Risk Services - AI Workflow & Automation Associate 2027 Intake (1-year contract)

What matters most

• Turn process pain points into AI use cases

  The role works with business stakeholders to map processes, identify improvement opportunities, and translate requirements into practical use cases and implementation plans.

• Design workable end-to-end automation

  A core duty is designing AI workflows with low-code and no-code tools, including ChatGPT Enterprise, Microsoft Copilot Studio and Power Platform, to improve efficiency, quality and user experience.

• Move solutions from proof of concept toward use

  The associate supports proofs of concept, pilots and scalable solutions, collaborates with technical teams on integration where needed, and helps with testing, documentation, stakeholder communications and change management.

• Demonstrate measurable value

  The posting calls for analysing business data and operational insights to measure solution effectiveness and value delivered.

• Build in responsible AI and controls

  The role must account for generative AI limitations and risks while applying responsible AI principles, data governance requirements and operational controls.

Likely challenges to think through

• Scoping an ambiguous request

  Likely preparation challenge, not a stated employer problem: stakeholders may describe a pain point without clear requirements. Prepare to explain how you would map the current process, clarify the desired outcome and select a bounded AI use case rather than automate too broadly.

• Balancing a quick pilot with a usable workflow

  Likely preparation challenge: a low-code proof of concept can be faster to build than a solution that fits existing systems and survives testing. Prepare an example approach that distinguishes pilot assumptions from integration needs, testing and documentation.

• Balancing automation benefits with AI risk

  Likely preparation challenge: an AI workflow may save effort while producing unreliable outputs or mishandling data. Prepare to describe how you would test output quality, identify data-governance constraints and decide where human review or other controls are appropriate.

Topics to prepare

• Process improvement and AI use cases

  The posting calls for understanding stakeholder pain points, mapping processes and translating requirements into AI use cases. For revision, take a hypothetical manual review process and explain which steps could benefit from automation, which need human judgment, and how you would judge the proposed change.

• Agentic workflows and generative AI

  The role focuses on agentic workflows and seeks familiarity with prompt engineering, AI agents, RAG and document intelligence. As a revision exercise, sketch a document-question workflow and explain what prompts, retrieved information and human checks would contribute. This is a practice scenario, not a specified project.

• Integration and data fundamentals

  The posting seeks basic knowledge of APIs, databases, data models and system integrations. Revise by drawing how an AI workflow could receive a request, read an existing record and return a result; identify the data fields and where an integration could fail.

• Responsible AI and operational controls

  The role calls for understanding generative AI limitations and risks and applying responsible AI, data governance and operational controls. For revision, review a hypothetical AI-generated answer for unsupported claims, inappropriate data use and missing human review, then propose controls.

• Measuring value from pilots

  The posting mentions proofs of concept, pilots and analysis of operational data to measure effectiveness and value. Practice choosing a baseline and measures for efficiency, quality and user experience, then explain what evidence would justify continuing or revising a pilot.

Skills to revise

• Elicit and map a workflow

  The role requires stakeholder communication, process mapping and working through ambiguity. Practice asking clarifying questions about a hypothetical process, then produce a current-state map, a pain-point list and a proposed AI-assisted version with assumptions marked.

• Design a low-code workflow

  The posting names ChatGPT Enterprise, Microsoft Copilot Studio and Power Platform as examples for end-to-end workflow design. Practice outlining a small proof of concept in one of these platforms if available, or diagram its inputs, steps, decision points and human handoffs without claiming platform experience.

• Test prompts and retrieved answers

  Familiarity with prompt engineering and RAG is sought, and testing is part of project delivery. As practice, write a prompt for answering questions from a short document, create a few answerable and unanswerable test questions, and check whether responses stay grounded in the text.

• Specify an integration handoff

  The role involves collaborating with technical teams on integration where required. Practice writing a brief handoff that lists the data needed from an existing system, expected outputs, access assumptions, error cases and questions for the technical team.

• Communicate pilot findings

  The posting calls for operational analysis, documentation, stakeholder communications and change management. Practice turning hypothetical pilot results into a short update covering the baseline, observed outcomes, limitations, user feedback and a proposed next step.

• Select a project walkthrough

  Prepare a concise account of an actual internship, coursework, research, hackathon or personal project involving AI, automation or analytics, if available. Show your contribution, decisions, testing and observed results; distinguish measured outcomes from estimates.

• Map a business process

  For a hypothetical process, draw the current steps, stakeholders, pain points and possible AI use cases. Add a proposed future-state map and explain how it could improve efficiency, quality or user experience.

• Sketch an end-to-end workflow

  Prepare a one-page design for a proof of concept using a low-code or no-code platform you know. Identify inputs, AI or agent steps, human decisions, outputs and what would need to change for a pilot.

• Explain the integration points

  Annotate the design with the data it needs, a simple data model and any potential API, database or existing-system connection. Mark assumptions rather than implying access to a specific client system.

• Plan evaluation and safeguards

  Draft a small test plan with baseline and pilot measures, sample test cases, and checks for generative AI errors, data governance and operational controls. Explain how results would inform a decision to refine or expand the pilot.

• Prepare a stakeholder update

  Rehearse a short explanation of the use case for a nontechnical stakeholder: the problem, proposed workflow, expected value, uncertainties and decisions needed. Include how you would document feedback and support change.

• Write questions for the employer

  Bring questions about how this Risk Services team selects AI use cases, collaborates with technical teams, measures pilot value and applies responsible AI controls. These are questions to ask, not assumptions about its processes.

• Check application materials

  Have a resume and academic transcript or degree audit ready as PDFs. Indicate this role as your first choice if it is, and add a second choice only where applicable, as the posting requests.