AI Workflow & Agent Library
Examples of the workflows and agents FPA designs for SMEs. Each one does the information work and stops where a person needs to decide. Most can be built bespoke or configured from a reusable FPA agent.
How to read this library: these are example workflows, not a catalogue of finished products. Every agent is scoped, tested and governed for your business before it goes live, and the library grows as we build.
| Category | Example workflow | The agent does | A person decides |
|---|---|---|---|
| HR | Policy and HR questions from staff | Answers from your handbook and policies, with sources | Exceptions and anything sensitive |
| Recruitment | Candidate pipeline admin | Organises applications against the criteria you set, schedules interviews, drafts replies | Who to interview and hire |
| Sales / lead | New enquiry follow-up | Researches the company, checks CRM history, drafts a reply and follow-up | What is sent and promised |
| Customer enquiry | Repeat customer questions | Answers common questions from approved content; hands off the rest | Complaints, refunds and exceptions |
| Finance reconciliation | Daily or monthly matching | Matches transactions, flags breaks and prepares the exceptions list | Adjustments and write-offs |
| Management reporting | Weekly or monthly pack | Pulls numbers from source systems and drafts the commentary | Final numbers and narrative |
| Procurement | Quotes and purchase requests | Collects and compares quotes against requirements | Supplier choice and approval |
| Knowledge | Internal Q&A | Finds answers across documents, SOPs and past work | Updates to the source of truth |
| Compliance | Regulatory checks | Checks documents and processes against a defined checklist | Interpretation and sign-off |
| Operations | Scheduling and coordination | Prepares rosters, reminders and status updates | Final schedule and changes |
Every agent in the library is built the same way
A written charter and narrow scope, approved information sources, defined tools and permissions, human approval for consequential actions, logged activity, testing before deployment, and measurement against business outcomes. Read more on our design principles and AI governance.
Don't see your workflow?
Most agents start as a conversation about one problem. Tell us yours.