Overview
In this section, we will be taking a look at an example of two different types of personas that can be used.
Key points
[Warning: empty required content area]Fluent Skills is structured around two core personas, each designed for a distinct stage of the product and operations lifecycle: the Investigator and the Builder.We will be going through what these personas are for, and what kinds of actions they are best suited to perform.The Investigator Persona
The Investigator persona is designed for support teams, operations staff, and anyone who needs to query and understand live data, all the while only using natural-language queries for ease of use and understanding.What It Does
Natural language data queries: Ask questions like "How many orders were placed in February?" or "What were the last 5 orders and the items in each?" The system interprets your intent and queries your data accordingly.Self-correcting GraphQL: If a query fails due to a schema mismatch or unexpected structure, the system performs automatic schema introspection and retries with a corrected query. No manual intervention required.Event log inspection: Investigate trade events, order histories, and workflow logs directly through conversation, enabling fast root-cause analysis without switching tools.Example
A support agent asks: "What were the last 5 orders and what items were ordered?" The system attempts the query, encounters a GraphQL schema error, introspects the schema, rebuilds the query correctly, and returns the accurate results — all without any action from the user.Agentic Potential
The Investigator persona is also a foundation for automated monitoring. Agents can be configured to watch for specific trade events, detect failures, classify whether an issue is a data error or a code error, alert the right person, and potentially apply a fix and redeploy automatically. The full diagnose–inspect–fix–deploy cycle can be run within a single session.The Builder Persona
The Builder persona is designed for business analysts, product managers, and developers working through the software development lifecycle. It has the ability to understand what has already been implemented, spec new features, generate the code and deploy it.What It Does
Implementation discovery: Ask the system to explain what features have been built into a given workflow. It reads the source code and workflow structure to produce an accurate breakdown of things like statuses, rule sets, user actions, and custom logic.Collaborative feature design: Describe a new feature in plain language and the system will guide you through the design interactively, asking targeted questions to define behavior, edge cases, and configuration options.Self-critiquing spec generation: Once an initial specification is drafted, a second agent automatically reviews it for gaps and inconsistencies by referencing the full Fluent knowledge base, existing implementation, and your custom rules before presenting the final spec for your approval.Code generation and deployment: After a spec is approved, the system generates the necessary rules and plugins, deploys to the Fluent environment, and uses the event API to test the implementation end-to-end.Feature Dashboard: A built-in dashboard tracks all features through their lifecycle stages: Discovery, Planning, and Deployment. Each feature is managed independently, with its spec accessible and editable at any stage.Example
A business analyst asks the system to add an approval step to the returns workflow. The system analyses the existing workflow, identifies that no user actions are currently defined, and asks a series of clarifying questions: Should the rejection reason be a free-text field or a drop-down? Should unapproved returns auto-reject after a configurable timeout? What timeout values are reasonable?Once the analyst has answered, the system generates a full feature spec, has it reviewed by a second agent, incorporates any suggested improvements, and presents it for approval. The analyst reviews the spec in the Feature Dashboard and approves it when satisfied.Context Awareness
The Builder persona works within the context of your specific implementation. By providing source code (such as back-end logic, compiled jars, and front-end components) alongside reference documentation gives the system a richer understanding of your environment and produces more precise, implementation-aware output. Complex requirements can also be supplied as documents and referenced directly during the design session.What Should be Used When?
The Investigator and Builder personas are complementary. The Investigator gives teams visibility into what is happening in production. The Builder gives teams the tools to design, validate, and ship changes.Together, they represent a full AI-assisted development and operations loop: observe → understand → design → build → test → deploy → observe again.
