Skip to content
Free AI Benchmark — see where you are exposed before you deploy AI.Start the benchmark

Why MergeOn

Turn the knowledge you already have into governed context AI can use.

Policies, procedures, manuals, regulations and operating documents were written for people — not AI.

MergeOn Document Intelligence transforms them into structured Tier-3 Review Packs, preserving meaning, relationships, dependencies and source context. Accepted knowledge can then be governed in the Knowledge Center and supplied to AI at execution.

Better context. Less repeated processing. More efficient AI execution.

Explore Governed Knowledge

Build on MergeOn

The application asks for a business outcome. The Runtime governs how it is produced.

An integration calls a published Business Capability rather than a model endpoint, so knowledge, policy, protection, human authority and evidence stay part of the activity instead of becoming your problem to rebuild.

Everything below describes the contract you build against and the architecture behind it.

Explore the developer platform

Know where you stand

Most organizations do not have an AI problem. They have a clarity problem.

Before deciding what to build, it helps to establish what your organization already believes about ownership, governance and decision-making — and where those beliefs disagree with each other.

Start with an honest read of where you are. Everything else follows from it.

Start the free benchmark
Platform/Govern/AI Model Governance

Govern the model.
Not just the application using it.

Enterprise AI depends on models that change, providers that change and operating contexts that change. MergeOn keeps the model, its qualification, permitted use and governing evidence connected to the activity in which it participates.

Know the model. Know where it may operate. Know what evidence supports it.

Governed model context

Model

Model identity

Qualification

Evidence available

Permitted use

Governed context

Change

Current state
Model governance

A model should not become trusted just because somebody connected an API.

Using a model and governing a model are different things. Enterprise teams need to know what model is participating, which provider or deployment it belongs to, what is known about it, what evidence has been established and the governed contexts in which it is permitted to operate.

MergeOn keeps that model context connected to governed AI activity rather than leaving it scattered across applications, configuration files and review documents.

Identify

Know which model or model version is being governed.

Qualify

Associate the evaluation and assurance evidence relevant to its intended use.

Admit

Determine whether it may participate in a governed context.

Observe

Retain the relationship between the model, its use and the evidence produced around it.

Model agnostic by design

Change the model. Keep the governed operation.

The model is not the operating system.

MergeOn keeps the business activity, knowledge, policy, protection, authority and execution controls outside the model. The model participates within that governed context.

That means teams can qualify a different model, change the model used by an activity and retain the governed structure around it — rather than rebuilding the application around a new provider.

Governed operation

KnowledgePolicyProtectionHuman ApprovalToolsExecutionEvidence

Model

Model AModel BModel C

The model can change. The governed operation remains.

Provider

Where the model or AI service comes from.

Model

The governed model or model context participating in the activity.

Governance

The controls, qualification and evidence associated with its permitted use.

Not a blind swap

Swap without rebuilding. Re-qualify without starting again.

Plenty of platforms can call several model APIs. That is not the same as keeping the enterprise control structure out of whichever model happens to be selected.

The governed operation stays put. The new model, however, is a new probabilistic participant — so its qualification and evidence position still has to establish whether it is suitable for that particular activity.

Make the operating system deterministic around a probabilistic model.

01

Govern the operation independently of the model

Knowledge, policy, protection, authority and execution remain controlled by MergeOn.

02

Qualify models for a specific use

AI Evaluation & Assurance establishes evidence about how the model performs in that governed context.

03

Record that position in AI Passport

Identity, version, qualification, permitted use and supporting evidence travel with the governed model.

04

Change models without rebuilding the operation

A different model can be introduced into the same governed activity and evaluated against the same operating requirements.

05

Know exactly what actually ran

Golden Thread retains the model and governed execution context associated with the operation.

AI Passport

Give every governed model an assurance record that can travel with it.

A model name is not enough. Its governance position depends on more than who supplied it. Teams need the relevant identity, version, qualification, evaluation, permitted-use and evidence context available when deciding whether that model is appropriate for a governed activity.

AI Passport is MergeOn’s assurance record for a governed model: the context that allows a model’s use to be examined against what is actually known about it.

AI Passport
Identity
Which model or governed model reference this record concerns.
Version
The model and version context associated with the assurance record.
Qualification
The assurance or evaluation evidence relevant to its use.
Permitted context
Where the model has been considered for governed participation.
Change
The model context against which the current evidence was established.
Evidence
The records supporting the governance position.
A passport is evidence of what is known and governed. It is not a claim that a model is universally safe or suitable.
Permitted use

The question is not “Is this a good model?” It is “Is this model suitable here?”

The same model may be appropriate for one governed activity and inappropriate for another. Its use can depend on the business capability, environment, release, policy, protection requirements, human authority and evaluation evidence surrounding the activity.

Model governance therefore belongs to the operating context — not to a static approved-model spreadsheet.

ModelGoverned activityOperating contextAssurancePermitted use

Qualification is meaningful when it is attached to the context in which the model will actually operate.

Qualification

Evaluation should change what you know about the model.

AI Evaluation & Assurance establishes evidence about behaviour, performance and the conditions being evaluated. AI Model Governance connects that evidence to the model and the governed context in which teams intend to use it.

A benchmark can describe a model in general. Governed evaluation can help establish whether that model is suitable for the specific activity being considered.

Evaluate

Measure the behaviour or condition that matters to the governed activity.

Associate

Keep the result connected to the model and context it was produced against.

Govern

Use the available evidence when determining the model's governed position.

When models change

A new model version is new evidence territory.

Models change. Providers change them. Configurations change. The activity around them changes. Evidence established for one model, version or operating context should not silently become evidence for another.

MergeOn keeps the governed model context explicit so teams can determine when previous qualification still answers the question — and when further evaluation or assurance is required.

Established state

The model and context against which the current evidence was produced.

Change

The model, version, provider, configuration or relevant operating context changes.

Review

Determine whether the existing evidence still answers the governance question.

New evidence

Where required, further evaluation establishes the position under the changed context.

At execution

Know which model actually participated.

Model governance matters most when it remains connected to real governed activity. The model context associated with an execution can form part of the evidence needed to understand what operated, under which controls, against which qualification context and with what outcome.

AI Model Governance

Establish the governed model context and supporting assurance.

Runtime

Execute the governed activity with its applicable operating context.

Golden Thread

Connect the model and governing context to execution evidence where recorded.

Over time

A model’s governance position should not end with approval.

Individual evaluations establish evidence at a point in context. Governed execution produces further evidence about how AI is operating. Where that evidence is available, ATHENA can reason across evaluation and execution evidence to help surface patterns, changes and areas requiring investigation.

AI Evaluation & Assurance

Evidence about behaviour.

AI Model Governance

Governed model context.

ATHENA

Intelligence across available evidence.

How it fits together

Know it. Qualify it. Govern it. Evidence it.

ATHENA

Reason over available model, evaluation and execution evidence.

Move past the pilot.

Put your first governed AI activity into production.