I wrote up how the Tanzu Platform product management team has been using AI, from talking with Chris McClanahan, one of the co-leads of product management, and Purnima Padmanabhan, the Tanzu GM.
The first thing you want to see, of course, is the speed-up. As Chris said:
[I]deation, which used to take, you know, six, seven, eight weeks to go through a single run of engineering is now a couple days at most.
Almost none of what got them there was tooling. It was org chart work – “culture” stuff. They put the product managers, UX, developers, architects, and end-to-end testers in one pod so that nothing has to be handed from one group to the next. As Purnima said:
You have to do some organizational changes. So, we reconfigured some of our development teams into a pod-like model, where we had the software developers paired with the product managers, the UX, the architects, the end-to-end testers all together in a small pod, and gave them more responsibility with AI, but also gave them more accountability so that they could discuss things faster, resolve things faster, and move much faster.
That is the same move we were selling in 2015 with a different justification, and the thing I was going on about a few weeks ago: AI amplifies whatever you already are. If an agent can go from idea to working prototype in a couple of hours and your organization still needs a meeting to pass the spec along, you have not sped anything up, you have just moved the bottleneck to the calendar.
Onboarding the new AI hire
Chris frames giving the AI context as a hiring problem:
It’s like a new hire, what am I going to give a new hire to get them going with the platform?
So they fed it knowledge base articles, documentation, customer information, historical blogs. The public stuff, the code and the docs, is already in the model. What is not in there is ten years of how your customers actually use the thing, and nobody else can supply that for you.
So the humans have to pass that along to the AI, just like you would a new human on the team. Once it has the context, the way Chris works with it is to make it interview him before it does anything:
Doing more work is exhausting and can cause errors
Being the human in the loop is tiring, and being tired makes you bad at it. One of the primary things humans do with AI is judge the results: use that human ability to quickly synthesize a huge amount of information, plus go with intuition and gut feel to answer the question “is this right?” That kind of judging takes a fresh mind. As Purnima said:
What we found is you can get into fatigue on the skeptic role. So, we have to keep changing that skeptic role, so that the brain stays fresh and you’re looking at what is being produced with the context.
Chris has the same problem from the other side, which he calls “staring ‘AI syndrome’” – sitting there reading the agent explain itself until your eyes go blurry.
Verification is the part of AI work that everyone says is essential and nobody staffs. Rotating people through it is a small, cheap fix that I have not heard anyone else mention.
The uninvested judge
That said, the AI is good at a type of judging humans aren’t always good at. Being an a-political, impolite judge in favor of reality. As Chris put it:
you have some great ideas as a product manager – because let me tell you, we have some great ideas, and about ten percent of them are good ideas.
If it’s not clear, Chris means “great ideas” as in eye-rolling, scare quote great ideas: ideas that actually aren’t great or, at least, are not practical to chase in the product. The agent has no stake in your great idea…and isn’t too concerned about the human-politics of telling you so. So, when it comes to “great ideas” it will tell you nobody is going to use it, that your, you know, baby is ugly.
Check out the whole write-up over on the Tanzu blog, and the conversation it came from is on YouTube.×
Leave a Reply