In the beginning, there were just compute and storage services — followed in the next 16 years by more than 200 other cloud infrastructure services offered by Amazon Web Services Inc. But ahead of his keynote at re:Invent Tuesday morning, AWS Chief Executive Adam Selipsky hinted at his ambitious plan to move Amazon.com Inc.’s cloud computing unit up the software stack.

Selipsky (pictured) told SiliconANGLE in an interview that the company he took over earlier this year will offer more higher-level services such as Amazon’s Connect cloud contact center, and also will expand into offering more complete industry vertical solutions such as HealthLake for the healthcare industry.

In this, the second of a four-part interview with me and Wikibon Chief Analyst Dave Vellante ahead of the company’s re:Invent conference this week in Las Vegas, Selipsky explained how embedding artificial intelligence and machine learning into all of its services will help reduce the need for enterprises to do Lego-style cobbling together of individual cloud services.

You can get the big picture from Selipsky’s entire interview here, and don’t miss the other installments  of the full interview in coming days. Also, check out wall-to-wall coverage of re:Invent by theCUBE, SiliconANGLE Media’s livestreaming studio, and SiliconANGLE all this week and beyond for exclusive interviews with AWS executives and others in the AWS ecosystem. If you’re at re:Invent, stop by theCUBE’s studio in the exhibit hall.

This interview was edited for clarity. (* Disclosure: SiliconANGLE and theCUBE are paid media partners at AWS re:Invent. AWS and other sponsors have no editorial control over content on SiliconANGLE or theCUBE.)

Leveling up

Furrier and Vellante: What does the next 15 years of innovation look like? How does Disruption 2.0 happen?

Maybe I’ll try and tear off a few pieces of it and we can see where it takes us.

I remember in 2006, 2010, we used to say that in the fullness of time, we believe that almost all workloads will move to the cloud. And that it could take 20 years or more. Now that it’s 2021, you can see just how dramatic the shift has been, and just how many organizations, big and small, private and public, have migrated very significantly to the cloud. We are well along the way.

On the other hand, it’s still early. Depending on which third-party estimates you’d like to look at, probably anywhere between 5% and 15% of IT spending has moved to the cloud. If you look forward a decade, you’ll see that most big enterprises, most government organizations, probably almost all startups are not only working in the cloud, but are substantially all in on the cloud. And so I think that overall adoption will just have deepened significantly. The acceleration will continue or arguably even speed up.

Furrier: What are some of the key trends in cloud?

One of them is the use of machine learning and AI in everything. I think we got off to a really good start with our ML and AI services. SageMaker is used by tens of thousands of developers and a lot of other ML capabilities as well. You’ll see us … continue to vend ML services that customers can consume directly. Things like SageMaker, things like our industrial ML services that we released last year, like Monitron and Lookout, which are already being used by companies like GE and Siemens.

But at the same time, equally exciting, I think you’ll see ML and AI baked into probably every other service that we have for customers. If you look today, you can already see powerful, but early, signs of that. If you take Amazon Connect, our call center solution, we don’t expose to those call centers, “Oh, here’s this primitive to do text extraction. Here’s this primitive to draw inferences from conversations that your reps are having.” We just expose that as high-level capabilities, which you don’t have to be a technical genius to interact with. 

And so if you take that metaphor and it just apply that across all of our services, our database services will be smarter. If you put data into the cloud, we will undoubtedly be able to know or predict which service should do the querying of your data.

Connecting the dots 

Furrier: Are you saying that that dynamic of enablement is going to be portable to other services?

Yeah. If you just take a step back, there’s 30 to 40 years of entrenched IT inside of data centers. And probably $4 trillion dollars a year of IT spent in data centers. And there’s a lot of functionality there. It is not one or two pieces of functionality. There are a lot of things. Now, nobody thinks it’s a good idea just to lift and shift all of that muck into the cloud, but there’s no getting around the fact that there’s a lot of sophisticated functionality that’s going to need to move to the cloud.

