Web3 CMO Stories

AI's Remarkable Role in Transforming Customer Service: An Insightful Dialogue with Sarah Al-Hussaini | S3 E29 (At Web Summit)

December 22, 2023 Joeri Billast & Sarah Al Hussaini Season 3
Web3 CMO Stories
AI's Remarkable Role in Transforming Customer Service: An Insightful Dialogue with Sarah Al-Hussaini | S3 E29 (At Web Summit)
Show Notes Transcript Chapter Markers

What if you had a sneak peek into the future of customer service? Brace yourself for an eye-opening conversation with Sarah Al Hussaini, the co-founder and COO of Ultimate, as she illuminates the remarkable role of AI in transforming the customer service sector. Sarah takes us on a deep dive into how AI not only makes customer service more efficient and personalized, but also increasingly data-driven. With illustrative examples from companies like Get Your Guide and Clue, we'll traverse the landscape of AI's potential to revolutionize customer service experiences, offering instant and scalable support in every language and channel.

The game of customer service is being rewritten by AI technology. It's not only shaking up jobs and markets but also empowering customer service agents to take on more sophisticated roles. We'll explore this seismic shift, featuring technologies like ChatGPT and LLMs, and how they create opportunities for startups and smaller businesses to compete with larger players in the market. We wrap up our discussions with insights into the journey of growth companies in the AI market, emphasizing the need for speed, long-term planning, and learning from customer feedback. Whether you're a startup looking for your footing in the AI industry or curious about the future of AI in customer service, this episode promises a fascinating journey into the realm of possibilities.

This episode was recorded at Web Summit in Media Village on November 16, 2023. Read the blog article and show notes here: https://www.cmo-stories.com/1903080/14135969-ai-s-remarkable-role-in-transforming-customer-service-an-insightful-dialogue-with-sarah-al-hussaini-s3-e29-at-web-summit

Unlock the Best of 'Web3 CMO Stories': Exclusive Season 2 Infographics!

Download your free copy now! 

Sarah:

It suffers terrible results. People don't like customer service, right? It's very normal to complain about it. It's a social norm, right? It is a trillion- dollar industry and it grows at 9% a year.

Joeri:

So hello everyone and welcome at the fifth interview I'm doing here at the Websum mit. It's the fifth and the last one, and my guest today is Sarah Al Hussaini. Sarah is the co-founder and CEO at Ultimate and, yeah, that's actually the world's leading customer service automation platform. Ultimate powers the support teams of global brands, including Zalando, Gorillaz, Finner, and Zendesk, and is 150 people strong. Hello Sarah, how are you doing?

Sarah:

I'm doing great. It's great to be here.

Joeri:

Yeah, isn't it every year at the Web summit these vibes? For me, it's already the seventh time that I've been here and I always get to meet amazing people and speakers like you. So welcome to the show, Sarah. Let's dive directly in. You know, with everything that's going on in AI, how do you see AI, I would say, impacting customer service? It's a broad question, but I'm interested to hear your thoughts about that and think about personalization, efficiency, and stuff like that.

Sarah:

Yeah, absolutely so. Firstly, AI has been impacting customer service for a long time. Right, we founded Ultimate back in the end of 2016, right off the deep learning wave and, in a nutshell, what happened with that deep learning wave was conversational quality, for bots became a lot better and conversations have really been increasing in quality ever since. But with that deep learning wave, we had bots become a lot better and customer service was a natural application for the technology, because it was a place where conversations were already happening. So let me give you a bit of a snapshot on the customer service industry. Right, it is, customer service is suffering from a huge labor shortage. Okay, people cannot hire into their teams. It suffers terrible results. People don't like customer service. Right, it's very normal to complain about it. It's a social norm. Right, it is a trillion- dollar industry and it grows at 9% a year. So it's just in an unsustainable situation today, and AI is the solution to that, right, so what we believe and we are a customer service automation platform. That is where we are applying our technology we see that AI is going to be the primary means of providing customer support in the future. Okay, and so there's three things that AI does for customer support. The first and the most important thing is that AI ensures that you're going to receive support instantly and on your terms, right. Ai is infinitely more scalable than human agents. It works around the clock in every single language, it responds instantaneously and it operates in every single channel. And so, to give you an example, one of our customers got your guide right. They are the world's largest marketplace for travel tours, and so a really common case for getting your guide will be you can't find your tour guide and it's 5:45 am and you're on the edge of the Sahara Desert okay, and that's stressful, and so a really bad customer experience for that customer is oh yeah, customer support is not open at 5:45 am, or that, oh, they only have an email support channel, and so you have to send an email and then just sit there refreshing your email inbox. But what is a great customer experience, then, you've got to ask yourself. A great customer experience, firstly, is that you're able to use WhatsApp or whatever app that you're using when you're traveling, that you're able to ask it in every language. Get your guide worked with. You know customers from around the world, you want to speak your native tongue and you want to have a response instantaneously, and so that's what AI can help you do, right. Two other things, to be very brief, because I think obviously the instant and the on your terms is the most important. But the second is AI makes customer service a lot more data- driven. Right, you're able to finally get insights on what's going on in your support to improve your products and services. So Clue, a FemTech company, a cycle tracking app, millions of customers around the world, users, and when they started with us, I think they had an automation rate of over 70% and we were so proud. We were like, oh my God, ultimate is the best. And then after a few months, the automation rate started dropping 50%, 40%, 30% and we were like, oh my God, what's going on? We get in touch and it turns out what they'd been doing is tagging the cases with the sorts of issues like product- related issues very specifically and feeding that back to their engineering team and improving the app and therefore support was going down. And that type of proactivity and support is only possible if you have the data and the insights to do it.

