ChatGPT Needs Adult Supervision

Will AI, machine learning, and ChatGPT replace product management? Not as long as human insights are needed.

Do we need people anymore? Just use Python and JavaScript to create your backlog. Use ChatGPT to write your blog posts.

Not so fast. Creating a backlog or writing a blog post requires insight, not just data.

In a Fast Company article, Alexandria Brown, a human resources consultant, said she would not be offended if she learned that someone used ChatGPT to help present themselves to her. “If you’re using technology to work smarter, not harder. I mean, I applaud you.” However, she warns, "You would still need to edit it before you create your final draft."

Microsoft Word and Grammarly give us helpful tips, not only spell-checking but grammar checking. Usually, they're right, but often they're not.

Have you ever worked with a consultant that does not understand your business or domain? You eventually learn to rely on experts for the content and rely on consultants for the mechanics. Use experts for the judgment and AI for the execution.

When ChatGPT was asked why NOT to use AI, it responded:

• Lack of originality: AI-generated content may lack originality and creativity.
• Lack of context: AI may not fully understand the context of the blog post and generate irrelevant content.
• Unnatural language: AI-generated language may sound unnatural or mechanical.
• Copying: AI may generate content that is similar or identical to existing content, which may result in copyright violations.
• Brand image: Using AI-generated content may negatively impact the credibility and image of your brand.

Also, because it pulls much (all?) of its information from the web, its results have the same prejudices and biases that exist on the web.

Content marketing experts Andrew Camel and Kyle Poyar tell us, "People have gotten away with writing thoughtless, repurposed content for too long. Value in the next decade will accrue to thoughtful humans and those who have helpful and original ideas."

Jakob Nielsen, a leader in defining the field of UX, recommends that you Write Articles, Not Blog Postings. He explains, “To demonstrate world-class expertise, avoid quickly written, shallow postings. Instead, invest your time in thorough, value-added content that attracts paying customers.”

Human judgment

AI, such as ChatGPT, must be guided by human judgment. By people who know the business and the domain. Those who have a hypothesis to be proven.

Those who have a point of view.

Start with interviews and observation to discover a problem and create a hypothesis. Then analyze the available data, including product usage, local data (such as found in your CRM system), and public data (such as information in LinkedIn). That’s where AI and machine learning will save you time and find correlations that are not readily apparent. These data insights will inform your experience and intuition to reach a conclusion.

For example, Walmart incorporates weather data into its supply chain analysis. They can predict spikes in demand for certain products, such as umbrellas or groceries for storm preparation, and adjust their inventory accordingly.

A good product manager is data-driven; a great product manager is data-informed.

ChatGPT agrees: “Yes, a good product manager uses data to make informed decisions and prioritize features, but a great product manager also considers user feedback, industry trends, and their own intuition to make the best decisions for their product.”

You must have thoughts to be a thought leader. — Steve Johnson