00:00You rely on AWS for basically all your infrastructure and you are an example of an enterprise scale software company, many of them in this building asking what's the value add that AWS is giving. You are taking some of their technology and you are giving to the market what we're calling AI first contact center as a service. Is there really a value add with that, Tom?
00:22We really see the Amazon relationship as strategic for our customers. First of all, we have a three-legged stool in the relationship. First of all, Amazon provides all of our infrastructure. We've been a customer for over 10 years. They do it at better quality, better security, at a cheaper cost that we can pass along to our customers. Second thing is that we are working with them as the Amazon Connect contact center is a key component of our overall solution. We have a go-to-market partnership with them. A lot of value add there.
00:48And third, we're diving into AI models. We are a user. A lot of our workloads, a lot of our AI business is through Amazon Bedrock and Amazon Nova. So we see it as a three-legged stool. Amazon's adding a tremendous value to not only our business, but most importantly, our customers.
01:04They have brought out a new generation of Nova, Nova 2, one variant of that. Omni can basically take different mode inputs and output in different modes. Is that actually useful to you?
01:15We're using them for workflows, things like transcription, sentiment. We're driving a lot more of our large language models through Nova right now. And so we're really bullish on the long-term prospects for Nova.
01:29We see the Amazon investment. They've made tremendous progress the past year with Nova.
01:33And if I may, let's go back to that first piece you talked about, a lower price point. You know, that is the AWS sales pitch, particularly for their own silicon and server designs. You've probably gone out to the market and said, okay, what's in the best interest of Zendesk?
01:47Was that a reasonable conclusion that you came to, that Nova and what AWS has to offer is better than doing multiple models from multiple labs?
01:56So we do use multiple models. But we think the great thing about Amazon Bedrock as well, we can go to one place for most of those models.
02:04And I'm sure they're going to continue to expand the Amazon Bedrock portfolio models. And so that's the great thing about Amazon. You can go to some of these specialized large language model providers, and you can go to models like Nova all in one place.
02:17And so I think that's a really strategic and competitive advantage for Amazon.
02:20So, Tom, with respect, there'll be large parts of the Bloomberg tech audience, like, banging the fist on the table, saying, stop speaking in abstract, Tom.
02:29Contact center is a service. Explain materially how AI improves the contact center, how it changes it.
02:38You know, all day long, AWS executives have been telling me, the story here is a transition from AI assistant to an AI co-worker.
02:44What are we talking about in Zendesk's case?
02:48So we've got a point of view that over 80% of interactions over the next five years are going to be completely solved by AI.
02:55That's AI agents. That's AI.
02:5880. 8-0.
02:588-0. 8-0. And we're seeing customers right now get 8-0.
03:02So our top, top leading customers, 20% human agents will be assisted by AI.
03:08So 100% of interactions, but 80% are going to be completely solved.
03:11And so it's not hype. It's happening now in some of our most advanced cutting-edge customers, and it's going to go through the whole industry over the next five years.
03:19Who benefits financially from that?
03:21Well, I think a couple people benefit financially from that.
03:24First of all, customers benefit. That's the most important thing, consumers.
03:29When you're able to have a low-friction customer experience with a company that you're talking to, it's a fantastic experience for you.
03:36So you benefit. It's lower cost for the company.
03:38And they're able to go reinvest some of that savings back into making the customer experience much better.
03:44And that's what we're seeing cutting-edge companies doing.
03:46They're actually not taking the money to the bottom line.
03:49They're creating a service dividend and focusing on how they can serve their customers better.
03:53It's not going to go away, the question of will my job be displaced by AI, or will it be replaced by AI, or is it sort of accretive to what I do?
04:04At one end of the technology spectrum, we're talking about call centers, humans building inquiries from customers.
04:10At the other, you've just said 80% of those interactions will be agents.
04:13Inevitably then, some roles will be eliminated.
04:16So what we think is going to happen is we're seeing the amount of interactions start increasing because consumers are much more likely to go to companies.
04:25And so all the repetitive, easy tasks are going to be solved by AI, but that's going to free up humans to do different things.
04:32And so I don't think roles are going to be eliminated.
04:34We think actually roles will increase, but what's going to happen is they're going to be doing different kinds of work.
04:39The other question that's come up all day is how difficult is this?
04:43You know, in some sense, we're talking about a specific Amazon product, but it's agentic AI as a broad term.
04:50If you run a massive enterprise with lots of different teams doing different tasks, is it as simple as switching, you know, flicking the switch and saying, okay, great.
04:58All these agentic tools just work now, or is it a massive structural undertaking for a company like yours?
05:03It depends what you're doing.
05:05I think there's everything from really easy to really complex.
05:07One of the things that we have is we have 30,000 customers with something called a knowledge base.
05:12They have all their policies, procedures, et cetera, in that.
05:16And you used to type in, how can I go return something?
05:19And you get a bunch of blue links back.
05:21Agentic AI is really, really good right now at taking that into a specific answer like Google Gemini, a very, very specific answer.
05:27But when you think about agentic overall, you're changing procedures, you're changing tasks, you're dipping into other systems.
05:36It's not simply a turn it on.
05:37You have to know prompt engineering.
05:39You have to understand workflows of a company.
05:41And so there is a big service component when you get into complex AI and agentic AI for sure.
05:47Tom, you've just told me that all of this is real and tangible and deployed in the real world.
05:51So you'll forgive me or maybe not for going to the AI bubble discussion.
05:56You know, the calculus for about two years has been very simple.
06:00How much money did you put in as an investment?
06:02And what kind of top line growth do you see as a direct result of that investment in AI?
06:07What's the result of the calculation for you?
06:10Yeah, for us, we put a lot of investment in.
06:13We'll end this year at about 200 million of AI, ARR, up from zero two years ago.
06:18And so we are seeing the growth there.
06:21And that's all new products.
06:22Some companies are saying a percentage of our revenue that we existing have and we send a feature.
06:27This is new AI product revenue that's gone from zero to 200 million AI, ARR over the past couple years.
06:33I don't think it's a bubble in the mid to long term.
06:36I think there's an old proverb, people overestimate the impact of technology in the short term, but they underestimate it in the mid to long term.
06:43I don't know about valuations for company, but what I see every day right now is AI impacting our customers and our customers' customers by automating low value tasks.
06:54A year from now, how big is that ARR, do you reckon?
06:56I think it's going to be three, four, five hundred million dollars a year from now.
Be the first to comment