AI that works on YOUR website. Not the other way around.
Most AI platforms require you to change your website to fit their system. GovAI's AI Engine integrates with any council platform in days — not months.
Integration methods
Deploy time: Minutes
Embeddable Widget
Add GovAI to any website with a single script tag. No backend changes, no IT project. Your residents get the same intelligent navigation whether you run WordPress, Drupal, Jadu, Firmstep, or a legacy system.
- • Any platform
- • Branded to council
- • No IT project
- • Automatic updates
Deploy time: Days
REST API
Full RESTful API for custom integrations. Connect your CRM, case management, or citizen portal. OpenAI-compatible endpoints so you can plug into existing tooling.
POST /v1/navigate GET /v1/services POST /v1/check-eligibility POST /v1/referral GET /v1/analytics
- • Full docs
- • OpenAI-compatible
- • Webhooks
- • Multi-tenant
Deploy time: Days
WordPress Plugin
Native WordPress plugin with Beaver Builder modules and Gutenberg blocks. Configure from the admin panel. The London Borough of Newham runs on this.
- • Beaver Builder
- • Gutenberg
- • Admin panel
- • Used by Newham
Core AI capabilities
The same engine powers our citizen-facing Resident Navigator and our Staff Assistant for officers.
Multi-Service Matching
Matches across financial, employment, housing, and social care simultaneously — not one service at a time.
Learn moreEligibility Checking
Checks criteria before surfacing recommendations. Residents see only what applies.
Learn more40+ Languages
Automatic language detection. Critical for diverse boroughs and inclusive access.
Learn moreService Triage & Routing
Determines the right department and pushes referrals into your CRM.
Learn moreSafeguarding isn't a feature — it's the foundation.
Sentiment analysis, distress detection, and keyword flagging run on every conversation. Automatic escalation to a human advisor when needed. Full audit logs and PII redaction in logs so you stay compliant.
Learn about the Safeguarding Layer →The AI engine that runs on infrastructure you control.
For security and compliance details see our Compliance page.
Generation model
Qwen3-30B-A3B (RAGAS faithfulness 0.91)
Embedding model
Qwen3-Embedding-8B (MTEB multilingual leader)
Inference
vLLM (OpenAI-compatible API)
Database
PostgreSQL + pgvector
Framework
Python + FastAPI
Fallback
Azure OpenAI UK South / Anthropic via AWS London
Infrastructure
UK-hosted GPU hardware, data never leaves UK