Module · SMART+
Vision AI
Artificial intelligence built into administrative processes and into platform configuration.
Vision AI brings artificial intelligence capabilities directly into the municipality's day-to-day work. The component combines a co-pilot for configuring the platform with an adaptive validation engine that assists officers in checking case files and running processes. The AI works within the context of existing data and rules, and human control over administrative decisions is preserved.
What changes in checking case files
The officer receives checks and recommendations relevant to the specific case, and the administrator can configure rules and workflows faster. Automation can be introduced gradually, from a simple suggestion through to controlled execution, with confidence thresholds, an audit trail and a return to human validation whenever needed.
What you can do
A co-pilot for configuration and an adaptive validation engine, in the context of the institution's data.
Conversational assistant
Built into the platform's configuration studio.
Rules from natural language
Rules generated from requirements expressed in words, not in code.
Its own context
Uses the entities, fields, types and rules already present in the platform.
Validation before proposal
Syntax and dependencies are checked before a rule is proposed.
Dynamic checklists
Generated according to the case and its context, not identical for every case.
Hybrid execution
Deterministic rules first, then AI for the uncertain situations.
Confidence thresholds
Configurable, with mandatory human validation below the threshold.
Flagging what is missing
Identifying information that is missing, inconsistent or needs attention.
See all features
- Progressive levels of automation: shadow, suggestion, draft and automatic with an audit trail
- Semantic search across existing rules and configurations
- Versioning of models, prompts and rules
- Inputs, model version, confidence score, result and final decision are all retained
- Monitoring of model performance and drift
- The option to run the models on-premise
- Orchestration of several specialised models within the same process
- Integration through microservices, plug-ins and APIs, without separating from the platform's processes
How it works
From a requirement expressed in words to a validated rule applied under human control.
ContextThe existing entities, fields and rules
ProposalA rule or check generated by the AI
ValidationSyntax, dependencies, confidence score
LevelShadow, suggestion, draft or automatic
DecisionThe officer confirms below the configured threshold
AuditInputs, model, score and decision, all retained
What you get as standard
The module is more than an interface. It comes preconfigured in its standard version, ready to use from day one.
- forms
- workflows
- registers
- reference data
- document templates
- business rules
- validations
- notifications
- reports
- dashboards
- access rights
- audit trail
Native integrations
AI results feed directly into the platform's interfaces, rules and workflows.
Part of the same platform
Vision AI is not a chatbot bolted onto the platform. It works within the context of existing data, rules and permissions, and every result stays versioned, explainable and auditable.
How you can start
You do not have to digitise the whole institution at once.
START ONE
Vision AI is switched on over a platform that is already running — it needs data and rules to work on.
START PRO + VISION AI
It is part of the VISION SMART+ premium extension, together with the IoT component.
Would you like to see how Vision AI would work in your municipality?
We show you how a rule is generated from natural language and how the AI assists in checking a real case file.