Since December, LinkedIn has been filled with positive reviews and prompt templates  about the generative AI tool, ChatGPT almost to the point that most of us are inundated with all the information. There are individuals who go to the extent of claiming that their jobs would be taken over by AI and I wanted to evaluate its effectiveness for Product Management as a function. Some of the responses were impressive, while others were not satisfactory.

After hours of exploration, I realized GPT-3 is not without its flaws. One of the most significant challenges with this model is its tendency to "hallucinate." This means that it may provide nonsensical or false responses depending on the prompt given to it. Another problem with GPT-3 is that it has a knowledge cutoff, which means that it relies on information available up to 2021. But with ChatGPT 4 on the horizon, I anticipate that the quality of responses will drastically improve and it can significantly transform the discipline of Product Management.

In this essay, we will explore the various ways in which product managers can leverage ChatGPT as a powerful tool with example prompts.

• 1. Creating product descriptions and marketing copy:  Most of us don’t have copywriters and end up writing copy on our own, which can be pretty time consuming. ChatGPT gives you the right head start. You can generate email templates, error and success message copies etc. Copy.ai  is another tool you can leverage specially for 1-liners or small descriptions in a tone that suits your brand.

‍

Example prompt: Write a short email to new users who join a weight loss program like Noom to excite them about the program.

• 2. Market sizing: I usually find this exercise quite tedious so having a helping hand is a boon. I wish it would gather the data and do the necessary calculations to give me a final answer. I will try that with Bing but for now, this should do.

Example prompt: Assuming TAM is defined as Total addressable market, how would you calculate the TAM for a weight loss product in India as a product manager?

• 3. Generate ideas/ Product ideation: You can take the help of ChatGPT to generate creative solutions to problems or ideate ways to expand your services/ build adjacent products. To avoid getting superficial answers from ChatGPT, it's essential to engage in brainstorming and critical thinking before using the tool. This will help ensure that the problem is thoroughly understood, enabling the creation of a detailed prompt that can yield insightful answers. While ChatGPT can be an invaluable resource for product managers, it should not be relied on exclusively without first taking the time to carefully consider the problem at hand.

Example prompt: What are some ways in which a ridesharing company like Ola can increase the number of rides booked per user in a month ?

• 4. Write a press release : Writing a PR helps one to focus on the final outcome with the customer truly at the center. Using ChatGPT you can get a headstart and boost your productivity. Make sure you add enough context in terms of the product or feature and any specific details or format to get precise results.

Example prompt: Assume you are building a digital weight loss product with a human coach. You are launching a new feature to connect users' fitness devices to the app. Using “Amazon's working backwards model”, write a press release for this feature. Add metrics and give it a catchy title.

• 5. Creating a survey: It can be an NPS survey or a way to get new product answers, using ChatGPT you can get a head start for the kind of questions you can add to your survey.  Personally I wasn’t happy with the answer that I got, but I could have improved the answer by refining the prompt further.

Example prompt: Using guidelines mentioned in Teressa Torres's book "Continuous discovery habits" formulate a survey to capture feedback for a fintech product that helps users buy stocks and mutual funds.

• 6. Competitor evaluation: When you have a new product idea or initiative, this can come in very handy in identifying the direct and indirect competitors.

Example prompt:  I am developing a SaaS product for the Indian market that helps restaurants directly sell to their customers through branded websites and mobile apps without going through a marketplace. We help users build a personalized website with their branding, customize the menu, launch a white labeled mobile app and run loyalty deals. Our unique value proposition is that we auto-suggest seasonal push notification and email campaigns. Can you write a competitive intelligence report comparing us to other players ?

• 7. Comparing two options, for example buy vs build decisions: These decisions can be complex, ideally I’d want a list of alternatives that I can consider with their pros and cons. As you can see the response is pretty generic.

Example prompt: Assume you are a product manager for a product that requires non-tech savvy admins to upload images, write text about places and events around them and this information appears on the app. There should be a preview feature as well to see how the data will look on the app. Should I build this custom content management system or use existing solutions in the market ?

• 8. Explain it to a 5-year old: I find ChatGPT to be incredibly helpful in explaining complex concepts. The explanation is always very easy to understand and saves me a ton of time by not having to go through several links.

Example prompt: What is a decentralized front end ?

Some other use-cases are as follows:

• - Generating user personas and their pain points for a product type.
• - Creating a customer journey map for a product
• - Help with prioritization of features based on a framework like KANO, RICE etc
• - Metrics to track for a dashboard
• - Writing release notes/ announcements
• - Generate a GTM strategy
• - Summarize something including books

There will be multiple products built on top of ChatGPT or using it to deliver value similar to no-code solutions available today. One can train the GPT-3 model on user feedback and reviews to identify common themes and pain points. Lenny’s newsletter featured an article on Lenny’s bot, where users can ask the bot questions, which is trained on Lenny’s newsletter archives. So instead of having to search across articles to get retention benchmarks, you simply query the information. To conclude, ChatGPT and other AI solutions will always have limitations, but those will shrink.

Additional reading:

• https://www.lennysnewsletter.com/p/i-built-a-lenny-chatbot-using-gpt
• https://tldv.io/blog/how-are-product-managers-using-chatgpt/
• https://levelup.gitconnected.com/chatgpt-guide-for-product-managers-top-20-most-important-prompts-a2c48dcc63c6

What use-cases have you tried ChatGPT for?