Home
CounterfoldPlatform

From your first question to a working system.

Request a demonstration
The executive decision twin

Keep the reasoning.
Put judgment within reach.

A company is shaped by decisions that never make it into a manual. Capture the choices, rejected alternatives, risk tolerances and exceptions behind them. Give your team a way to consult that reasoning against the facts of today.

POLYDEUCES × SOLONIllustrative case · no live AI

Would we take this customer at this margin?

01 / The operation

18% contribution margin. Repeat order, prepaid. Capacity is available.

Illustrative order and capacity records
02 / The decision memory

A 20% floor, with a documented exception for prepaid repeat orders when capacity is available.

Illustrative, authorised decision record
03 / Recommendation for review

Recommend the exception for review.

The recorded exception fits these facts. Confirm that the costing is current before relying on it.

Confidence is limited to this precedent. An authorised person still approves the exception.
Rejected alternative and evidence limits

Rejecting every order below 20% would ignore the recorded exception. Taking the order without checking its costs would ignore the exception’s conditions.

These records are invented to explain the proposed interaction. They are not customer evidence or a measured model result.

Fixed examples. No founder has been modelled here; no order is accepted or changed.

Fewer decisions waiting on one person.

Staff can consult relevant precedents, see why an exception was made and know when the executive must decide.

Your company facts. Your decision principles.

The operational twin supplies current facts and dependencies. The decision twin represents the executive’s documented judgment about those facts.

A succession record the company can use.

Retain authorised reasoning when a founder steps back. Access, permitted uses and updates are agreed; a successor can challenge or retire old principles.

Discuss a decision-history evaluation
How we would prove it works

With several consenting founders, collect historical cases and reserve unseen decisions for testing. Compare recommendations and reasons with their actual choices, a written playbook and a retrieval-only baseline. Measure costly disagreements, uncertainty, abstention and authority breaches, then test whether teams use it when the founder is unavailable. Matching a past choice does not establish that it was a good choice.

Research on simulated people supports investigation, but has not established fidelity for consequential company decisions. Read Stanford’s account ↗

Proposed extension to evaluate. Source-linked recommendations, visible uncertainty and explicit action limits; no substitute for executive authority. Explore the operational foundation ↗

From each employee to the board

An agent for every role.
From understanding to action.

Give every employee a custom agent that understands their responsibilities, authorised working context and dependencies. Connect these digital counterparts of the work into one operational picture, so leadership can understand what is happening, why it matters and where to act.

  1. 01

    Each employee, supported

    An assistant for the work of each role: find evidence, prepare decisions and carry out authorised tasks. Daily work becomes easier while commitments, constraints and dependencies become visible to the right people.

  2. 02

    Every dependency, connected

    Department agents connect the handovers between teams. A local exception can be traced to the customer promise, budget or strategic objective it affects, before it disappears inside a summary.

  3. 03

    Leadership, able to act

    The CEO and board can move from a company-wide question to the underlying tasks and evidence within their access rights. Compare interventions, set priorities and spending limits, then route approved action back to accountable owners.

Evidence flows up. Approved decisions and priorities flow back to the responsible teams. Sources, uncertainty and dissent stay visible; human authority and access boundaries remain in force.

The executive advantage

Ask why. Test what changes. Direct the response.

A connected agent organisation closes the loop between what leadership decides and what teams actually do. Explore the intended workflow:

Which commitments are at risk, and why?

Trace an executive signal through departments to the affected workflow, source record and responsible role. Preserve uncertainty and conflicting assessments instead of averaging them away.

A connected explanation, with evidence behind each dependency.
Company-specific models · deployment choice

Your expertise. In models you can host.

Why local matters: keep authorised workloads within your chosen infrastructure, encode your terminology and working methods, and decide when models change. Tested open-weight models can reduce dependence on a single provider. Quality, latency and total cost are measured before choosing the deployment.

A deployment can include locally hosted open-weight models such as Qwen, adapted to your terminology and recurring tasks. We can fine-tune model weights or specialist adapters using approved, reviewed examples from your organisation.

Connected sources supply current facts. Training adapts task behaviour. We compare prompting, retrieval and fine-tuning on held-out work before choosing the model, hardware and deployment. On-premises, private European hosting and approved frontier services are options to assess together.

Discuss your deployment requirements

This is our custom-delivery proposition. Scope, permissions, model licences and performance are established per engagement. The public demonstrations do not run an employee agent network or a company-trained model.

Why Counterfold
Our mission
Give European organisations the ability to understand change, rehearse consequential decisions and act with authority over their intelligence.
Our vision
A Europe whose institutions and businesses can turn their knowledge into systems that anticipate, adapt and improve under their own control.

See beyond
the expected.

The next disruption may not resemble the last. Connect changes across systems, challenge the assumptions behind your plan and rehearse responses before the choices narrow. Show what is known, what is inferred and what would change the decision.

Counterfold designs sovereign intelligence systems around the decisions that define your operation. We connect evidence, model dependencies, explore possible futures and build governed AI workflows that carry a reviewed decision into action. Each outcome can become evidence for the next improvement.

01. Understand what happened.

Operational systems, live events and documents hold different parts of the story. Bring them together, with their source and history intact.

02. See what is connected.

A delay affects an asset. That asset belongs to a rotation. The rotation depends on people and time. An ontology makes those relationships usable.

03. Explore what could happen.

Surface unusual changes, trace their possible consequences and compare responses. Bring evidence, domain models and human judgment together, with assumptions and uncertainty visible.

The anatomy of a decision
Individually, fragments.Together, intelligence.Scroll to connect the pieces
Human authority stays in the loop.Illustrative architecture · scoped for each project
Custom by design

Your operation
defines the system.

Your decision horizon, operational vocabulary and authority model become the blueprint. We connect your systems, encode the real constraints and design agents, simulations and interfaces around the people who use them.

01 / THE DECISION

Start where value is clear.

Agree the question, its operational owner and the outcome that would make the work worthwhile. Establish success measures before choosing the technology.

02 / THE FOUNDATION

Build from the evidence.

Confirm data access and quality. Define the business objects, domain assumptions and operating constraints with the people who know them.

03 / THE WORKFLOW

Make it fit the operation.

Select integrations, models, interfaces and controls for the task. Test representative cases, failure conditions and permission boundaries with your team.

04 / THE NEXT QUESTION

Expand from what works.

Replay real decisions, establish a baseline and run in shadow mode. Extend the system as evidence supports it, with operating ownership and evaluated releases agreed from the start.

What you can shapeSources and update cadence. Business definitions and models. Deployment and access. Operator interfaces and approvals. Evaluation, support and ownership.

Begin with the question

What could your data make possible?

Start with a scenario and the systems involved. An initial conversation does not require confidential data.

Request a demonstration