Gary Melling is the President and CEO of Acquired Insights, a firm that designs customized AI applications and tech-driven strategic initiatives. He’s also a speaker and thought leader widely sought out for his insights into the digital economy. We spoke to Melling about the rise of automation and the imperatives it creates for today’s companies.
We hear about companies becoming “data-driven.” What’s distinct about working with digital data compared to the insights of the past?
At an IBM lecture a few years ago, the speaker famously noted that “90 percent of all the data that exists in the world was created within the last 24 months.” People often forget his next statement: “90 percent of all that new data is unstructured.” So if we think historically about companies with an ERP, they’re typically using structured data (strictly defined and classified), and they’re not very proactive about pushing insights toward users. What that means is that many companies are only working with 20 percent of the insights available to them. Accessing the other 80 percent that’s unstructured or semi-structured is going to be key.
In addition to structuring it, what other challenges arise from managing, securing, or monetizing data?
Knowing that technologies like AI and machine learning were coming, many companies started storing data preemptively even before they knew when or how technology would turn it into something valuable. Now we’re finding out that a lot of that data in storage doesn’t contribute to business decision support. Instead of thinking about collecting as much data as possible, companies are stepping back to ask: What data do we have, how clean is it, and how current is it? What are the obstacles to integrating different aspects of that data? A lot of organizations think their problem is not having enough data when it’s really about how they utilize what they already have.
Why aren’t they utilizing it correctly? Are there institutional obstacles?
AI hasn’t been mainstreamed for business leadership. The executives who are a year or two away from retirement haven’t built their careers on AI and machine learning, and that’s an obstacle. I think there’s about to be a changing of the guard, both in terms of who leads and how. Previously, executives often relied on “squishy metrics” to prove they were generating revenue, reducing costs, optimizing CapEx or OpEx, or retaining customers. Now that executives can and must prove performance with hard numbers, we should see a lot more automated technologies become mainstream.
What, specifically, do AI and machine learning do to help companies learn more from data?
I was recently working with an oil and gas company worried about predictive maintenance, namely whether they could delay equipment maintenance past manufacturer recommendations to save time and costs without putting unnecessary stress on the equipment. AI and machine learning along with plenty of data helped provide an answer to which the company felt confident committing. Healthcare is another example. Formulating complex diagnoses means analyzing reams of data, and that can take longer than a patient is able to wait. With AI and ML, we can reduce 12 weeks of analysis to 2-4 hours. So there are many levels of benefit when using these technologies.
Are AI and machine learning too complex for smaller companies? How can they leverage their data?
I think a lot of business leaders see AI on the cover of a magazine, read a brief article about it, and never really grasp why or how it could benefit the business. At worst it seems like an unproven risk, and at best it seems like a nightmare of vetting partners and picking the right products. The mystique around AI can often obscure the reality of using it, but for people who are interested, there are plenty of resources to help executives get up to speed about AI quickly.
Speaking of speed, do companies put themselves at a competitive disadvantage when they don’t embrace technology quickly enough?
It varies by industry, but I think most companies have a window to implement things like AI and machine learning. There might be some pushback at people-focused companies as workers worry about robot replacements, but I think the key is to frame AI as something that augments certain tasks and complements the work humans do, not something that brings wholesale changes or redundancies. I think people generally overestimate the speed and scale of technological change. That fact aside, if companies don’t start preparing now, they might not have enough time before the window closes, and then they end up obsolete.
What can companies do to maximize the effectiveness of new technology?
Back in the mid to late 90s, fears over Y2K drove a lot of ERP implementations. I remember seeing teams of hundreds of consultants working on $400 million implementation projects. Even with all those resources, many of those projects failed, between 75 and 90 percent. One of the things I observed in the midst of all this was that if organizations spend a minimum of around 20 percent of their total budget on organizational change management, they significantly increase their probability of success. You can have the best technology in the world, but if the workforce doesn’t understand why a change is being made, they won’t buy in. It’s about communicating and managing expectations.
What’s your advice for 2020 in terms of data and technology?
The best time to begin is in the beginning. Before companies can do anything else, they need to explore the viability and ROI of new tech projects. I heard a quote recently, “No one wants to be first to be first, but everyone wants to be first to be second.” Yet I see a lot of companies still sitting on the fence instead of taking those important initial steps. That delay could have real consequences in terms of profitability and competitiveness in as little as 18 months, so now’s the time to start having some hard, serious conversations about data and technology in your business.
For more information about how to leverage your organization’s data to power your business, download our ebook: A Guide to Driving a High-Performance Organization today.