AI integration
We add AI to web applications: assistants, suggestions and natural-language questions about your data. Every feature respects user permissions, can be switched off, and has its cost tracked.
- Adds
- Assistants, suggestions, AI search
- Control
- Per-feature switches
- Security
- Role-based access
- In production
- Keruja
What we add
- In-app assistant: a help and question-answering panel that knows your product.
- Suggestions: for example, which available staff best fit a shift, with the reasons shown.
- Questions in plain English: ask about your own data without building reports.
Built to be safe in production
- Each AI feature has its own switch, plus one master switch.
- Server-side permission checks; some user groups can be excluded entirely.
- Usage tracked so costs can be capped or billed on.
- People confirm decisions; AI suggests.
Proof
These patterns run in production in Keruja, where admins get AI staff suggestions and plain-English questions, while client and staff portals have no AI at all.
Questions
Can you add AI to our existing product?
Yes. We add AI features to the app you already have, behind switches, so they can be rolled out to some users first and turned off at any time.
Will the AI see all our data?
No. AI features follow the same permissions as the user who calls them, and whole groups of users can be excluded from AI entirely.
How do you control the cost?
Usage is tracked per organisation and feature, so you can see what AI costs, set limits or bill it on to your own customers.
Does the AI make decisions on its own?
Not unless you want it to. We design features where AI suggests and explains, and a person confirms.