Evidence Infrastructure
for Controlled AI
DAAI helps organisations deploy AI inside their own approved environment, under defined evidence rules, with deterministic enforcement and an inspectable audit trail.
AI is becoming operational. Control must become operational too.
As AI moves from experimentation into operational use, organisations need more than fluent outputs. They need control over where proprietary information is processed, which rules govern the workflow, what evidence supports a result, and how the process can be reconstructed afterwards.
DAAI combines a version-controlled DAAI Evidence Assurance Standard with a deterministic DAAI Evidence Assurance Engine. Together they allow organisations to define, enforce and record the evidence rules that govern their own AI-assisted workflows.
DAAI has been shaped by regulatory, compliance and technical experience. Our objective is to combine practical workflow knowledge with controlled AI deployment, so that models may assist while accountable professionals retain interpretation, approval and release authority.
Regulatory experience behind the architecture
Dr Anzal Qurbain
Founder & Chairman
Dr Anzal Qurbain is a GMC-registered pharmaceutical physician and Final ABPI Signatory with more than 25 years' experience across medical affairs, compliance, safety and governance in global and local organisations.
He has made and defended thousands of real-world approval decisions, covering claim substantiation, fair balance, digital content and omnichannel materials. This experience brings practical regulatory judgement to the design of DAAI's evidence-assurance workflows.
As Founder and Chairman of DAAI Systems Limited, Anzal applies the same principle to AI: material outputs should be evidence-anchored, explainable, reviewable and accountable by design. DAAI is building the infrastructure that allows organisations to use AI without surrendering control of their data, rules or acceptance decisions.
The control gap in operational AI
Fluent outputs are not the same as defensible decisions.
In regulated and evidence-sensitive work, an answer is not defensible merely because it sounds plausible. Material statements may need to be linked to approved sources, correct versions, precise locations, applicable rules and an accountable review decision.
Organisations are under pressure to adopt AI while protecting proprietary information, controlling third-party processing, managing model changes and demonstrating how AI-assisted decisions were reached. Cloud access to a general-purpose model does not, by itself, create an auditable control framework.
The practical question is how AI can be introduced inside a controlled boundary, under organisation-specific rules, with a clear separation between model assistance and accountable acceptance.
Defined rules. Deterministic enforcement. Accountable acceptance.
DAAI applies models within structured workflows governed by a DAAI Evidence Assurance Standard. The DAAI Evidence Assurance Engine enforces the applicable rules, records the evidence used, identifies exceptions and prevents unsupported outputs from being accepted as verified. Models assist with extraction, retrieval, comparison and drafting; qualified professionals retain interpretation, approval and release authority.
The result is an AI-assisted workflow that can be deployed inside the organisation's own control environment and evaluated against its documented governance, quality and regulatory requirements.
What DAAI provides
DeepLink Citation
Extracts material assertions from a document.
Checks each assertion against approved source documents.
Identifies unsupported, partially supported, contradictory or stale citations.
Records precise source anchors and evidence receipts.
Supports regulatory, scientific, legal, financial and technical documentation.
Initially developed for pharmaceutical promotional and medical review workflows, with wider application across evidence-sensitive documents.
AQP Digital
Structures review stages, roles, decisions and escalations.
Applies organisation-specific rules to repeatable workflows.
Records reviewer comments, exceptions and dispositions.
Supports version-controlled process definitions.
Creates a consistent and inspectable review record.
AQP Digital was initially developed around pharmaceutical MLR and ABPI-related workflows. Its underlying structure can be adapted to other controlled review environments.
Evidence assurance and auditability
Version-controlled rules define what may be accepted.
Evidence receipts link material outputs to sources and processing steps.
Cryptographic records support reconstruction of the review.
Exceptions, overrides and human decisions are recorded.
Fail-closed behaviour prevents missing or contradictory evidence from being treated as verified.
Every material result should be reviewable,
challengeable and defensible.
Built at the intersection
of regulation and engineering
DAAI combines regulatory and workflow experience with software and AI engineering. That combination matters because evidence assurance is not solved by model selection alone. It requires defined rules, controlled data flows, deterministic checks, version control, human authority and an audit record.
The DAAI approach separates probabilistic model assistance from deterministic evidence acceptance. Models can help retrieve, classify, compare and draft. The DAAI Evidence Assurance Engine determines whether the defined conditions for acceptance have been met.
Regulatory insight and technical rigour are built into the Standard, the Engine, the sector adapters and the deployment controls.
DAAI selects, constrains, governs and verifies AI-assisted workflows.
Where DAAI can add value
Compliance Functions
Increase review capacity while preserving accountable human control.
Improve consistency and transparency across repeatable activities.
Introduce AI support with documented evidence, rules and escalation paths.
Reduce avoidable rework and unsupported acceptance.
Legal & Governance
Define how AI may be used within the organisation.
Maintain version-controlled rules and approval authorities.
Reconstruct material outputs and review decisions.
Support DPIAs, supplier reviews, model-risk assessments and internal audit.
Maintain evidence of exceptions, overrides and human acceptance.
Medical Affairs
Check scientific and medical claims against approved source material.
Identify unsupported, incomplete or contradictory citations earlier.
Reduce late-stage rework before MLR, publication or regulatory review.
Produce a clearer evidence record for accountable reviewers.
Senior Leadership
Establish a clear control boundary for enterprise AI.
Protect proprietary information and reduce uncontrolled third-party processing.
Create evidence that AI-assisted workflows are governed and reviewable.
Improve predictability around model changes, deployment and operating cost.
Support board, regulator, auditor and insurer discussions.
Designed for controlled deployment across jurisdictions
DAAI Systems Limited is based in the UK and is developing customer-controlled evidence-assurance infrastructure for organisations operating across the UK, EU and other regulated markets. Deployment, data location, governance and applicable rules can be defined with the customer and its advisers.
Discuss a controlled AI workflow
Tell us about a workflow where evidence, confidentiality, repeatability or accountability matters. We can explore whether DAAI's Standard, Engine or sector modules could support a controlled pilot.
We welcome conversations with prospective customers, design partners, platforms, systems integrators and investors.