AI has become a core focus in my analyst practice, and now all the industry players have an AI story
: Cisco, Microsoft, Avaya, Genesys, Mitel, Five9, Talkdesk, Twilio, and Zoom, to name a few. The capabilities for each are impressive, and they address a specific set of challenges facing enterprises and contact centers.
Our market space absolutely needs these solutions, and they will keep getting better as AI matures. To a large extent, these vendors are helping customers solve problems in better ways than workers or agents can without using AI. As AI’s track record improves, it will also be applied to problems humans can’t solve efficiently – such as detecting customer dissatisfaction in real-time across millions of call records instead of the handful that a supervisor can review manually. Going a step further, AI will identify new problems that we didn’t even know existed, such as identifying unknown forms of fraud arising from home-based workers accessing meetings over unsecure networks.
Thinking About Bigger Problems with AI
These capabilities will certainly support the business case for AI, and many vendors will make healthy profits when the benefits are finally delivered. Overall, this is a welcome development in our space, but it’s important to keep in mind that communications and collaboration is just one of many applications of AI, and in the grander scheme of things, these are actually rather small problems.
Of course, they’re mission-critical for any business, but it’s not the only way to think about AI’s potential. I’ve long advocated that use cases should drive IT decision-making rather than going with a solution that’s “AI-enabled.” AI itself is not a solution, and its value should be based on the problems being solved rather than how cool the technology is.
The problems of worker productivity, agent engagement, and customer satisfaction
need to be addressed, but my mantra lately has been about thinking bigger. AI is about scale and speed and is valued for its ability to solve big problems.
Humans can solve small problems fairly well, but AI does it better.
Conversely, the bigger problems remain problems precisely because we can’t solve them ourselves, and in today’s world, there is absolutely no bigger problem to solve than COVID-19. Our space is poised to benefit nicely from AI, but the stakes are much higher in healthcare and medical science, and I would encourage IT decision-makers to look in this direction for inspiration to think bigger with AI.
Listen to the Science: Michio Kaku Says AI Can Defeat COVID-19
Kaku’s name will be familiar to many, as he's a renowned physicist and acclaimed author, and I was lucky to see him speak at a recent NEC virtual event
. Kaku showed how AI can be applied to larger problem sets and how it can perform tasks that humans simply can't do while at scale.
During his talk, Kaku discussed various ways that AI can defeat the pandemic, both with new capabilities as well as leveraging existing technology that’s widely used today. Some of this is about having the data science expertise to utilize AI applications, but it’s also about connecting the dots among things in plain sight and layering AI over that to glean new insights to address the problem — such as in this case, to stop COVID-19.
In other words, the coronavirus itself can be very harmful — but only under the right conditions. The real harm is caused by human behavior, and AI can track our behavior and identify risk scenarios before they become dangerous to others, as in super-spreader events. Along those lines, he talked about using AI with Internet platforms like Google search and Facebook to detect anomalies that provide clues where COVID-19 is present.
As Kaku explained, someone experiencing COVID-19 symptoms might type “I’m having difficulty breathing” into Google, searching for answers. Taken in isolation, that’s just another search query, but it’s a pretty telling indicator of what might be happening. Then, by tracking their Facebook activity, AI can get a good read on others whom that person will likely be in contact with, and in turn who each of those will be in contact with soon after. Layer on that smartphone tracking, and it becomes pretty easy to pinpoint exactly where that person is going, at what time, for how long and with whom.
Each of these applications is an important piece of the puzzle, but in isolation, they don’t tell you much. Reading across all this, AI can draw a pretty reliable heat map of where pandemic spread is likely to occur, and if corrective action is taken, it can be curtailed significantly.
Naturally, there can be false positives, as that query for “I’m having difficulty breathing” may simply be searching for a song lyric or a movie title. On the other hand, one of AI’s virtues is not making the same mistake twice. Once an error is identified and corrected, it never happens again with AI, and the same can’t be said for humans.
Kaku takes this another step further by looking at global air traffic patterns. While community spread is very localized, what makes COVID-19 a true pandemic is the ease of transmission by virtue of human travel. No doubt, a great deal of virus spread comes from traveling by auto or train, but air travel takes things much further, much faster, and on a greater scale. In fact, he explained that 60% of COVID-19 entry into the U.S. came via air travel, specifically at either JFK or Newark airports.
Layering concentric rings of human behavior from local out to continental circles of travel, AI pieces together a much bigger, global picture of how COVID-19 spreads. AI was built for providing this level of understanding with such massive, global data sets — and doing it in real-time. The key to defeating the pandemic is stopping large scale spread, and that has nothing to do with the virus itself and everything to do with understanding human behavior – which en masse is highly predictable. Big problems need big solutions, and for context, Kaku cited this as a key reason why the Spanish flu was so deadly. The conditions are no different than today, but in 1918, we simply didn’t have the technology to detect spread before it gets out of hand.
What Should You Be Thinking About?
In contrast, the challenges faced by IT decision-makers may seem pretty small, but they still need to be addressed. Current applications of AI in our space become no less valid, and I’m sure Kaku would agree. However, I think he would also say that the most important thing remains how you think about the problem set.
This is the essence of the scientific method, and so long as you only focus on specific use cases, the impact of the solution will be fairly small. As stated at the outset, AI is really about solving the big problems – the complex problems that we just otherwise can’t even begin to think about or solve.
As a takeaway, I would encourage IT decision-makers to think about the bigger problems that go beyond making individual workers more productive or specific customers happier. Some vendors have made moves in this direction, such as with Microsoft Graph or “cognitive collaboration” from Cisco. These models have power by virtue of connecting millions of data points from our everyday activities to identify patterns to improve workflows and detect anomalies before becoming problematic. AI is really good at this, and as we edge closer to the world of IoT, those data sets will become even bigger but also will become more accurate predictors of our behavior.
Of course, the potential is there for a hard right to the workplace becoming a surveillance lab where every move is monitored, and workers have no choice but to conform to the operating methods chosen by the enterprise. However, Kaku seems to have only good intentions, and unless you’re a big pharma cynic, his lessons learned are noble.
There’s no reason why enterprises can’t be thinking the same way, especially multinationals with massively distributed workforces and customer bases. Some are already well down that path, but for many, the focus is on the small stuff. Dare to dream and think bigger — if AI can curb this pandemic, there’s no business problem too big for it to handle. However, it’s not about technology; it’s how you think about the problems.
This post is written on behalf of BCStrategies, an industry resource for enterprises, vendors, system integrators, and anyone interested in the growing business communications arena. A supplier of objective information on business communications, BCStrategies is supported by an alliance of leading communication industry advisors, analysts, and consultants who have worked in the various segments of the dynamic business communications market.