AI Solution Architect

About the role
You are the squad's business brain and the bridge between business and technology. You look beyond surface-level requests to find the real business problem, redesign the end-to-end process where needed, and work with AI Product Engineers to define and validate the solution. Your role is not just to produce requirements, but to use data, AI and rapid prototyping so engineering capacity is focused on the highest-value problems.

Key Responsibilities

Observe — understand the business and process

• Build a deep understanding of customer journeys, end-to-end processes, operating metrics and frontline pain points. Use data and observation to identify where change is really needed.

• Stay close to business leaders, operations teams and key stakeholders, and maintain an opportunity backlog grounded in real problems and value.

Orient — define the right problem and outcome

• Don't simply document requests. Challenge assumptions, find root causes, and determine whether the answer is a process, policy, people, data or technology change.

• Define the target outcome, success metrics, business rules, edge cases, data needs and acceptance criteria, and co-author the specification with engineers.

Decide — support prioritisation with business value

• Build concise business cases using revenue, cost, customer impact, turnaround time, risk, compliance and effort.

• Challenge low-value requests with evidence and recommend priorities and trade-offs; final portfolio decisions follow company governance.

Act — validate, launch and realise value

• Own business validation / UAT, launch readiness and stakeholder expectations so what is delivered remains aligned to the original business outcome.

• Measure adoption and business results within 7–30 days after go-live, and feed the learning into the next cycle.

Build — use AI and prototyping to validate faster

• Use AI, no-code, low-code or vibe coding to create lightweight automations and rapid prototypes that test an idea before engineering capacity is committed.

• Prototypes are for learning and validation, not a substitute for production engineering. Production solutions are rebuilt by AI Product Engineers under AI, information-security and data-compliance governance.

What This Role Does NOT Do

• Build or own production core-system code, or own API, data-model or technical architecture design.

• Manage developers / QA day-to-day tasks, assignment or sprint execution; the engineering team owns technical delivery.

• Act as an order-taking BA that writes requirements first and hands them to IT; discovery and design are collaborative from the start.

Requirements

Must-Have

• 3–5 years in business analysis, product, digital transformation, process improvement or a similar business-facing role.

• Strong problem solving and end-to-end process design; able to turn ambiguity into clear outcomes, processes and requirements.

• Data-literate and comfortable using dashboards and analysis; basic SQL is a plus.

• Able to work with senior stakeholders and challenge assumptions with evidence while maintaining constructive relationships.

• Hands-on with Jira, Confluence or equivalent project / delivery management tools; able to structure backlogs, track dependencies, manage requirements and coordinate UAT / release readiness.

• Practical AI use beyond writing or summarising; willing to build lightweight tools or prototypes with no-code / low-code / vibe coding.

• Business-value mindset: measures success by outcomes and metrics moved, not documents delivered.

• Mandarin proficiency is required for stakeholder engagement across Hong Kong and Shenzhen.

Nice-to-Have

• Fintech, lending, payments or financial-services experience; familiarity with origination, KYC, credit assessment, drawdown, repayment, monitoring and collections.

• Product discovery, roadmap prioritisation, business-case or measurable process-improvement experience.

• Lark/Feishu or similar workflow automation / no-code platforms.

• Conceptual technical-architecture understanding: can read system diagrams, understand system boundaries and ask intelligent questions (no coding required).