What do you do all day when you are jobless, and job hunting is getting boring?
I talked to ChatGPT, of course.
In the name of “training”, today’s question is about validating my business ideas, and I enlisted ValidatorAI and Speedy Marketing to the table for the group discussions. I called it a “Blue-Sky session”.
“Act as a Venture capitalist and startup expert; you are tasked to evaluate and validate this business idea. Ask me 10 questions so you can start to evaluate and validate the business idea.” - prompted Jessie
Predictably, ChatGPT asked me a bunch of questions that are standard in the evaluation process. Back in the good old days, I was the one who asked these questions.
- What is the core product or service your business will offer?
- Who is your target market?
- What problem does your product or service solve for your target market?
- Who are your main competitors?
- What is your business model? How do you plan to generate revenue?
- What is your marketing and sales strategy?
- Do you have a prototype or MVP (Minimum Viable Product)?
- How do you envision growth over the next 5 years?
- What is your financial projection for the next 1-3 years?
- What kind of team do you have (or plan to build) to execute this idea?
The replies from my digital colleague were sensible, so I pushed on:
Recommend strategies for "cons and important things to consider" - provide a step-by-step guide to plan for these potential drawbacks and build on existing foundation and goodwill.
Then I also asked the team to develop a business plan including all sections based on Business Model Canvas and Marketing 7PS.
Once again, the replies were sensible and logical. No bugs nor hallucinations. It’s almost as good as sitting with colleagues bouncing ideas around.
I followed up with this question several times to address the “cons” and the “challenges”: Critically evaluate this Business Plan. Outlines challenges and the strategies to overcome them.
Finally, to address the elephant in the room, I prompted the team once again:
- write a well-defined USP
- develop a robust and well-defined marketing strategy for the first 6 months
- develop a customer retention strategy to reduce the churn rate
A few more discussions in between to address these challenges, redefine business goals and add more context. The last question I asked was, “Rate the validity of the forecast on business performance in 24 months.”
The team concluded with this statement at the end of our Blue-Sky session.
“Given the well-defined USP, marketing strategies, and customer retention strategy, it's possible that the business could perform well. However, reaching $xx in annual revenue in the first 2 years will still require a strong marketing and customer retention strategy. Therefore, on a scale of 1 to 10, with 10 being the most confident, I would rate the validity of the forecast an 8. Regularly reviewing and adjusting the forecast based on actual performance is essential.”
The goal needed to be revised to start with. With a modest tiered pricing plan, the hairy goal was $2 million in recurring revenue in 24 months.
The lessons learned in the past 3 months of working with my digital colleagues, ever-expanding ones with hundreds of plug-ins added to the ecosystem weekly.
- Having a learner mindset, don’t expect you will get the correct answers at the first prompt. It’s a learning process akin to onboarding a new colleague. Context matters.
- Well-defined roles and instructions help with communication. No surprise here.
- Reframe the questions and ask them multiple times. These exchanges build a richer context to dive deeper into the topic with a mutual understanding.
- Keep up to date with the plug-ins and its limitation. AI is evolving rapidly. It’s not about access to the information anymore. It’s about our ability to “evaluate” the information that matters. Not all information is good information.
- Evaluate the outputs all the time. I can ask the team to do almost anything in a business setting because I can evaluate the ideas. Even though the team can generate Python code, it’s not something I would prompt simply because I cannot understand the context myself. As an amateur “Prompt Engineer in progress”, it’s also good to know our own limitations.
What are your thoughts?