Does AI Make TBM and TBM Tools Irrelevant?
The Question: Won't AI make TBM and TBM tools irrelevant?
What's Going On: AI will soon know what every org spends and be able to tie that spend to the apps and services each business consumes.
What I'm Afraid Of: That TBM practitioners won't be needed in the future because AI can do it all.
This is becoming a popular question. The short answer is no, AI has not made TBM and TBM tools irrelevant. AI cannot do TBM end to end. It can make parts of it a hell of a lot easier.
An effective TBM capability follows a 7-step process:
Define your objective (use case)
Gather the data you need
Ingest and process that data
Define rules, logic, and assumptions
Calculate the numbers
Produce an insightful report
Embed this process into the culture and operating model
The foundational models or AI capabilities within data platforms can (and are) making some of these steps easier and reducing the dependency on a TBM tool for doing these steps. In many ways, it's reintroducing the "best in class" tooling vs. "platform" argument on how to solve a problem. Do I want a specialized tool to solve this problem, or can I use the tools I have to make a solution that works for me?
This may be self-serving because this is what we specialize in at Falconbridge, but defining the objective / use case and adoption steps in the process are not prime for AI to take over. You still need to bring people with differing perspectives and views of the problem together to navigate a decision. That piece of the equation isn't "commoditizable."
A few months ago, I shared my top 5 tactical ways of using AI for TBM programs on this blog post. Those are meaningful ways to push things forward for programs. For those new to the space, evaluating solutions, it's fair to ask where AI and the AI powered tools you already have fit within this equation. You'll need to decompose the 7 steps above and really evaluate whether your AI tools can (and should) be doing each step.
If you're a complex organization, don't minimize the effort to train and maintain your own data and allocation processing engine. People have done it, I don't think any would say it's simple or easy like AI prompting.
So if you're a practitioner worried about this, look at where your week goes. If most of it is cleaning data and rebuilding the same mappings, that's the work AI is coming for, and I'd let it. If more of it goes to getting people to agree, explaining what the numbers mean, and pushing a decision across the finish line, you're in good shape. That's the work practitioners have always been needed for, and I bet it stays that way for us.
Hope that helps!
—
Have a question of your own? Submit here.
