Business analysts write the rules other people follow: process maps, requirements, acceptance criteria. In AI operations, those same skills shape what an agent may do, what it must hold for a person and which real cases prove it works. This CV section translates your analysis work into those terms, with a template and a filled sample from retail banking.
Where does a business analyst fit around governed agents?
Most review roles sit downstream: a person checks what an agent proposes. A business analyst often sits upstream, deciding what the agent should be allowed to propose. That means defining its role, the sources it may use, the limits per capability and the cases that test whether it behaves. In Pryme's terms, it is closer to building and certifying an agent than to reviewing its queue.
So your CV section should point there. A bank hiring for agent operations will value someone who can turn a policy into precise limits and a set of test cases. The guide on how a business analyst should show AI agent skills on a CV covers the positioning in more depth.
How should you write the bullets?
Translate each piece of analysis work into its agent-era equivalent:
- A process map becomes a map of which steps an agent drafts and which a person approves.
- Acceptance criteria become the standard an agent's answers must meet.
- User acceptance testing becomes replaying the agent against real past cases.
- A requirements traceability matrix becomes the link between a policy rule and the limit that enforces it.
Write bullets that show you have done the left-hand version well, and, where true, the right-hand version too.
The CV section template
Agent design and controls analysis
Business analyst in [sector] ready to define agent [roles, limits and test cases] for [type of process].
- Mapped [process] end to end; identified [number] steps suitable for [automation / an agent] and [number] that must stay with a person because [reason]
- Wrote [acceptance criteria / requirements] for [system or change]; [how they were used or what they prevented]
- Built [test packs / UAT scripts] from [real historical cases]; found [type of defect] before release
- [If applicable:] Defined [limits / escalation rules] for [an agent or automated process]; [result]
Methods: [process mapping], [requirements], [testing from real cases], [control mapping]
Certificate: [name], issued [month, year], expires [month, year], verify: [link]
A filled sample
A sample for an invented analyst, Dev, at an invented retail bank:
Agent design and controls analysis
Business analyst in retail banking ready to define agent roles, limits and test cases for servicing and payments processes.
- Mapped the card dispute process end to end; identified the steps an agent could draft (case summary, customer letter) and those that must stay with a person (any refund over the team's limit, any case with a fraud marker)
- Wrote requirements for a new complaints workflow, with a rule linking each regulatory deadline to an automatic reminder; no deadline missed in the first quarter
- Built UAT packs from 200 anonymised historical disputes rather than invented scenarios; found three routing defects that synthetic tests had missed
- During an agent pilot, drafted the escalation rules for the servicing agent and reviewed its held decisions weekly with the operations lead
Methods: process mapping; requirements and acceptance criteria; testing from real cases; mapping controls to policy
Certificate: Certified Pryme Operator, issued [month, year], expires [month, year], verify: [link]
The third bullet is the one that matters most. Testing on real past cases rather than invented ones is how governed agents earn their authority: Attest replays an agent against real cases from the customer's own operation and scores every answer. Dev did the same thing with a workflow, which an interviewer will recognise at once.
What should a business analyst avoid claiming?
Avoid "AI strategy" language unless you set strategy. Avoid listing every methodology you know. And do not claim to have built an agent if you only wrote requirements for one; say what you did. In a bank, accuracy about your own role is itself a signal.
If the advert is for a reviewer rather than a designer, adjust. Lead with a bullet about reviewing decisions, and move design work lower. The guide to rewriting a business analyst CV for AI operations jobs shows both versions.
Where the Academy fits
For most business analysts, the Pryme Intelligence Academy route runs through Operator to Builder. Operator teaches how a governed agent decides and how review works; Builder, about 24 hours plus a capstone agent certified in Attest to depth 2 or higher, is the track for people who design, train and certify agents, which is where an analyst's skills point.
Questions people ask
Should I start with Builder rather than Operator?
The tracks run in order, each the key to the next, so Operator comes first. It also gives you the reviewer's view, which makes your design work better.
How does this differ from a compliance or KYC analyst section?
Compliance and KYC analysts usually review decisions against rules; a business analyst designs the process and its limits. Compare the CV section for a compliance analyst and the CV section for a KYC analyst.
Do I need to code?
Not for most agent design work. Clear writing, precise limits and good test cases matter more. The AI skills for job seekers guide lists what employers ask you to prove.