Recently I attended a tech meetup on the challenges of product managing AI products. And wanted to share some relevant highlights for us tech product users.
Firstly, I should explain most attendees were people who work for vendors that create products, rather than companies that use products. So, this kind of skewed their perspective that AI is something you develop.
In a pre-session networking breakout, someone threw out a question for discussion: did people think AI was more often found at a startup than at an enterprise?
I said I look at it differently because my experience includes applying technology to business problems, not just the back end job of creating technology. Also, because I focus on helping marketers, I know there is a real need and use for AI products in marketing. Everyone is looking to make use of it. Though, unfortunately, at times it’s enterprise that has the budget to really do it. So, no not just for startups.
It also depends on what you mean by AI. If you mean self-driving cars. Then yes, maybe it’s the realm of startup tech or big tech. It’s another story if you mean chat bots, predictive analytics, and personalizing content. For Marketing teams, AI is here and we are using it now. Everyone wants to use it.
Then… time ran out in the breakout. We transitioned back into the main room for the presentation and I climbed off my soapbox, lol.
Though the first challenge discussed was inline with my way of thinking. The challenge was to make sure that the AI product they are developing is solving a real-world business problem in need of AI. Not just because the tech team thinks it’s a cool thing to do.
The speaker pointed out that AI is often used when the task at hand needs to be automated or streamlined. AI is of benefit, when the job to be done requires speed, scale or quality that is beyond what a human can do. (Though we could debate that at lot of tech does that to a certain degree.)
Which is why another requirement around AI products is trust and transparency. If it’s something we can’t do manually, we need to at least understand the product enough to believe it’s doing the job we need done. I’d add, especially for marketers, if we are putting it between us and our customers. Otherwise, we will hear about it!
The final challenge was another big one for AI consumers. Delivering the right end-user experience! Making sure it was easy for us who apply it to use it. Otherwise, it will fail as a product.
I hope this is some food for thought on what we should expect from our AI products. The developer’s challenges are our must haves!
If you enjoyed this, check out a previous newsletter on making sure your AI products are providing the AI value.
