The platform
One governed platform from source system to AI answer.
Kenerix makes your company data ready for AI. It connects every source, models it once into shared meaning, governs every request at runtime, and gives people and agents a safe way to ask, build, watch and share.
Architecture
Read the platform from the bottom up. Select a layer to see what it does.
The governance gateway runs alongside every layer: each request from a person or AI client is checked against policy before data is returned.
Your source systems
The systems you already run stay exactly where they are. Kenerix connects to them with read-only credentials over encrypted connections.
- Relational databases such as SQL Server and PostgreSQL
- Finance, CRM, HR and operational systems, through their databases or APIs
- Files and sheets: CSV, Excel and shared drives
- Online services, APIs and live feeds
Federation and storage
Two ways to bring data in, with one result: the layers above never need to know which one they are reading.
- Federation: query sources in place, with no copy and no migration project
- Hosted storage: import files, sheets or slow systems into Kenerix when speed or availability matters
- Every source, kept live, under the same governance
Data map
The single modelling surface for your whole estate. A federated table and a hosted table are the same kind of node.
- Entities, joins and relationships modelled once, across systems
- Links marked as exact, or as matching only after cleaning, with a note of the rule
- Foreign keys imported automatically where they exist
- The map tells every agent how your systems relate, so it never has to guess
Semantic layer and knowledge
Where data gets its meaning. Every source has its own semantic definition, and the business adds what the source never knew.
- Inherit: types, names and keys come from the source
- Override: business names, descriptions, classifications and units take precedence, without touching the source
- Virtual views: saved logic that behaves like a table but lives in the model
- Knowledge base: concepts, measures, reports and guides, certified by their owners
- Rules and monitors: quality and drift checks attached to fields and datasets
Gold layer
Curated datasets and business logic that give fast reads for reports, apps and agents. It sits inside Kenerix, so it is never built on your source systems.
- Business logic held in one place, not repeated in every report
- Pipelines, federated where possible, keep it current on a schedule
- Today shaped with our team; on the roadmap, built on demand by the report, app and agent builders
Development environment
Where people and agents turn data into something useful. Builders request data from the gold layer and the gold layer serves it back.
- Reports and dashboards: built from a plain-English description, checked and certified
- Agents: created from a prompt or an existing app, with workflows, routing and schedules, acting only within their knowledge, tools and access
- Apps: web and mobile apps from a flow chart or a description, on governed datasets (in development)
Capabilities
Nine capabilities, one platform. Use what you need now and grow into the rest.
Register each system once. Kenerix reads through read-only credentials over encrypted connections, browses schemas and tables, and samples rows so your team can see what is there before modelling.
- Databases, SaaS systems, APIs, files and sheets
- Federated in place or imported into hosted storage
- Read-only by default, encryption required
- No migration project to get started
Build a semantic model for each source: the entities that matter, how they join and what they mean. Save reusable logic as virtual views. Link systems on the data map so a customer in finance is the same customer in operations.
- Entities, relationships and meanings in business language
- Virtual views for logic the database does not hold
- Cross-system links, exact or normalised, with match notes
- Overrides that never touch the source system
A knowledge base that tells people and agents how your business works: the concepts, the measures with their exact rules, the reports and who they are for, and house rules for how answers are written. Owned by the team that knows, certified, versioned and searchable.
- Concepts, measures, reports and guides
- Certified content managed by its owning team
- Every change versioned; author marked as person or AI
- Backlinks and full-text search across the base
- Agent rules: tone, format, default scope, what never to change
Ask in the Kenerix portal or in the AI tool you already use. The agent reads the knowledge base first, uses certified definitions, and answers in your house style: a clear title, a short summary, the table and what stands out.
- Chat in the portal or through Claude, ChatGPT, Copilot and other MCP clients
- Answers grounded in your data and your definitions
- Each person’s answer limited to the data they may see
- Business words in the answer, not table names
Describe the report you need. The builder writes and validates the queries, previews real data, checks how the page renders and saves it in your private reports until you choose to share it.
- Filters that cascade, accept several values and search as you type
- Queries checked as read-only before anything is saved
- Private by default; share with people or move into team folders
- Certified reports for the numbers the business runs on
- Full revision history, compare any version with today’s data, restore in one step
Turn any measure into an alert in a sentence. Alerts run on the same certified definitions as your reports, so the number that triggers the alert is the number in the board pack.
- Business alerts on certified measures
- Data quality and drift rules on fields and datasets
- Routed to the person who owns the fix
Every request, from a person or an AI client, is checked against policy before data is returned. Access follows the job, not the tool: define roles and permission groups once and they apply in the portal, in reports and through every AI connection.
- Governance gateway for every AI client, with per-client grants
- Role-based access to brands, sites, rows and columns
- Read-only by default; write changes only as reviewed proposals
- Data retention set to your agreement
- A record of who asked what, through which client
When an agent thinks data or logic should change, it does not make the change. It raises a proposal with its reasoning and the exact change, runs a dry-run check with read-only credentials, and waits for a person to approve it.
- Proposals as tickets with title, rationale and SQL
- Dry-run plan check without touching data
- Human review before anything is applied
Publish to a secure portal of company reports and alerts. White-label it for your customers and partners, so each one sees their own data, governed by the same rules as your staff.
- Company, team and personal report spaces
- Shared-with-me view for colleagues’ work
- White-label branding for client-facing portals
- Certified badge on reports people can rely on
Security and trust
Built so your security lead can say yes.
Governance is the primary feature, not a checkbox. These are the controls and assurances in place and on the way.
Independently verified UK baseline for cyber security.
Privacy by design, retention to your agreement.
Independent attestation of our operational controls.
AI management system and information security certification.
Controls in the platform
- Read-only database credentials and encrypted connections
- Policy checks on every request, per person and per AI client
- Role and permission groups that follow the job
- Agents act only within their knowledge, tools and access
- Human approval for every proposed change
- Version history on reports and knowledge
Ready for regulated buyers
For financial services, healthcare and public sector customers, we map platform controls to the frameworks your risk team uses, including UK GDPR, the EU AI Act deployer obligations and FCA Consumer Duty, and provide the evidence they ask for during procurement.
Ask us for the current security pack.
Roadmap
Where the platform is going next.
From governed answers to governed action. Roadmap items are indicative and shaped with customers.
Governed answers
- Connections, semantic models and data map
- Certified knowledge base
- Chat, report builder and alerts
- Governance gateway for AI clients
- Change proposals with human review
Builders and the closed loop
- App builder for web and mobile, from a flow chart or a prompt
- Agent builder with workflows, routing and schedules
- Gold datasets built on demand by the builders
- Monitoring agents that re-run pipelines or rebuild datasets when rules trip
Assurance at scale
- Tamper-evident audit trail across every agent step
- Pre-execution risk scoring and blast-radius limits for agent actions
- Automated compliance evidence packs per sector
- Multi-region data residency
See it on your own data.
A live demo on a sample of your data, under your rules, in about an hour.