Revenue Cloud Intelligence

Ask Aria 
anything.

Aria turns your products, orders and customers into instant, plain-language answers — no dashboards, no queries. Just ask.

Just ask, in your words

Type a question like you'd ask a colleague. Aria understands the intent and answers in seconds.

Grounded in your data

Every answer comes straight from your own products, orders and customers — never made up.

Answers you can see

Ask for a trend or breakdown and Aria replies with a chart — not just a wall of numbers.

For developers

Build on Aria

The same client the Aria widget itself runs on, packaged for your app. Typed, ESM, and no protocol to reimplement.

1

Install

Point the @revenexx scope at the registry once, then add the package.

# .npmrc
@revenexx:registry=https://npm.pkg.github.com
npm install @revenexx/aria-sdk
2

Ask a question

One call, one answer. Your app supplies the token.

import { createAriaClient } from '@revenexx/aria-sdk';

const aria = createAriaClient({
  getToken: async () => session.accessToken,
});

const answer = await aria.ask({ message: 'Top 5 products last month?' });

console.log(answer.text);

Need tokens as they arrive? aria.stream({ threadId, message }) yields typed events instead. Failures always arrive as one ApiError, carrying the server's own code and message.

The roadmap

The path to autonomous agents

Six steps that turn connected data into intelligence — and intelligence into agents that run the work on their own.

1

Infrastructure

Done
The foundation

Set up the secure, scalable backbone everything else runs on. No intelligence without a solid base.

  • Scalable, secure cloud infrastructure
  • Access control & governance from day one
  • The platform Revenexx Intelligence is built on
Milestones
Cloud environment provisioned
Security & access baseline
CI/CD & monitoring
2

Data and Capabilities

In progress
Connected via MCP — continuously extended

Data flows in through MCP connectors. Sources are added and refined in an ongoing loop — the knowledge base keeps growing and improving over time.

  • MCP-based connectors to every relevant source
  • A looping process — new sources added continuously
  • Clean, structured data ready for AI to use
Milestones
Revenexx Workflows MCP
Revenexx Knowledge Base (Developer) MCP
Revenexx API MCP
Revenexx Analytics MCP
Custom app MCPs & vendor MCPs
3

Aria

In progress
The AI Assistant

Infrastructure and data come together as Aria — an AI assistant that answers questions and surfaces information on demand from the connected data.

  • Natural-language access to all connected data
  • Instant answers & smart search
  • The first face of Revenexx Intelligence
Milestones
Assistant prototype
Natural-language querying over data
Rollout in Revenexx Cockpit & integration in other apps
4

Co-Pilot

Planned
From answers to suggestions

Aria grows into a co-pilot — it doesn't just answer, it proactively suggests next steps based on the data. The human stays in control.

  • Proactive, data-driven suggestions
  • Recommendations tailored to context
  • Human in the loop — you decide, it advises
Milestones
Contextual suggestion engine
Recommendations in the UI
Approve-and-apply workflows
5

Agents

Goal
From suggestions to action

Agents take on tasks and carry them out end-to-end within defined boundaries — turning recommendations into completed work.

  • Task-focused agents that execute, not just advise
  • Defined scope & guardrails per agent
  • Humans supervise, approve & override
Milestones
Single-task agent pilot
Dedicated agent workspace with direct Git access — e.g. build & edit a shop storefront
Guardrail & scoping framework
Supervised end-to-end execution

Autonomous Agents

The final goal
Work that runs itself

Agents set their own sub-goals, reason across systems, and act independently — continuously optimising the business while people set direction and boundaries.

  • Goal-driven agents that plan & execute on their own
  • Multi-agent collaboration across the stack
  • Continuous self-improvement within human-set guardrails
Milestones
Goal-driven planning
Multi-agent orchestration
Self-improvement loop