, welcome to The Une, Weekly #10. Last week we gave Stewy, our AI Value Steward, the mic. This week we follow-up by sharing our vision for the kind of AI we’re aiming to build.

The Interview
Weekly #9 gave a little taste of the Steward's design in action.

We asked it to interview us about our own value compass. It chose to do so by exploring a dilemma with us. It looked for a possible tension, introduced the compass when the conversation needed it, gently pushed our proposed solution rather than affirming it, and generated one option. Then it handed the judgement back to us.

Make trade-offs visible, name what's under pressure, leave the decision where it belongs.
These are examples of some of the design choices we are making with the Value Steward.

In this week’s weekly we dive a bit deeper into our reasoning behind it.

The Agency Problem
There's a fear in conversations around the interaction with AI. That the machine will think for us. Decide for us. Taking away our agency, creativity and judgement.

People are starting to sense this is a real risk. Not in the last place because our human system is optimised for energy efficiency, and thinking is an expensive act.

We are primed for convenience. AI that replaces judgement will offer exactly that.

The result? Too little internalising and learning.
Decisions get made, but no one really knows why, or whether they'd make them again.

The Values Problem

Every AI already has an implicit value frame. An internal compass it follows. Values that are baked into the foundational model. Preferences, priorities, and resolutions of the organisations that built and trained it.

Those values might be largely invisible, but they are real and have real consequences. Slowly but surely the logic behind AI-supported decisions deviate away from the organisation’s core principles.

This is called value drift.
(As if organisations don’t already struggle enough living their values. AI scales this problem.)

But what if AI helps to find the way back? Again and again?
That’s what we are building the Value Steward:

From value drift to value coherence.

The Steward

The Value Steward is the AI tool of Value Design. At the heart of it is your Value Compass — your unifying intent, your commercial, social and ecological principles. Stewy carries your compass into the moments where decisions actually happen.

It doesn't optimise. It doesn't score. It never decides for you.

It exists to make people and teams more capable. More agency, not less. Sharper thinking, individually and together. The judgement and decision making stays where it belongs: with you.

The Partnership

The Steward is a partner that challenges, reflects, identifies tensions, runs scenarios, names trade-offs, and checks against the compass.

In practice, the Steward works with you in three ways:

Co-produce: Turning intent into concrete output. A positioning paper, a proposition, a ritual design — drafted from the compass, with the reasoning visible. Output that carries your principles

Co-create: Exploring what doesn't exist yet. Options, scenarios, directions — generated with your principles as creative constraints. The Steward enlarges what can be inspected. So you’re well informed when deciding what to turn into reality.

Co-conscience: Testing what's on the table. Proposals, decisions, documents — checked against the compass. Trade-offs, boundaries being crossed, norms being created: the Steward makes it explicit.

Most conversations move through all three. Draft together, explore alternatives, test the result. An integrated process.

You will notice Stewy will bring friction. That friction is what it is engineered for. If a choice strengthens one value, the Steward names which value comes under pressure — and leaves the weighing to you.

It’s not about the most convenient experience. But about the best possible decisions that contribute to the impact you aim to make: commercially, socially and ecologically.

(Tip: Dark Matter Labs founder Indy Johar recently wrote a reflection on thinking with LLMs. He describes working with AI as carving into a generative marble. The intelligence lives in the friction between you and the AI. Read the full piece here: Thinking with LLMs: A Reflection)

The Applications

We are developing various playbooks for applications — standard included and custom for clients. In each of these contexts, the Steward carries the compass into the room. It makes the relevant trade-offs visible, checks where a choice strengthens one value at the cost of another, and holds the direction.

Some playbooks work at the level of identity and expression — how the compass becomes visible and felt in the world, through brand, content, and positioning. Others work at the level of value creation — how the compass shapes the way the organisation grows, sources, and builds propositions that hold for all stakeholders. And some work at the level of daily life inside the organisation — how the compass lives through the people who join, the rituals they share, and the strategic choices they make when urgency pulls in a different direction.

One coherent value logic that helps the organisation to move as one.

The Aim

Indy Johar closes his piece with a warning: “organisations that use AI only to accelerate document production will produce enormous volumes of articulated knowledge — without increasing their collective capacity to know.”

We aim for the opposite. For every conversation to grow the team's own capacity to think and judge, together. To enhance value coherence. As a co-producer, co-creator and co-conscience, directed by the compass, carrying it into daily work.

Move as one. For the benefit of all. — Laurens & Jorn 👋 hello@une.eco