SupportLogin
Talk to us

AI in MyRapidi

AI Agents: Background Workers for Your Data Integration

An AI agent in MyRapidi is a worker that runs in the background. You write its job into a system prompt and a user prompt, set how often it runs, and it gets on with the work without anyone opening a chat.

Rapidi AI agents are background AI workers. Each agent runs on its own schedule and does what its system prompt and user prompt tell it, and a skill can be preloaded so it starts with the expertise its job needs. Albert, the AI chat, is the one chat agent in MyRapidi, and the Error Resolver starts from an error message when you click Ask Albert for help on it.

How an agent runs
On a schedule you set, such as every 30 minutes or every weekday morning. No chat window, and nobody at the keyboard.
Who sets one up
MyRapidi users, under the service. What each user can see and change follows their access rights.
What it leaves behind
A transcript of every run, so you can read what the agent did and why.

Which kinds of agent does MyRapidi have?

Three. They work from the same skills and functions. What differs is how each one starts.

The agents in MyRapidi and how each one starts.
Background agentsRun on a schedule and do what their system prompt and user prompt say. A skill can be preloaded, so the agent starts with exactly the expertise its job needs. Examples: a daily email of new data errors, a report on transfer activity, or a planned change to your integration at a set time.
Albert, the AI chatThe one chat agent. Albert starts with the list of skills available to him and loads the one that suits the task in front of him, so a question about a connection gets the connection skill and a new schedule gets the schedule skill.
Error ResolverStarts from an error message. Click Ask Albert for help on the message, and the AI chat opens on the page with the error troubleshooting skill and the error already loaded, and works through it on its own.

What does a background agent need?

An agent is set up once and then runs until you stop it. These are the parts you give it.

The parts of a background agent.
System promptThe standing instructions: the agent's role, the rules it works to and what a good result looks like.
User promptThe task for each run, such as "Email the new data errors from the last 24 hours to the integration team".
ScheduleHow often it runs. You can write it in plain words, such as "every 30 minutes" or "every Monday at 9am", or as a cron expression.
Preloaded skillOptional. The AI Skill the agent starts with, so it does not have to find the right expertise first.
The user it runs asAn agent acts as a MyRapidi user and can do what that user's access rights allow, and nothing more.
TranscriptsEvery run is saved step by step. You can also start an agent straight away to test it, then read its transcript.

What can you manage under your service?

AI Agents, AI Skills, AI Memories and AI Transcripts are available to MyRapidi users under the service. What each user can see and change depends on their access rights.

The AI features available under a service in MyRapidi.
AI AgentsCreate background agents, set their prompts, schedule and preloaded skill, and start one straight away.
AI SkillsThe knowledge packages Albert and the agents work from: instructions, background information and evaluation criteria.
AI MemoriesThe context the AI keeps, such as preferences, scoped to a user, an account, a service or an agent.
AI TranscriptsThe record of each agent run, so you can check what an agent did.

Can you use the agents from Claude or another AI tool?

Yes, through the Rapidi MCP Server. It gives AI tools that support the Model Context Protocol, such as Claude, the same functions and skills Albert and the agents use, signed in with your own MyRapidi login, and limited to what your access rights allow.

Frequently asked questions

What is an AI agent in MyRapidi?

A background AI worker. You give it a system prompt and a user prompt, set a schedule, and it runs on its own, optionally with a skill preloaded for its job. Every run is saved as a transcript.

How is an agent different from Albert?

Albert is the chat agent: you talk to him, and he loads the skill your task needs from the list available to him. A background agent has no chat. It runs on its schedule and does what its prompts say.

What is the Error Resolver?

A specialized agent that starts from an error message. Click Ask Albert for help on the message, and the AI chat opens with the error troubleshooting skill and the error already loaded, and works through it on its own. It does not run in the background or on a schedule.

Can I build my own agent?

Yes. AI Agents are available under the service in MyRapidi, together with AI Skills, AI Memories and AI Transcripts. Which of them a user can see and change depends on that user's access rights.

Do the agents run my transfers on their own?

The Rapidi transfer engine still runs your transfers as configured. An agent does what its prompts ask, within the access rights of the user it runs as, and every run is saved as a transcript so you can review it.

What does it cost?

The AI features are part of MyRapidi, and Albert is available to everyone who logs in. Rapidi is priced per subscription plan, so ask us about your own setup.

Have a job you want an agent to take on?

Tell us the task, and we will show you how an agent for it is set up in MyRapidi.

Find out if this fits your setup

Tell us which systems you run and we will tell you plainly whether Rapidi is the right tool.

Free, 17 pages

Get the Salesforce and Business Central solution sheet

How the integration is built, what the RapidiConnector secures, the pre-configured templates, and the eight transfers most customers start with.