Design AI agents from reusable, versioned building blocks
Also listed as: Agentic framework & configurable prompts
What is it?
Xelence 10.0 will provide a framework for agentic AI, meaning AI that carries out a series of steps toward a goal. Instead of hard-coded prompts, an agent's instructions are assembled from reusable parts: who the agent is, how it behaves, what it can do, and the instructions for this use case. Your team writes these in a guided way, drafting, testing, reviewing and rewriting, and keep them in a shared, versioned library.
Why does it matter?
When AI instructions are built into code, each new AI feature starts from scratch, and changing how it behaves needs a developer and a new release.
A common, governed foundation for AI agents, quicker tuning of AI behavior as needs change, and instructions the next AI feature can reuse.
Generative AI Services and Xelence AI Engine work together. Building AI and running it explains which does what.
Getting it into your solution
- Needs an upgrade to the Xelence release
- Needs configuration
- Needs implementation work
Where people stay in control
People write, test and review instructions, and only published versions can be used by an agent. Each environment keeps its own published version, and administrators must review any code in a skill before it can run.
Start thinking about
- Which multi-step tasks in your programs might suit an AI agent?
- Who should be allowed to change agent instructions, and who should approve and publish those changes?
- Which existing guidance, such as letter templates or procedure notes, could become reusable AI instructions?
Terms used on this card
- Agentic AI
- AI that can carry out a sequence of steps toward a goal, such as checking, deciding and acting, within limits people set, rather than only answering one question.
- AI skill
- A defined ability, such as looking up a record or summarizing a document, that AI agents can reuse across workflows.
- AI use case
- A specific business task where AI helps, built so it works alongside your forms, rules and workflows.
- Environment
- A separate copy of your solution used for one stage of work, such as development, testing or production. Production is the live system that staff and constituents use.
- Large language model (LLM)
- The kind of AI model that understands and writes text, used behind AI assistants and agents.
- Prompt
- The instructions that tell an AI feature what to do and how to behave.