Q&A with Jill Dyche: I'm Ready for AI. But My Company's Not!
In which Jill makes nice with the engineers.
Sometimes we’re victims of our own competency. You lose the promotion because “you’re too valuable where you are.” You’re asked to deliver the same talk again—you know, the one that killed in Atlanta! The same conversation takes place in multiple meetings, like a workplace Groundhog Day. Your boss says, sunnily, “Keep doing what you’re doing.”
This is true of Carlos, who’s been working on an analytics team for what seems like forever.
I work in the Analytics Office at an automobile company. I can modestly say I’m an expert on analytics, having worked with several vendors, large and small, on dozens of projects. My colleagues are all just as and in some cases more talented than I am. We all like the variety of work and the credibility our team has. Plus we’re proud that our company has become more data-driven because of us.
However we all share the same frustration: we’re ready to adopt more advanced technologies. I recently saw a demo of a deep learning algorithm that call tell the difference between humans and non-humans. Think about the potential for an autonomous driving vehicle to distinguish between a traffic cone and a child. We could be saving lives!
A team member in the cube next to me just wrote a white paper on virtual and augmented reality. We all agree we could be doing some cool new stuff with AI. But our leadership isn’t receptive. Any advice?
-- Carlos, USA
Carlos, I’m not sure whether to offer my congrats or my condolences. Your team is certainly great at what it does. Let me suggest a gut check before answering your question.
Are you absolutely sure your team isn’t succumbing to BSOS (Bright Shiny Object Syndrome)? It’s been known to happen. A crack team of smart people will often get bored with the same ole’ same ole’ and hunt for opportunities to learn new things or try out new tools.
That’s not an indictment, it’s reality. Just be clear-eyed about whether you’re seen by others as truly delivering value. Even if there’s a one hundred million dollar opportunity for deep learning, you might not be involved if your team is considered too unfocused or academic.
Honestly—and I say this with love—I suspect there might be a little of this happening. Otherwise, why did your colleague write a white paper and not a business case?
Which brings me to my suggestion: Write a business case. White papers, faster algorithms or hiring more data scientists won’t necessarily lead to new opportunities. You need to justify the extra investment (in money and resources) by explaining in concrete terms what the benefits are. The business case should include:
- A problem description: what business problems will the proposed technology solve? Why is it unique?
- Proposed functionality: what does the new tool do? Stay away from technical jargon and overdone functions and features. Simply explain what it does that your incumbent solutions can’t do. Give an example like sensors on cars distinguishing between various objects on the road.
- Rough costs: This doesn’t need to be exact. Somewhere between napkin math and vendor proposals will do. Estimate spend for software, hardware, additional skills needed, and ancillary costs like integration or fresh data sets.
- Risk assessment: Create a list of realistic-but-avoidable risks. Be frank without sounding dire. Include not only technical risks (what happens if the software doesn’t do what the vendor promised?) but market risks (will the introduction of the new capability cause confusion with customers, partners, or the street?), and organizational risks (will implementation take resources away from another planned initiative?).
- The pitch: What value will the new capability bring to the company? Will it drive new revenues? Save costs or introduce operational efficiencies? Maybe introduce a new business model? Include a discussion of ROI if you can, and provide “soft” benefits such as employee retention and improved market perception if they apply. This is the “it’s worth it” pitch.
- Opportunity cost: It’s cheeky, but when I wrote business cases I always included the cost of doing nothing. Don’t wag your finger, just paint the consequences of not being a first mover, losing a potential customer segment, or incurring a lawsuit.
- The Ask: Don’t leave approvers hanging. Ask for what you need. Perhaps a software license for a new AI module, some development talent, and some servers. Or all of these. Avoid asking for the “bare minimum”—the solution you acquire should scale. And you want the appropriate level of commitment from approvers.
For teams like yours, assembling a compelling business case is more of a challenge than building the edgiest of algorithms. But it’s a critical exercise every analytics executive should go through, at least once. If you succeed, you’ll learn a new technology AND a new skill!