AI for Problem Solving: Personas, Challenges & Solutions

Personas, Problems, Pains, & Gains Emphasizing People over Technologies © 2025 PRODUCTSIDE, © 2025 PRODUCTSIDE 1 LLC  What Problem are we Solving? For Whom are we Solving it? © 2025 Productside...

Personas, Problems, Pains, & Gains Emphasizing People over Technologies © 2025 PRODUCTSIDE, © 2025 PRODUCTSIDE 1 LLC  What Problem are we Solving? For Whom are we Solving it? © 2025 Productside 2  ComfortAIble Scenario Simplifying smart home systems for users. Enhancing tenant-domicile automation experiences. Reducing management complexity & maintenance costs. Tackling tactical challenges with strategic AI solutions. Scaling and improving smart home usability. © 2025 Productside 3  Personas define the “who” A Persona is a composite profile of a group of persons. It acts as a reference point, providing context for who our users and buyers are. © 2025 PRODUCTSIDE, 4 LLC  User Personas Buyer Personas Focus is on Focus is on Usage Behavior Buying Motivation Pains & Goals Budgetary Constraints UX Preferences Business Goals & Pain Points Behavioral Insights Value Proposition & Benefits Empathy & Understanding External Factors of Impact Qualitative Feedback Quantitative Data © 2025 PRODUCTSIDE, © 2025 PRODUCTSIDE 5 LLC  Personas Problems Agenda Prompt Engineering Jobs-to-be-Done Stay Calm Feel their Stake Out Stand on Unpack Build on Socialize and Focus Pains, an AI the Assumptio Tiny Bets an on the Learn their Playing Shoulders ns & for Big Outcomes Problem Gains Field of Giants Unknowns Wins Blueprint Space © 2025 Productside 6  How might we use Generative AI to Explore our Problem & Personas? © 2025 Productside 7  Leveraging AI Generative AI ML GAI is a creative branch of AI: DL NLP Made up of multiple AI: GA Deep learning, neural networks, & I NLP Makes new things: text, images, music, etc. Makes things up: Hallucinations, confabulations, errors © 2025 Productside 8  Understand the AI Chatbot Dialog There are 3 roles when to consider 1. User: You, the person making requests 2. Assistant: the AI interface providing responses 3. System: governs the Assistant’s behavior © 2025 Productside 9  Prompt Engineering Essentials 1. Models can’t read minds. 1. Be clear with your request. 2. Brief replies lack depth. 2. Request an expert-level role if needed. 3. Format won’t fix itself. 3. Show the format you prefer. 4. Jumbled input = jumbled 4. Use delimiters & data for request clarity. output. 5. Provide clear, step-by-step instructions. 5. Steps need to be explicit. 6. Give examples to improve understanding. 6. No examples = no clarity. 7. Quality input creates quality output. 7. Garbage in, garbage out. © 2025 Productside 10  How do you prompt as a team with different sessions? DATA SCIENCE END USERS CHANNEL OPERATIONS SALES CONTENT STRATEGIST FINANCE UX DESIGNER SUPPORT UI/UX EXECUTIVE CUSTOMERS DEVELOPER LEADERSHIP BUSINESS ANALYST MARKETING PRODUCT PRODUCT COMM BUYERS MANAGER OWNER USERS PRODUCT DESIGNER ENGINEERING PRODUCT MARKETING LEGAL REV USABILITY DEV TEAM OPS TESTER DEV OPS © 2025 Productside  Set the Session Context Let the session know: Who you are Who your audience is What your goals are What task you need completed What constraints you face What formats you want More Context == Less Hallucinations © 2025 Productside Sources: Microsoft Support and OpenAI via Noureddine Haouari  Art of the Session Starter Prompt Let's start a new session to explore and as a strategic product management expert to address a specific persona's needs based on What? the context below: PERSONA: [Describe your selected persona, including their name Sets roles, context, and tone. & key characteristics.] PERSONA ROLE: [Detail the persona’s primary function or job responsibility.] Why? CHALLENGES: [List the key obstacles or inefficiencies faced by Keeps the session aligned on outcomes. the persona.] OUTCOMES: [Define the desired results or goals for the persona.] Reusable & Shareable: CONSIDERATIONS: [Highlight any important constraints, Keeps team sessions consistent. priorities, or limitations] TONE: [Specify the tone/style for AI responses, e.g., clear, engaging, professional.] Focus on Problems: OUTPUTS: Write responses in a format that is concise, Start with problems, not solutions. actionable, and easy to share (e.g., bullet points, slide-ready text, or step-by-step instructions). As the AI Assistant for this session, do you have any questions or need clarification before starting? If not, provide three concise bullet points summarizing your understanding of the session context. © 2025 Productside 13  ACTIVITY Pick-a-Persona GOAL Choose a problem statement and customer persona. This persona and their corresponding outcomes will guide your exploration of potential solutions in future