John Roese: I would argue that the other two types of AI are much more important even though they’re much less visible. In addition to AI to improve the human condition, the second domain that we think is probably significantly more important and probably more valuable is applying artificial intelligence to transform every single business process that exists today.
Business processes always have a level of automation that’s done by the computing infrastructure and the software. In almost every case, that automation stops at a point when human thinking is required. What happens with AI when you apply it to business processes is that boundary between when the human being has to intervene in the business process versus when the computers do their work suddenly shifts dramatically.
With that, you start to improve the ability to reason across larger datasets, to think through more scenarios, and to analyze the data in more interesting ways. That has a profound impact. It’s exactly the same type of impact as we had when we first digitized or computerized business processes a long time ago. It used to be a general ledger that told you your financials. Suddenly, you can put it into a computer and you could use a spreadsheet and do pivot tables. You can think through things in more meaningful ways.
When you went to Salesforce for a CRM system versus just trying to keep a database of customers, you could start to run analytics against it. Machine intelligence applied to business processes doesn’t replace them. It just dramatically improves them almost to the same degree as when we first put them into computer systems the last time around. That one, unfortunately, is much less visible. Very few people see the business process mechanics. They just benefit from them. It requires, primarily, IT expertise, business process expertise, and computer science.
While that is harder to do, it actually has a bigger impact than simply creating a better natural human experience than that first example would solve for. We’re extremely excited because if we think about how many business processes exist in the world, it is almost innumerable. There is an interesting statement that says that every business process can be made better, more efficient, more productive, and have a greater probability of success with the incorporation of machine intelligence to accelerate the thinking tasks that, historically, had to happen by human beings after the business process ended.
We’re extremely excited about that because business processes aren’t just about CRM and financial services. Business processes are things like having an IT infrastructure know where to place a work load – to know that this particular application should be placed in cloud A because it has better economics. It could be around understanding how to automate your compliance and regulatory obligations.
Today, there are great tools that will give you lots of data. Ultimately, a human being has to think through this. If we can push that thinking task into the automation layer to change that business process, we take away human error and we also accelerate the process, and we can start to scale beyond what human beings typically can do when managing IT systems or business processes as a direct participant.
Sramana Mitra: If you look at our Thought Leaders in Artificial Intelligence series coverage, I’d say the bulk of the stories are focused on business process automation in various domains. We’ve gone from mortgage lending, cyber security, ad tech, to all sorts of niche business processes that are domain knowledge heavy. To be able to impact specific business processes and domains will require enormous amounts of domain knowledge that you can then algorithmically design into an application.
From an entrepreneur’s point of view, those are really the business-building opportunities of the next generation. There may be some opportunities in autonomous cars and stuff like that, but I think a large number of business-building opportunities are in the realm that you are talking about.
John Roese: Absolutely. I stood up in front of about 20,000 of our customers and partners at Dell Technologies about a month and a half ago and walked them through these three domains. One of my punch lines was, “While we’re obsessing about the first one because we can see it and touch it and it’s good in popular culture, but it’s not really a business opportunity for many people.”
The people who run IT systems have that deep level of business process understanding, but they’re now going to get a whole new set of tools around AI and machine learning to try to reinvent it. The two pieces necessary to actually do these kinds of transformations in the second domain of business processes is deep understanding and a technology understanding and open-mindedness to using these new tools and capabilities to solve the problems in a different way.