Turn company knowledge into a foundation for AI.
Start with an assessment of your data, priority workflows and access controls. Mickai can then digitise, clean, classify and organise the agreed records on your own hardware, creating a governed foundation for useful AI assistance. Deploying the Mickai Sovereign Intelligence Operating System is an optional next step.
The problem is never ambition. It is the foundation.
Useful AI depends on reliable company knowledge. Scattered files, duplicate records, outdated information and paper archives can make answers hard to trust. Unclear permissions can put sensitive material in the wrong hands. An AI Readiness assessment identifies these gaps, connects them to the work you want to improve and sets priorities for data preparation, access controls and human review.
Five stages, from filing room to AI-ready.
A staged, gated engagement. It starts with a fixed-scope, on-site assessment and ends with a governed estate you own. Every step runs on your own hardware, sealed to the Open Audit Record.
Assess
We review the agreed systems, drives, inboxes and paper archives against a readiness baseline. The fixed-scope, on-site assessment identifies useful workflows, data gaps and the controls they need.
Digitise
On-site scanning, on-device extraction and structured indexing make records in scope searchable. We retain provenance and flag unreadable material for human review.
Organise and clean
We normalise formats, reconcile duplicates and review conflicting records with your data owners. Unresolved gaps remain visible for a decision.
Classify and compartmentalise
We classify records by sensitivity and define who may access each group. Permissions are established before those records are made available to AI workflows.
Govern and hand over
We document lineage, ownership, retention and access controls, then hand over the agreed data foundation. You can stop at this stage or plan a SIOS deployment.
How the engagement runs.
The engagement starts with a fixed-scope Readiness Assessment. Each subsequent stage ends with a deliverable pack and a go or no-go decision. You can stop at an agreed gate with the work completed to that point: a score and report, an estate map, a captured corpus, a classified corpus or a governed foundation. Processing boundaries and audit coverage are defined in the scope.
Survey
The fixed-scope, on-site Readiness Assessment. Stakeholder interviews across the departments that hold data, a walk of the filing rooms, archives and data rooms, and a read of the digital estate, producing a first readiness score on sampled evidence.
You receiveAI Readiness Score and report, a data-estate summary, risk and exposure notes, and a scoped programme proposal.
Map
Exhaustive discovery across the agreed scope. An on-hardware crawl records what exists, where, in what format, how much, last-touched, apparent owner and duplication; the physical estate is inventoried to the box; systems and data flows are mapped; and a first cut of redundant, obsolete and trivial data is flagged.
You receiveThe data-estate map, a system and data-flow diagram, a redundant-data register, and confirmed scope for the stages that follow.
Capture
Records in scope are ingested into a staging area on the client's hardware. On-site scanning, OCR and document understanding extract text and capture provenance, with careful handling for fragile originals. Unreadable material goes to a human review queue.
You receiveThe staging corpus, text-extracted and provenance-stamped, and a digitisation report.
Refine
The corpus is cleaned, normalised and de-duplicated. Conflicts and gaps are reviewed with data owners, with unresolved issues documented. Records are classified by document type and sensitivity, then grouped according to permitted access.
You receiveThe cleaned, classified, sensitivity-tiered corpus, a data-quality report, and the compartment and permission model.
Govern
Lineage, ownership and retention are documented. The permission model and audit coverage are configured and reviewed against the agreed workflows, ready for the client's handover review.
You receiveThe governed data foundation, the permission and retention model, governance documentation, and the final AI Readiness Score with supporting evidence.
Deploy
Optional, and only if the client chooses it. Plan how SIOS will use the governed foundation on the client's hardware, including model selection, capacity, integrations, permissions and acceptance checks.
You receiveThe SIOS deployment plan, or a clean handover of the AI-ready estate if the client stops here.
An AI Readiness Score tracks discoverability, digitisation, quality, classification, governance and sovereignty. It is assessed first on sampled evidence, then against the cleaned data and at handover. The score and supporting evidence help you review progress against agreed acceptance criteria; they are not a certification.
The groundwork that makes AI work.
Data estate audit
We inventory structured, unstructured and paper sources in the agreed scope. A readiness baseline shows the gaps, risks and work required before deployment.
Paper and data-room capture
On-site scanning, OCR and document understanding produce searchable records with provenance. The originals stay under your control and unreadable items go to review.
Cleaning and de-duplication
We reconcile duplicates, standardise formats and flag conflicting entries. Data owners review exceptions so unresolved issues are visible at handover.
Sensitivity tiers and permissioning
Records in scope are classified by sensitivity and grouped by permitted access. The permission model defines which company knowledge each workflow may use.
