Composed goals are the difference between an organisation that uses AI and one that runs on it.

Decidr is an operating system for goals. Every person, every business and every organisation gets a structure to define what they are working towards — measurable ranges, at every level, all active at once — with rules, conditions and budgets built underneath, and the work assigned to whoever is best placed to do it. A person, a specialist AI, or conditional logic. Nothing runs that cannot name the goal it is for.

Strategy linked to execution. Every function, every process, every task type running under its own goals simultaneously. Results measured, evaluated and improved every period. Only the largest organisations in the world could operate this way. That is what Decidr is built to change.

Goal composition is the operating system of the agentic organisation.

Agents can do the work. Models can produce at a scale no human team can match. None of that answers the only question that matters: what is the work for?

Goals answer that. And when every goal is broken into smaller goals — across every function, every process, every task type, all running at once — something more powerful emerges. The organisation knows what it is trying to achieve at every level, in every period. Every agent, every automation, every specialist is working inside a structure that is connected all the way down.

That connection is what builds proprietary intelligence. Every result, every evaluation, every period adds to what the organisation knows about itself. Private, compounding, owned. The deeper the goal composition, the higher the automation, the sharper the decisions, and the faster the organisation moves toward what it is actually trying to achieve.

That structure is what Decidr builds. For a global enterprise, a growing business, or a single person running their life as an organisation for the first time.

Without structure, decisions follow the path of least resistance

Goals sit in people’s heads, get argued over in meetings, drift as they travel down an organisation, and are rebuilt every planning cycle. In a household or a life they are rarely defined at all. A decision with no structure takes the nearest path.

Tools get added to cope. Email, accounts, a CRM, project management, three chat apps. Each does its job. None of them know what you are working towards. The work runs. The goals don’t.

Running work is not the same as directing it

Automations

An automation repeats a step exactly. It has no way to tell that the step stopped serving the goal.

Assistants

An assistant answers well, often. It does not know your goals, so it cannot say which answer matters.

Decidr

Decidr starts from composition: your goals as ranges, in periods, with rules and budgets beneath them. Running the work comes second. Nothing runs that cannot name the goal it is for.

From a goal to work that earns its place

Define the goal as a range.

An outcome, a measure, a period and the levels between: green where you want to be, orange as a warning, red when it is off track. Goals are quantities in periods. A wish is not a goal.

Compose the goal downward.

Break it into function goals, process goals, task-type goals. Each smaller goal governs a narrower span and produces evidence more often. The more finely you see it, the better you know it. The more goals, the less uncertainty. Statistical confidence is what an organisation earns from its own evaluated periods — and it stays in your organisation, private and compounding, for as long as you use it.

Set the rules and the tolerances.

Plain conditions under each goal, with limits built in. “When a supplier invoice arrives, match it to its purchase order within one hour. If it matches — quantity within 5% and price within 2% — schedule payment for the last day of the discount window, up to A$10,000 without approval. If the discount window has already closed, schedule for the standard 30-day term instead. If it does not match within four hours, hold it and assign it to accounts. Never release a payment above A$10,000 without a second authorisation. Never pay an invoice with no matching purchase order.”

The more goals you define, the smarter the organisation gets.

Every goal is assigned to a person, a specialist AI or a conditional logic rule. As results land inside the goal range, autonomy and budget grow. As they fall outside it, scope narrows and a better specialist is procured.

This is continuous performance management. Every result measured against its goal range, every period, at every level. Nothing waits for a review cycle.

The more goals defined, the more focused each task type becomes. Focused task types attract hyper-specialised AIs and specialists with narrow spans and measurable outputs. Hundreds of goals evaluated every period build statistical confidence fast.

That confidence belongs to the organisation, not to the model that helped produce it. It lives in the goal structure, in the results, in the evaluated periods — sovereign, portable and compounding. No foundation model owns it. No platform can take it. It is the organisation’s own intelligence, built from its own goals, and it gets more precise every period.

Broadcast what you need.

Any goal can become an opportunity — your result range and your budget, broadcast anonymously into the Decidr marketplace. Specialists respond with the result range they will commit to and a price. You choose on the Decidr score: a clear gap between first place and second, at the defined purchase time. If that gap does not exist, the decision waits.

Evaluate every period.

Results land green, orange or red against the range. Green earns confidence and continued budget. Orange procures a specialist task from the marketplace. Red stops and reallocates. Every evaluation feeds two places: your organisation’s own earned confidence, kept private, and the group prior — shared anonymously in ranges, never in records.

Same goals. Yours differently.

Every organisation has the same structure: where you are, where you want to be, and how far apart those are. What differs is everything you link to those goals — the people, assets, beliefs and choices you have made, and the history of every evaluated period behind them. That set of links is your ontology. Decidr builds it from the questions you answer as you go. Those answers train your own AI. They stay yours, for as long as you keep using it.

See where you stand, without anyone seeing you

Every answer puts you in a group — the people and organisations whose answers match yours on that question. When a group reaches at least 50 members, it publishes ranges, not records. You see where you sit against the median, the minimum and the maximum for the same result type. No one can be singled out. Benchmarks that once needed a research budget arrive every period, as part of membership. There could be millions of groups. Decidr forms only the ones that change what you do next.

One structure for an organisation and life

Agents and robots can do the work. Something has to say what the work is for. That something is an organisation: goals, periods, budgets and results. Every person will need one.

BusinessOSGoals, rules, data and flows for an organisation in one structure. Every function, every process and every task type runs under goals that can be measured, evaluated and improved. The organisation learns from its own results — privately, continuously, period by period.
LifeOSA personal organisation for every person. Goals as ranges across every dimension of a life — health, wealth, relationships, learning, time. An entourage of specialists and agents works on those goals in the background.

Join innovators circle