Blue Cross Blue Shield of Michigan
Manager Conversational AI Design
- Compensation
- Not disclosed
- Location
- Detroit, MI, United States
- Work setup
- Hybrid
- Seniority
- Not specified
- Employment
- Contract
- Experience
- Not specified
- Category
- Design
- Posted
Skills
Must have
- Bachelor’s degree in Communications, Content Strategy, English, Journalism, Marketing, Human-Computer Interaction, User Experience, Information Science, Business, Health Administration, Computer Scien
- Minimum seven (7) years of progressively responsible experience in conversational design, content strategy, user experience writing, digital content, prompt design, applied AI, or a related role requi
- Minimum of one (1) year hands-on experience working with large language models, including prompt development, testing, evaluation, and refinement is required.
- Experience in healthcare communications, health insurance, benefits, clinical content, health plan operations, or another regulated consumer environment preferred.
- Experience designing and evaluating structured prompts, templates, conversational examples, model instructions, test scenarios, quality criteria, grounding requirements, and guardrails; ability to ana
- Experience using research, analytics, experimentation, quality reviews, and user feedback to improve conversational content and model behavior preferred.
- Equivalent or relevant BCBSM experience may be considered to satisfy the requirements outlined above. Demonstrated expertise in digital content, AI-enabled solutions, strategic thinking, and business
- Demonstrated experience creating consumer-facing conversational content for complex digital products, self-service experiences, chatbots, virtual assistants, or AI-enabled solutions.
- Ability to translate complex domain, policy, operational, or technical concepts into clear, accurate language that reflects how consumers naturally communicate and meets approved plain-language, acces
- Strong knowledge of conversational design, content hierarchy, progressive disclosure, intent recognition, clarification strategies, response architecture, tone, and consistent voice across supported d
- Working knowledge of large language models, generative AI, model limitations, grounding, retrieval-augmented generation, evaluation, monitoring, human oversight, and responsible AI practices.
- Ability to identify unsupported responses, ambiguous member intent, high-risk topics, information dependencies, downstream impacts, and situations requiring clarification, escalation, disclaimers, or
- Strong stakeholder leadership and influence skills with the ability to align Product Management, Product Owners, design, engineering, data and AI teams, operations, legal, compliance, privacy, securit
- Strong knowledge of privacy, consumer communication, accessibility, language-access, and regulatory considerations relevant to consumer-facing AI experiences.
- Excellent writing, editing, facilitation, presentation, and storytelling skills, with exceptional attention to clarity, accuracy, consistency, tone, and detail.
Full description
This role is responsible for designing, optimizing, and governing the conversational content and prompt architecture that powers AI-enabled member self-service. The role translates complex healthcare, insurance, benefits, and operational terminology into clear member language and influences model behavior through grounded prompt design, conversational examples, testing, feedback, monitoring, and continuous refinement.This role will also partner with Product Management, Product Owners, design, engineering, data and AI teams, member experience, operations, legal, compliance, privacy, clinical, and business subject-matter experts to ensure conversations are understandable, accurate, trusted, accessible, and aligned to product outcomes and organizational standards.
- Lead conversational content and prompt strategy for AI-enabled member self-service. Define member-friendly wording, conversational tone, response length, response structure, progressive disclosure, follow-up patterns, and a consistent, trusted voice across supported self-service channels.
- Establish plain-language and health-literacy standards, including target reading levels where appropriate, aligned to member needs, product objectives, brand expectations, accessibility, and trust.
- Translate complex healthcare, insurance, benefits, clinical, policy, and operational terminology into clear language that reflects how members naturally ask questions.
- Map conversational intent to the appropriate health plan concepts, processes, actions, escalation paths, and approved information sources, including evidence of coverage and other trusted enterprise content used to answer questions about coverage, cost sharing, enrollment, claims, networks, prior authorization, and related topics.
- Design, test, and refine grounded system prompts, prompt templates, conversational examples, response instructions, clarification prompts, and guardrails that influence model behavior.
- Determine what information should be presented immediately, progressively disclosed, offered through follow-up prompts, or escalated to another channel.
- Establish response boundaries and human handoffs to reduce, detect, and appropriately manage unsupported or inaccurate responses and prevent the experience from providing medical advice, guaranteeing coverage, or communicating beyond approved benefit and policy information.
- Lead evaluation of conversational quality using member research, analytics, test scenarios, model outputs, failure patterns, feedback, and operational insights.
- Define quality measures and acceptance criteria for clarity, accuracy, relevance, tone, health literacy, completeness, grounding, safety, accessibility, consistency, appropriate escalation, and multilingual or language-access needs where applicable.
- Partner with Product Management, Product Owners, experience design, engineering, data and AI teams, operations, and subject-matter experts to translate conversational requirements into implementation-ready content, prompt assets, test cases, model-evaluation criteria, and backlog priorities.