The Role:

You will work directly with clients, including CXOs, to understand their business, identify the problems worth solving, and shape the AI solutions that solve them. We expect you to push the boundary on how far AI can take the analyst's job: AI assistants, agents and tools in every step,

from preparing for a client conversation to producing, testing and maintaining the requirement set.

You will also design harnesses for the business problems you identify: the prompts, context, data,

tools, checks and human review points that turn a model into a working solution for a specific business process. You will work independently and carry several accounts and projects at the same time.

What you will do

- Work directly with CXOs and business users to learn how their processes run today, where they break, and what outcome is worth paying for. Keep the conversation anchored on business value.
- Use AI assistants to prepare for discovery: research the client's domain, draft question sets, and summarise prior material before each session.
- Generate requirements with AI tools from meeting transcripts, client documents, process artefacts and sample data. Review, correct and own the output; the tool drafts, you are accountable for what goes to engineering.
- Design harnesses for the problems you identify: what the model sees, which tools and data it uses, how output is checked, and where a person reviews. Build working prototypes yourself where that is the fastest way to prove the approach.
- Write requirements that reflect how AI-based solutions behave: variable outputs, confidence thresholds, exception handling, human review points, data and context needs, and failure modes.
- Define acceptance criteria and evaluation sets that let engineering and QA measure whether an AI component is ready for production.
- Map current and target workflows,



showing where the solution automates and where people stay in the loop.
- Work with engineering on feasibility, scope and trade-offs, and raise it early when a requirement does not fit what the solution can do reliably.
- Run UAT with clients, triage feedback, and keep the requirement set current as changes come in.
- Keep reworking your own process. Build reusable prompts, agents and AI workflows, and share them so the whole team works faster.

What we look for

- 3+ years as a business analyst on software delivery projects, including direct client-facing requirements work.
- Daily, hands-on use of AI assistants (Claude, ChatGPT, Copilot or similar) in your own work, with concrete examples of using them to produce requirements, analysis or documentation, and of where you took the process further than most.
- A working understanding of how LLM-based solutions are put together: prompts and context, retrieval over documents, data extraction, agents and tool use, evaluation, and why outputs vary between runs. Enough depth to design a harness for a business problem and to write requirements engineers can build and test.
- Confidence with senior client stakeholders. You can hold a conversation with a CXO about their business and tie a proposed solution to the value it delivers.
- The habit of reading documents, data and AI-generated output critically, and finding the gaps, contradictions and edge cases.
- Independence and flexibility:



you can work across several engagements with competing deadlines and switch as priorities change.
- Bachelor's degree in engineering, computer science, IT, business or a related field.

Positive to have

- Domain experience in commercial real estate, insurance or financial services.
- Experience writing evaluation criteria or test sets for AI features.
- Basic SQL or scripting, enough to inspect client data and build a prototype yourself.
- Experience building automations or agent workflows with AI or low-code tools.

Compensation There is no fixed band, and pay is not tied to years of experience. We set compensation on proven skills, competence and flexibility, as shown during the selection process and in how you handle multiple projects and changing priorities. Strong candidates are paid accordingly.

How we hire

1. A practical exercise: turn a realistic client brief into a requirement set and a harness outline.
2. A discussion of your submission, where you explain and defend your choices.

3. A final conversation with leadership.

AI tools are allowed, and expected, in any exercise. You must be able to explain and defend every part of your work yourself.

Before you apply

This role is demanding. Problems arrive half-defined, priorities move with client needs, and you are accountable for the outcome. There is little handholding. Apply only if you are up to the challenge. [email protected]

If your experience is mostly writing documents from templates, or you have not yet used AI tools seriously in your own work, this role is not the right fit.

How to apply

Send your CV to [email protected] with Business Analyst (AI-Native) in the subject line. Include a short note on one business problem where you used AI to change how the work was done, and what you built to do it.

📌 Business Analyst (AI-Native) (Chennai)
🏢 Perfecter.ai
📍 Chennai