Overview
— ROLE
Content Designer
— TIMELINE
Q1 2026 (3 months)
— TEAM
Product Manager
Product Designer
Product Marketer
I designed an AI-powered Product Strategy Assistant using Gemini Gems and NotebookLM that combined product documentation, customer personas, strategic planning materials, and organizational knowledge into a centralized resource.
The goal was to help stakeholders quickly access accurate information, generate strategic recommendations, and accelerate product decision-making.
My Contributions:
- Audited and selected knowledge sources
- Structured information for AI retrieval
- Designed system instructions and behavioral guidelines
- Defined communication standards
- Tested outputs for usefulness and accuracy
- Iterated on prompt design and knowledge sources
- Supported adoption among stakeholders
Challenge
Product knowledge existed across multiple sources, including help documentation, customer personas, roadmap materials, business strategy documents, and historical product information.
While this information was available, finding and utilizing it often required searching through multiple documents and repositories. General-purpose AI tools like Gemini and Claude could assist with ideation, but lacked the context needed to provide recommendations grounded in Text Request’s products, customers, and business goals.
I wanted to explore whether a specialized AI assistant could provide faster access to institutional knowledge while supporting strategic product work.
Process
Gathering
I began by identifying the sources of truth that contained the most valuable and accurate knowledge about the product, which included each page of our help center, as well as a training video of the Text Request platform. As the writer of all of our help documentation, I had a pretty good idea of the accuracy of this training data and knew what questions to test the NotebookLM with to see if it was pulling data and answering correctly. The main purpose of this NotebookLM was to use it as the basis of the Gem, allowing the Gem to fully understand how Text Request works and what it does.
Once I was content that the NotebookLM was answering my questions correctly and had a fundamental understanding of Text Request, I began adding other documents I (and our Product Manager and Product Marketer) had created. These included:
• User personas
• Our growth strategy for 2026
• Reports on the various industries we were targeting in the upcoming years
• Competitor research reports
Purpose & Goals
I then had to create the Gem and using the NotebookLM documentation was a start, but not enough to fully create the Gem. I defined its Purpose and Goals with the following instructions:
- • Act as a 'Text Request Product Manager and Strategist', serving as an expert resource on the Text Request business texting platform.
- • Utilize deep knowledge of user personas, competitor landscapes, and industry-specific targeting strategies to provide strategic advice.
- • Support the 5-year vision of success by analyzing potential paths for growth, new feature development, and product updates.
- • Assist in the creation of high-quality product documentation, including product briefs, wireframe concepts, and testing protocols.
Behavior & Rules
I defined its Behavior and Rules with the following instructions:
- Product Strategy and Analysis
- • Provide insights based on the Text Request help center knowledge base and business model.
- • Evaluate new feature ideas against the existing product ecosystem and target user needs.
- • Conduct mock competitor analyses to identify market gaps and opportunities.
- Document Creation
- • Generate detailed product briefs that include goals, user stories, and success metrics.
- • Describe wireframe layouts and user flows for new or updated features.
- • Develop comprehensive testing protocols and QA checklists to ensure product quality.
- Communication Style
- • Maintain a professional, analytical, and data-driven tone typical of an experienced Product Manager.
- • Be concise but thorough, ensuring all technical and business requirements are addressed.
- • Use industry-standard terminology (e.g., KPIs, MVP, User Journeys).
- Overall Tone
- • Professional, strategic, and collaborative.
- • Insightful and forward-thinking, focusing on long-term product success.
- • Resourceful, drawing upon internal knowledge bases to provide accurate information.
Knowledge Architecture
Use Cases
Product Planning
Challenge: Evaluating new feature ideas against user needs often required reviewing multiple documents.
Solution: The assistant synthesized user personas, roadmap priorities, and product knowledge into a single evaluation.
Outputs:
- • target users
- • opportunities
- • risks
- • success metrics
Example Document: Assign Conversation/Internal Notes Feature Planning doc
Documentation
Challenge: Create a product brief for a new messaging feature.
Solution: The assistant created a detailed brief that was specific to the needs of our user personas and generated initial images, wireframes, and even prototypes for design to use to start brainstorming.
Outputs:
- • goals
- • user stories
- • KPIs
- • rollout considerations
Competitor Analysis
Challenge: Compare our offering against competitors and identify market opportunities.
Solution: The assistant created detailed reports of what our competitors offered with links to the help center sources it pulled the information from, and included easy to scan tables comparing each competitor to not only each other, but Text Request as well.
Outputs:
- • differentiators
- • gaps
- • recommendations
Outcomes
Results
The assistant dramatically decreased the time it took for all stakeholders to complete various tasks, including drafting product briefs for the Product Manager, conducting competitor reports for the Product Marketer, and building wireframes for the Product Designer.
During roadmap discussions, stakeholders used the assistant to quickly retrieve persona information and competitive context that previously required searching multiple documents. The assistant could synthesize information from user personas, roadmap documentation, and competitive research to create a single, cohesive response that was relevant exclusively to Text Request.
Stakeholders also used the assistant during meetings to quickly answer questions from other departments instead of having to reschedule or say, "I'll get back to you on that." This was especially useful when testing wireframes or design concepts with internal, customer-facing stakeholders from the Success, Sales, and Support teams.
Before & After
Before
To evaluate a feature idea, stakeholders often needed to:
- • Search help documentation
- • Review user personas
- • Consult roadmap materials
- • Reference business strategy documents
- • Synthesize findings manually
After
The Product Strategy Assistant could retrieve and synthesize this information in a single interaction, dramatically reducing time and friction.
Takeaways
AI Is Only As Good As Its Context
The biggest lesson from this project was that AI quality depended far more on knowledge architecture and source selection than prompt engineering alone.
As the knowledge base expanded, maintaining high-quality source material became increasingly important to ensuring reliable outputs.
Information Architecture Matters
Organizing content effectively proved just as important as creating it. The most useful outputs came from connecting product documentation, customer research, and business strategy into a single knowledge ecosystem.