And so, that’s a reason why we have over 200 services today. Because there are lots of different types of customers, lots of different types of applications and therefore, lots of different types of needs. And we will continue building new, as we call them, primitive services like silicon and chips.

At the same time, we think it’s really important, and customers are asking us more and more, to provide them with higher-level abstractions on top of our services. Again, not instead of — it’s in addition to releasing these powerful primitives.

Furrier: Examples?

Amazon Connect is a great example that’s out there today. It’s a complete call center solution. The demand is incredibly high. The service has really been on fire. You’ve got examples like Barclays, who at the beginning of the pandemic stood up a 6,000-person call center in 10 days. In fact, during March and April of the pandemic, there were 5,000 Connect contact centers set up by customers. And the benefits of Connect are amazing, on average 31% lower cost, 20% less supervisory time required. And a lot of other really impressive productivity gains.

In the coming years, I think you’ll see us continuing to look at horizontal use cases such as the call center. 

What other horizontal use cases are big needs for enterprises or other customers? We’ll undoubtedly try and help by releasing higher-level capabilities that both have brand-new functionality as well as bundling up, if you will, some of our existing capabilities. We’ll continue to look at opportunities beyond call centers.

Vellante: Can I play that back? It’s not a change in strategy. It’s a selective deployment of solutions that abstract those primitives based on, I think I heard you say, the horizontal total available market essentially, right? The application that the customer needs there.

I’m not sure I’d say it either is or isn’t a change in strategy. I would say more that we have over 200 very powerful and important services for customers [and] we continue to go in that direction. At the same time, we will accelerate the number of horizontal use cases that we directly can address, like Connect, with higher-level capabilities.

We’re already doing it, but I’m saying, you’re going to see more of it. It’s not some fundamental left turn in strategy. It’s more like, hey, this is what customers are demanding from us. We do 90% of our product development based on what customers are demanding. We can do that because we’re working with all of the most sophisticated cloud customers in the world. They are the best source of what needs to come next. And this is one of the things for which they’re consistently asking that we innovate.

It sounds like it’s the fruit from the labor of having all the cloud services.

That’s exactly right. If you just did that, it’d be like trying to ice a cake and then build a cake underneath the icing. You got to start with the cake and then you ice it.

The Connect phenomenon was very situational. COVID forced that. What’s going to be enabled next? What’s the next Connect?

We’re looking at a whole bunch of those horizontal use cases. Things in marketing. We’re doing a lot with manufacturing and industrial capabilities now. It wouldn’t surprise me at all if customers continued to press us there.

I think if you just look at all of the different horizontal use cases that happen, particularly inside of enterprises, we’re just talking intently with customers. And I know there will be more and we’re just figuring out where do customers need us to go next.

Then there are the industry vertical cases. And more and more you’ve already seen from us. We’re releasing industry-specific capabilities. You could call it purpose-built, not custom. Because we don’t want the cloud to be custom, then you destroy all the economics and even impact things like operational excellence and probably security at the end of the day. That’s an important tenet. But we are doing more and more that’s purpose-built for specific industries.

Vellante: What are some examples of that? (NEW QUESTION TO BREAK UP COPY)

You’ve seen HealthLake, which is a data lake for the healthcare industry. You’ve seen FinSpace, for the financial services industry. You’ve seen us work with large telcos and new entrants of the telco space. You have DISH to deploy native 5G network in the cloud, [and] Verizon, with our Wavelength service, to provide 5G services to others.

You’ve seen in the automotive industry where we’re doing everything from working with VW to build their industrial cloud and redefine their manufacturing, to working with a lot of automakers on building the software to find vehicle emission.

You’ll continue to see us put the pedal down in developing those solutions, which are purpose-built for specific industries and solve for unique requirements of specific industries. We’ll put a significant amount of effort into that and have some amazing partnerships going forward.

Tomorrow: Adam Selipsky goes back to his Tableau roots and talks about the changing role of data management in the cloud and how AWS aims to apply AI and machine learning to all of its services.

Photo: Tableau