(Photo by Annie Spratt on Unsplash )

(I wrote this post a little more than a year ago but never got around to publishing it. A LinkedIn post from Solomon Kahn jogged my memory. One year on, it still holds true.)

Be honest with me: how are things going with your company's ML/AI efforts? Are things not turning out as planned?

Maybe you overestimated your company's needs for this technology. Or your predecessor had a huge appetite for AI, and they've left you holding the bag. Now your CFO is hounding you about this six- or seven-figure hole in the balance sheet.

You need to come up with answers. Fast.

Here are four steps you can take to chart a new path, three of which don't involve layoffs:

1/ Take a second look at the data team's work

Data scientists don't often brag about their work. So it's entirely possible that the team is providing value, but they just haven't told you. This is a good time to check in with them to see how their work aligns with the business mission. (While you're there, why not remind them to be a little more boastful?)

You may as well check with the other department heads, while you're at it. AI might be a key driver of your loyalty program and marketing campaigns, or otherwise providing significant value. And that's precisely the message to take back to your CFO: "if we cut back on AI, we'll be cutting back on our business."

2/ Check for other opportunities to use AI

Maybe the data team could stand to contribute more, but they just don't know how. This would be a good time to review your company's overall business model and challenges, in search of new opportunities for AI. And since you already have a team of data scientists on-hand, you'll be able to execute on those ideas in short order.

Speaking of the team, be sure to involve them in this search. They can meet with key stakeholders and product owners to see where AI could take things up a notch.

Having gone through steps 1 and 2, you may determine that your data team is indeed overstaffed. Fair enough. It's time to think about restructuring.

3/ Find other spaces in the org chart

The members of your data team already work in the company. They know the product and hold valuable institutional knowledge. Assuming this exercise is just about the data team (not a company-wide downsizing) you may be able to arrange moves to other departments.

Some will be small shifts, say, from data engineer to backend application developer. Both roles involve writing a lot of code for machines to talk to other machines. There's a good chance that your data engineers came from backend developer jobs, too, so this move would be like slipping into a comfortable old pair of shoes.

Larger shifts are also possible, especially for the data scientists who are interested in a career change. Perhaps they'd like to join the finance or marketing teams? (Bonus: with enough time, they may uncover new use cases there for AI.)

I promised you three steps that would not involve layoffs. If you've gotten this far, it's time to consider reducing headcount. If so:

4/ Get creative

Your company doesn't have enough work to merit a full-time data scientist or machine learning engineer. But what about a part-time role? Someone who is heading back to school or needs more time for parenting might be happy to cut their hours. Can the two of you carve out a mutually-beneficial arrangement?

Similarly, I'll bet someone on your team would like to build a data consulting business. Could they work with you on a part-time, contract basis? You'd get a data scientist you can trust, and they'd get their first client. Win-win.

Next steps

My hope is that you were able to stop after step 1 or 2. If you still find yourself needing to reduce headcount, please remember that it's not your data scientist's fault that the company overhired. It's now up to you to take responsibility.

As these dedicated employees transition to their next roles, show them the dignity and respect they deserve. That means supporting them with hiring assistance, severance, and anything else they'll need to move on.

If you need help assessing and restructuring your data team, please contact me. One of my AI Assessments can surface gaps as well as missed opportunities.