exercises. TOOLS Mural with Case Study Document TIME 10 minutes © 2025 Productside 14  ACTIVITY Gen AI Session Start a new session using the Gen AI Starter chatbot of your choice, including ChatGPT, Claude, Copilot, Gemini, etc. GOAL Start a Gen AI session that includes structured data and Use the session starter, Personas Details external resources that we'll use for the class exercises YAML, and 1 external data source to kick that follow. off a chatbot session we’ll use for the rest of our class exercises. TOOLS Mural & Gen AI Chatbot of your choice TIME 15 minutes © 2025 Productside 15  How do we minimize hallucinations, confabulations, and errors? © 2025 Productside 17  Session Context Windows 1. Setting initial context improves subsequent responses. 2. Context window slides, affecting response accuracy. 3. Forgets earlier context based on token count. 4. Token count incremented by prompts + responses. 5. Manage context by evoking summarization of key points. © 2025 Productside 18  Adding Outside Examples Smarts (picking the right tool for the job) AI-Assisted Search Copilot or Perplexity for near-real- time, relevant data. Documents (e.g., PDF, PNG, etc.) Upload PDFs or Images for domain- specific knowledge and insights. Outcome: Improves focus on relevant, accurate information. Why It Matters: Ensures AI works with current, precise context. © 2025 PRODUCTSIDE, © 2025 Productside 19 LLC  Leveraging Structured Data Think Beyond Markdown: Use formats like YAML, JSON, CSV. Why Structured Data? Reduces ambiguity, ensuring clearer, relevant responses. Shareable & Reusable: Helps teams maintain aligned context across sessions. © 2025 Productside 20  How might we use common product management templates & canvases to get our job done? © 2025 Productside 21  Using Prompts by Reference Examples: Explore ideas using a Teresa Torres Opportunity Solution Tree. Map customer pains, gains, and JTBD with Osterwalder’s Customer Profile. Visualize key interactions with a Jeff Patton User Journey Map. Define product positioning using Geoffrey Moore’s framework. Frame testable hypotheses with Tim Herbig’s UX Hypothesis Template. © 2025 Productside 22  Using Prompt by Template Clarity and Consistency: Outputs follow a clear, repeatable structure. Collaboration Ready: Align teams via a shared framework. Reusable and Scalable: Can be adapted across projects, saving time & ensuring consistent quality. © 2025 Productside 23  Framing the Problem Articulating Underserved Needs © 2025 PRODUCTSIDE, © 2025 PRODUCTSIDE 24 LLC  What is your Problem? © 2025 Productside 25  Frame the Problem Use a problem statement to communicate what you have learned in the problem space: Focus efforts from the outside in Tell the customer’s story to stakeholders Help align & prioritize engineering efforts © 2025 Productside 26  Problem Frame Example I am a busy airport security officer at a large US airport who is trained in security procedures. Trying to ensure efficient screening processes while maintaining safety But passengers often arrive at the security checkpoint unprepared for the screening process Because they lack awareness, understanding, or adherence to security procedures Which makes me feel frustrated and overwhelmed. It’s a bad situation for everyone. Optimal Product Management | Unit 04 | Lesson 02 | Solutions Hypothesis | 27 © 2025 Productside  How Might We Prompt to … What are the persona’s key pain points? What outcomes is the persona trying to achieve? What barriers are preventing the persona from succeeding? How does the persona feel about the problem? © 2025 Productside 28  ACTIVITY Personas: Pains, In this exercise, you’ll identify the pains, Gains, & JTBD gains, and jobs-to-be-done for both personas. A provided prompt can be used optionally GOAL to guide your brainstorming, but feel free Identify and understand the pains, gains, and jobs-to-be- to generate your own ideas as well. done for both personas. TOOLS Case Study Document TIME 10 minutes © 2025 Productside 29  ACTIVITY Frame the Revisit the Pains, Gains, and Jobs to be Problem done for your user and buyer personas to surface their top problems. GOAL Based on your persona and target Document the problem you have outcomes, articulate a problem worth investigated and develop a strong problem solving that will support Product statement. Outcomes. TOOLS Whiteboarding tool TIME 10 minutes © 2025 Productside 30  Rules of the Road DO NOT DO Start prompting without a plan Establish context in new sessions Prompt using solution speak Explore problems before Ask for features with no solutions context Iteratively evolve the session Craft a tactical plan from scratch Generate conversation starters for humans Ignore your organization’s AI & Privacy Policies Engage in ethical use of AI © 2025 Productside 31