Get a reply from this employer
Your resume and cover letter, tailored to exactly what they’re hiring for.
Badger SA | ARC AI is seeking a Business Analyst to translate unclear briefs into structured, testable requirements for AI-enabled systems. You will sit at the center of product delivery, aligning business intent with feasible technical solutions while managing stakeholder expectations.
You will write PRDs, map processes, and define acceptance criteria, edge cases, and failure modes. Experience with data-driven validation and agile delivery is essential to succeed in a fast-moving, regulated AI
Badger SA | ARC AI | On-site | George, Western Cape
Shape ambiguity into products engineers can build.
At ARC AI, we build AI agents that operate inside real businesses, handle real interactions, and deliver measurable outcomes in a regulated insurance environment. This is not theory. We ship.
We are looking for a Business Analyst who can take unclear, high-level thinking and turn it into structured, testable requirements for systems whose outputs are not always predictable. That combination demands something most BA roles do not: comfort with uncertainty at the specification level.
You will sit at the centre of product delivery, where business intent becomes working AI systems.
Your stakeholders are business owners who know the outcome they want but cannot always articulate what the system should do to achieve it. Your job is to manage that ambiguity upstream: ask the right questions, challenge assumptions, and consistently distinguish between what a stakeholder asked for and what they actually need the agent to do.
That means going beyond surface-level discovery, understanding how processes really work (not just what stakeholders say happens), and defining what good looks like in a world where AI outputs vary.
What a Typical Day Looks Like
This is not a documentation role. It is a thinking role.
You are comfortable operating without a script. You take a vague brief, structure it, and produce something engineers can build from. You actually prefer working this way.
You will need to
You will stand out if you
Most importantly: you think in systems, not tasks. And you treat failure modes as requirements, not edge cases to figure out later.
The interview process includes a practical task with two components.
First, you will be given a vague executive brief and asked to produce a structured discovery plan: the questions you would ask, the workshops you would run, and the first-draft requirement categories you would define.
Second, you will review a sample AI product specification and identify what is missing, particularly around failure modes, constraint specifications, and requirements that describe features rather than behaviour.
There is no single correct answer to either component. We are assessing clarity of thought, comfort with ambiguity, and how you think about risk.
Skilled
Permanent
Market Related
No
George