I recently stumbled across "Lead Bullets," an gem of a Ben Horowitz post on the Andreessen Horowitz blog.

It really resonates with me. You may enjoy it, as well.

The gist is that we tend to look for the fabled, magical "silver bullet" of a solution to a big problem. In "Lead Bullets," the message is clear: you can't wish away a problem with a magic solution. You need to roll up your sleeves and get to work on addressing the core issues.

Reflecting on the data field, we often see companies that want a silver bullet for their ML/AI work. That, or, they treat ML/AI itself as the silver bullet.

Instead of waiting around for the magic solution, you need to identify and address your ML/AI problems head-on:

If your company treats ML/AI as a silver bullet: Start by reviewing your business model, revenue streams, strengths, and weaknesses. Develop a realistic and viable data strategy that outlines concrete plans for using ML/AI to improve your business.

If you're lucky, you'll be able to shift focus and keep moving ahead. You may also face the difficult job of restructuring or downsizing your data-related teams. This goes beyond the data science team itself. You must also consider product managers or software developers you hired to build out your AI-powered dream.

If your company needs a silver bullet for its ML/AI implementations: Your first step is to pause any ML/AI activity so you can review the team, the projects they are working on, and the data they are using. What specific problems do the projects solve for your business? Assuming you still need to undertake these projects, are the team members -- data scientists and product managers alike -- the best fit for the work they're doing? And is ML/AI even the best approach to the problem you're trying to solve?

Again, you may have to make some uncomfortable changes. It's possible that you've over-hired on your data team, or that the current slate of projects has no real business value. Maybe you need to ditch that under-performing ML model in favor of a straightforward software approach. (Twitter's use of an ML model to crop images comes to mind.)

By revisiting your data strategy (or creating one, if you haven't done this already) you may find that you can to move people to different teams or projects instead of downsizing. But you won't know until you work through this exercise.

Doing the hard work

Whether you're a data scientist, product manager, or executive, you will face tough situations.

If you find yourself looking for a silver bullet, that means you know deep down what you need to do: re-read Ben Horowitz's "Lead Bullets," And then roll up your sleeves.

There's no time like the present.

Need an unbiased review of your company's data strategy, team, and processes? Maybe you require due diligence for an upcoming merger? Contact me to schedule an ML/AI assessment.