Trust but verify – how I use ChatGPT for competitive analysis.

Trust but verify – how I use ChatGPT for competitive analysis.

It’s been a few weeks since I did a post – I was heads down, working on client deliverables.   I guess I shouldn’t complain – it’s the life of a consultant!

I recently received an email from a SaaS software company. The email copy and the whitepaper (as a download offer) clearly showed that they were both written using ChatGPT without much verification. The email copy was very generic, and the white paper spent a lot of real estate educating me about the problem that I, as a former CMO, already knew so much about and had dealt with daily. However, the white paper lacked any insight and was filled with buzzwords when it came to providing a solution to the problem. It appeared that ChatGPT processed some brochureware and regurgitated it in its response to the marketing person’s query. Then, the marketing person cut and pasted the information into the whitepaper. However, I don’t blame ChatGPT for this. The marketing person was either lazy or did not know the domain well.

Ideally, a better approach would have been to use the white paper’s table of contents (TOC) to formulate a series of queries for ChatGPT (or your favorite Gen AI tool). Ask the AIL chatbot a question for each section in the TOC and then refine that question iteratively by adding additional criteria until you are happy with the response. Review the listed sources to ensure the responses are correct and credible. This ensures that the tool is not hallucinating or using a source you cannot trust. This ChatGPT-created content then becomes the starting point for the white paper. Remember, ChatGPT creates the recipe for you – not the dish itself.  You still need to put in the effort to refine it and add to it so it is meaningful for your target audience.

You can similarly use generative AI tools (such as Microsoft Copilot) to create a starting point for competitive analysis. Start with a prompt and refine the query to get your desired answer. For example, you may start with something like: “How is Salesforce Marketing Cloud better than HubSpot? List the sources of the information.”  This will give you high-level strengths and weaknesses. Then, pick one of the points listed and dive deeper. For example, “How is the salesforce marketing cloud more scalable than HubSpot?  Show me sources of information.”  You can review the response & sources and follow up with “Show me technical and architecture reasons that show how salesforce marketing cloud is more scalable than HubSpot.  Show me sources of information.” You can then pick one of the points and go deeper. For example, “How does salesforce marketing cloud offer better customization than HubSpot?  Place the responses in a table format. Show me the sources”. Below is the response you would get. And, of course, you can go even deeper until you are happy with the response.

This is just an example of how you can iterate to the nth degree and get the initial data set to start doing competitive analysis.  Remember – never ever share any proprietary information with Gen AI products. They can use it to build their ‘knowledgebase,’ and now your company’s (or client’s) proprietary information is public. And you may be in violation of your NDA. I use the GenAI products to get all the publicly available information about the competitor and various sources, including white papers, customer reviews (such as G2), third-party papers and research, competitor documentation, etc., in one place.  Then, I can begin to do deep dives to get those competitive nuggets that sales and product management are looking for.  Thanks to GenAI, I can spend more time digging deeper into the competition for my clients rather than spending a lot of time searching for sources of information.  In short, I can speed up the process while building a better data set for competitive research and analysis. I do competitive analysis for a living for my clients, so getting a head start using GenAI tools is very helpful.

If you have any questions about this approach, please reach out to me via LinkedIn at https://www.linkedin.com/in/applicationsmarketing/