The governed foundation
Lineage, ownership, retention and access are documented for review. The handover gives your team a data foundation and evidence to assess against its requirements.
Preparation on your hardware
Data preparation is designed to run on your own hardware, with offline processing where required. Source connections, transfers and audit coverage are agreed before work begins.
- A map of the agreed data estate: what you hold, where it lives, its condition and the gaps to address.
- Paper and data-room material in scope digitised and indexed, with provenance and a report of items requiring review.
- A cleaned and de-duplicated record set, with unresolved conflicts and quality issues documented.
- A governed, sensitivity-tiered foundation under your control, with a documented permission model.
- Lineage and audit evidence for the agreed workflows, including Open Audit Record verification where configured.
- A scoped path to SIOS deployment, or a handover of the completed data foundation if you stop there.
On-premise from the start
Plan assessment, digitisation and data preparation around hardware you control. Define the processing boundary, permitted source connections and handling of sensitive records before work begins.
Deliverables at each stage
The assessment provides a readiness report and scoped proposal. If you continue, the later stages produce captured, cleaned and governed records, with a handover at each agreed gate.
Audit evidence you can inspect
The Open Audit Record supports hash-chained records and ML-DSA-65 signatures, verifiable offline with the operator public key. The engagement defines which workflows are recorded and how that evidence is checked.
Deployment remains your choice
The completed data foundation is yours, on your hardware. A SIOS rollout is a separate decision based on the assessment, workflow needs and deployment plan.
Designed for sensitive work
Finance, healthcare, law, defence and the public sector inform the design. Permissions, sensitivity tiers and reviewable records support your governance work; compliance still depends on your deployment and operating practices.
A documented patent portfolio
The Mickai platform is supported by a portfolio of 104 filed UK patent applications and 2,340 claims. These are applications, not granted patents or evidence of certification.
AI readiness, in plain terms.
Can we start an AI readiness assessment before choosing a model or buying hardware?
Yes. Start by defining the business processes you want to improve, the data they depend on and the controls they require. The first assessment produces a readiness report, a data-estate summary, risk and exposure notes and a scoped programme proposal. Model selection and hardware sizing should follow those findings, rather than determine them.
Which business tasks should we automate first?
Prioritise frequent, time-consuming tasks with accessible source material and a clear way to check the result. Document search, drafting, meeting preparation and administrative handovers are useful candidates to assess. For each workflow, record the owner, current effort, expected benefit, failure cost and the point where a person must approve an action. Test a bounded workflow before expanding its authority.
What if our company records are paper documents or scanned PDFs?
The Capture stage covers scanning, OCR and document understanding for material in the agreed scope. It produces a text-extracted staging corpus with provenance and a digitisation report. Unreadable material goes to human review. Extraction alone does not make records ready for AI: cleaning, classification and access permissions follow in the Refine and Govern stages.
How much does AI readiness cost, and how long does it take?
Scope depends on the number of sources, data volume, document condition, access constraints and the governance work required. Mickai starts with a fixed-scope assessment and uses its findings to define the programme. Ask for the deliverables, timetable, responsibilities and price at each gate. There is no single published price or duration that applies to every organisation.
Do we have to deploy the operating system after the assessment?
No. Deployment is an optional stage. The engagement is organised around deliverables and go-or-no-go decisions, so you can stop at an agreed gate with the assessment or data foundation produced to that point. The assessment should help you decide what to do next, not commit you automatically to an operating-system rollout.
What is AI readiness?
AI readiness means having the data, workflows and controls needed for a useful AI deployment. Records need to be discoverable, usable and classified by sensitivity, with clear access permissions and a way to review results. Mickai assesses these foundations and scopes the work needed to prepare company knowledge for the workflows you choose.
How do I get my business ready for AI?
Start with a business workflow and an assessment of the data it needs. Mickai maps the agreed sources, including paper archives, then can digitise, clean, classify and organise them on your own hardware. Processing boundaries, source connections and transfers are agreed in scope. The result is a governed foundation and a deployment plan if you choose to proceed with SIOS.
What happens in an AI readiness assessment?
We review structured, unstructured and paper records in the agreed scope, interview data owners and assess sampled evidence against a readiness baseline. You receive an AI Readiness Score and report, a data-estate summary, risk and exposure notes and a scoped programme proposal. You can stop there, act on the findings yourselves or engage Mickai for the next stage.
Request an AI Readiness assessment.
Start with a fixed-scope, on-site assessment of your data, priority workflows and permissions. Get a readiness report and a scoped plan for preparing company knowledge on your own hardware. You can act on the findings yourselves or ask Mickai to deliver the next stage.