Have you ever been passionate about a product idea, but unsure how to build a compelling, evidence-backed business case? You’re not alone. As product managers and entrepreneurs, we often grapple with the uncertainty of whether our ideas will resonate with our audience or truly deliver value.
The Backstory: My Jira Plugin “At-A-Glance”
Many of you might not know that, I am also a product manager of At-A-Glance, a Jira plugin available on Atlassian’s marketplace. https://marketplace.atlassian.com/apps/1217549/at-a-glance-telemetry-dashboard-for-jira?tab=overview&hosting=cloud At-A-Glance stitches together data from various tools—like Jira, continuous integration tools such as Bitbucket and GitHub, build automation tools like Travis CI and Jenkins, and quality analysis from SonarQube—to create a comprehensive story from a team’s initial development to customer delivery. The dashboards display real time data from all these systems and enable product teams to visualize impediments in real time
I recently envisioned an exciting new feature: adding AI-driven observability to our plugin. My hypothesis is:
- Increase team productivity by 35%.
- Boost software downloads by 20%.
The challenge? I needed solid, scientific, and market validation for my hypothesis into a convincing business case that I can use to start having a conversation with Customers and Stakeholders.
While most Product people use AI for writing User Stories and analyzing usage data, I wanted to see if I can use AI for helping me with creating an initial business strategy
So I started experimenting…
Experiment 1: Initial Market Research using Claude and ChatGPT
My first step was to use Claude prompt and let Claude. While ChatGPT is often my go-to AI, this time, I experimented with another AI tool called Claude. Given Claude’s reputation for stronger reasoning and more human-like outputs, I initially expected it to deliver crisp insights.
Claude did a fair job, providing public data, pros and cons, and high-level market considerations. But its insights were too generic—too “safe.” I needed more depth and precision.
Switching gears, I tested the same prompt on ChatGPT. Surprisingly, ChatGPT’s response was more thorough, detailing:
- Customer segmentation
- Clear market sizing
- Customer needs and preferences
Yet, I was still not fully satisfied. While ChatGPT provided broader market clarity, it still lacked deep, scientific insights.
The Prompt that I used for both Claude and ChatGPT was
Perform comprehensive market research based on the given observability with AI Jira Plugin. Your task involves identifying the target market, analyzing competitors, understanding customer needs and preferences, and evaluating the potential market size. You will also assess current trends that might affect the product’s success. Your findings should inform product development, positioning, competitive analysis, customer segmentation and marketing strategies. Compile your research into a detailed report, including data analysis, insights, and actionable recommendations to ensure the product’s competitive edge in the market
First I wanted Claude to focus on what the scientific community thinks about my feature, so I decided to collect this information myself
Experiment 2:
a. Scientific Validation Using “ELICIT”
To truly validate my idea scientifically, I turned to a specialized AI research tool called ELICIT. ELICIT uniquely scans scholarly literature, summarizing key findings from relevant scientific papers.
Using Elicit.com I received a summary based on ten research papers around AI-driven observability. Even with the free version’s limitations, the summaries were incredibly valuable. I downloaded each paper cited, diving deeper into their content, gaining a clearer understanding of the scientific community’s view on AI-driven observability. https://elicit.com/review/ee907df3-5b1f-4abd-8a57-9c21b9cc2bf2 is the information that Elicit found.
The prompt I used was
Can you share techniques for observability tools using AI and how they have effectively solved customer needs
This step was crucial. Now, my business case wasn’t just market-informed—it was scientifically validated.
b. Market Research with “Perplexity”
Next, to complement scientific validation with market intelligence, I used Perplexity, another AI-powered research tool.
I use the prompt “Help me with the latest market research on Observabiliity for AI tools “ on Perplexity.ai
Perplexity provided detailed market research, complete with source citations. This allowed me to download and organize eight insightful market sources neatly into a dedicated directory. I now had thorough market and scientific insights ready for synthesis.
c. Synthesizing Insights into a Professional Business Case
I took all the research—scientific papers, market data and any videos or links that I found—and uploaded them into Claude. I used Claude 3.7 Sonnet Reasoning Gen. AI tool that has deep reasoning capabilities.
Claude uses the concept of Projects and I uploaded all the files into the project knowledge
I also provided specific instructions at the project level “You are an experienced and strategic Product Manager working with customers and stakeholders. Think step by step, break down large tasks and ask questions as necessary “
I then used the same prompt as in Experiment 1
Perform comprehensive market research based on the given observability with AI Jira Plugin. Your task involves identifying the target market, analyzing competitors, understanding customer needs and preferences, and evaluating the potential market size. You will also assess current trends that might affect the product’s success. Your findings should inform product development, positioning, competitive analysis, customer segmentation and marketing strategies. Compile your research into a detailed report, including data analysis, insights, and actionable recommendations to ensure the product’s competitive edge in the market
Claude didn’t disappoint. It provided me with (check out below):
- A powerful executive summary synthesizing all insights.
- Detailed customer segmentation and identification of customer needs.
- A clear analysis of direct competitors and differentiation strategies.
- A precise SWOT analysis comparing my tool (with the new feature) against competitors.
- Recommendations for product positioning, pricing structure, and even a foundational go-to-market strategy.
d. Communicating Clearly with Stakeholders (McKinsey-Style Presentation)
Finally, a business case means nothing if it can’t be communicated clearly to stakeholders.
I instructed Claude to transform the detailed business case into a visually appealing, professional presentation—specifically, adopting the structured, persuasive McKinsey style. Claude generated the presentation content, which I seamlessly imported into PowerPoint and Google Slides.
My prompt is “Create a professional presentation from the report uploaded using React artifacts. Please show your thinking using <thinking> tags, then create the presentation. Make it McKinsey-style with a clear story, headlines, icons, visuals and cues.”
I can use this professional presentation to start having stakeholder conversations and setting the stage for successful internal buy-in and external market validation conversations.
What’s Next: Validation and Beyond
With a compelling AI-powered business case now in hand, my next step is to rigorously validate the idea with real customers. I’ll share more about the validation process—again leveraging AI—in my upcoming videos and blogs.
Key Takeaways & Next Steps for You
Here’s how you can replicate this powerful process for your own product ideas:
- Clearly articulate your hypothesis and goals.
- Leverage multiple AI tools (Claude, ChatGPT, ELLICIT, Perplexity) to collect market and scientific insights.
- Synthesize these insights into a clear, compelling business case.
- Translate this business case into an impactful presentation to communicate effectively with stakeholders.
- Move forward confidently into customer validation and implementation.
Try This Yourself
Want to try this method for yourself? Use the ideas and the prompts to research your feature
If you found this guide insightful, please like, comment your thoughts or questions, and subscribe to my newsletter. I’m always eager to learn from your feedback and help you innovate effectively.
Let’s keep innovating, validating, and building impactful products—together.
Until next time,
