AI and Product Roadmaps
Its interesting watching hype cycles run their course and all of the chaos it brings. On one hand, if you add all AI things into the public roadmap your overall valuation will go through the roof. On the other, if you focus on customer value and solving their problems (maybe with AI) it will increase overall revenue.
While these two things can happen at the same time it seems like the focus is on the former and not the latter. How can I get people excited about the hype… maybe I can add AI to my washing machine… yeah people are going to love that. There are so many people with a solution looking for a problem.
With the cheap money thats been around for close to 20 years now, its hard to blame folks for trying to boost valuations and show user growth… who cares about making any money. This can be seen in some start up darlings heading into bankruptcy this year like Bird or WeWork.
I see a lot of these problems as a Product Management problem or a complete lack of product thinking. Or to put it a bit more bluntly, its a lack of market knowledge, understanding your customers needs, complete disregard for revenue vs costs, or the desire to just ignore what the market says.
Throughout this year, I have talked to many different clients around AI and it can be used in their products. I have seen some really great ideas and some that are well… not so great.
One of the ideas was around using GenAI to automate the creation of workflow templates. On the surface, this doesn’t sounds like a terrible idea. The problems really start to show up once you dig into how users might actually use this. The first question that comes to my mind is “how many times does a given client actually do this?”. The answer is that most clients may have ~10 or maybe 15 of them. Another question could be “How long does this task take now?”. Even the most complicated might take 1-2 hours at the most.
How often is this done and how long might it take
So let me understand here… You want to build a model that helps to automate something that might take a client a total of ~20 hours. This also completely ignores if you do get GenAI to make it happen, it will still need to be checked over and understood by a human to help if things go wrong. I just cant understand why this would even be considered as a feature to build, other than just adding AI to the roadmap. It doesn’t add much value and I would bet 6 nickels that very few would actually use it.
This is where a good product leader would step in and paint the picture of reality or at least understand what could really happen once its released. Maybe its still worth doing, but I am a bit dubious.
A great example I have seen is injecting AI into the enterprise knowledge management problem. This is a problem most, if not all, have and could provide immense value across any engineering organization.
Product people need to more focused on actual revenue and sales targets. They needs to be the ones that highlight some of these issues and help to steer an organization away from chaos. If we as a product community don’t step up, especially in these times, we will become redundant, show little value, and get pushed out of organizations.
Don’t forget to check out the Think Big Code Small Podcast!
You can find us on all major podcast platforms like Apple podcasts or YouTube!.