Joeri:

Exactly, makes total sense. But then also the AI gets in better because it's learned from the data insights and you can as a human, interpret it, but also AI can learn from that. Because that was when I was thinking, when you were explaining all of that this sounds wonderful, but

Sarah:

Firstly, absolutely right for a case like that. If you're thinking about one of that's one of get your guys' most common questions, I can't find my tour guide, and so, for a case like that, you would have a lot of data. Firstly, lots of examples of that being asked and your customer support. The AI would be very accurate on that case. And, secondly, you would have really thought about what is the process for handling this and you would have mapped that workflow out into Ultimate, and so you would have a very high degree, very high degree of accuracy and understanding, right Intent detection and then also in resolution, and then, finally, what we always, always say to our customers everything the customer is, it's not automation at all costs right. So the average automation rate for Ultimate in 2022 was 45%. Now, with Gen AI and I know we're going to talk about Gen AI, the customers using Gen AI with the workflows is more like 60%, which is very exciting, but still, let's say, you know it's hovering around 50% so that 50% that goes through needs to be escalated to an agent. You need to allow your bot to fail gracefully. If it can't fail gracefully and your customer is stuck and you go straight back into terrible customer experience.

Joeri:

Yeah, I love that because people are sometimes afraid about AI losing their job. But if AI can make their life better for them and for their customer and it's the combination, I guess how knowing how to use AI and how to handle AI, that it helps, really helps you. And so, Sarah, because it helps people. It helps making their job lighter. I would say you'll be more efficient, but it also helps with their job satisfaction. Do they have less stress, or how is that?

Sarah:

Yes. So this is a great question and I've actually been talking about this. It's a very sensitive issue, right, in the concept of automation and AI. What is the impact on jobs? What is the societal impact? And so, first, some stats. I'm very happy on the stats, I love it. Some stats 4% of employed people in Western Europe and in North America are customer service agents. 4% it is one of our biggest employers, right? 70% of customer service agents are women. Okay, and it is a job which has it's challenging from an upwiz mobility perspective. Okay, because there are so many roles and that's what I said. There's a labor shortage at the kind of more entry level end, and then it really does. There are much fewer jobs at the top and actually, interestingly, of course, top jobs are dominated by men still, which just makes no sense. Makes no sense. So that's the industry that we're dealing with, and so customer service, the job of the agent, can be incredibly stressful You're dealing with. It's very normal to be dealing with a backlog. I've had customers come to me and they've got 70,000 tickets in their backlog that they have to be resolving and dealing with, and what that's like for an agent is from the second you get into work in the morning, you are behind, you work flat out all day in what feels like a never- ending task that you have no control over, and then you finish your day and you kind of go back to the same cycle. You're dealing with people, with problems, which can be incredibly stressful, and it can be very repetitive Some of the stuff. It can be very, very repetitive for the agent. So it can be quite a stressful job. And the final thing to add is agents often have to work overtime to deal with this backlog, support the customers. Because you go into customer support if you care about your customers, you care about people, you're empathetic. That's when you're successful in the role. And so during peak seasons, which for many companies can be six months out of a 12 month calendar year, people work overtime. And I've heard more often than I ever should about customer service teams not having additional budget so they don't get paid for that overtime. Right, and that's what the industry's like. So it's a tough job. So when we come into businesses firstly it is 99% of the time the agent team that is the most excited about us coming in we take this repetitive burdensome really. Where's my order? Where's my order. Where's my order? Over and over again, we take that, okay, the basic stuff. And so what ends up being escalated to the agent? It is the sophisticated tasks and it is the revenue opportunities. And so you, as an agent, turn into someone that provides a white glove service. So you are proud of your work. You aren't dealing with the backlog anymore, right? Otherwise we've ultimately failed at our job and you should be able to command a higher salary, right? Because everything that you deal with is a revenue opportunity or a sophisticated case. And then the second thing I would say is, if anyone that's listening to this works in the customer experience industry is, AI is changing customer experience and it's creating completely new roles. And what we see in businesses almost every single business that we enter within the first six or 12 months, someone will be promoted to some sort of product owner of Ultimate, some sort of automation manager, because, remember, we're taking 50% of support. It's a job, and these people again come under higher salaries and are visible within the company as a technical role, an AI specialist. Sometimes you're the only AI specialist in your business and you've done that from a completely non-technical role, and that's a beautiful transition to be able to go on.

Joeri:

Wow, that sounds amazing, All of these possibilities that we have with AI and also the way that you explain to the passion and results of the impact that you can have with your technology. You mentioned it. We also will talk about GenAI, and there is ChatGPT, which is coming out with all these new possibilities, and I guess it opens opportunities for smaller businesses and startups to fight against those bigger companies that take a lot of the market. Any thoughts on that or maybe advice for smaller startups or companies on how to use AI for their growth or for their competitiveness?

Sarah:

Absolutely so. I love this topic. I think it's been the most interesting time ever, ultimate, the greatest year, with the release of ChatGPT and LLMs. And that's so interesting to say because at the start of company founding, it's so exciting. So, to have this during your growth phase, this innovative year, this fast-paced year, is just the greatest gift. So when a completely new, disruptive technology is released, it levels the playing field. Nobody has incorporated yet, nobody's using it, and it is light years best than anything else that's out there. And so you're absolutely right, it creates a huge opportunity for new market entrants if you're able to capitalize on it. So, from an ultimate perspective, we want a new market entrant. Right, we're not a big, slow-moving enterprise, but we are a growth- stage company. We have almost 200 people in the team and product roadmaps and the whole plan business plans, and what we needed to do was level the playing field. We needed to execute upon this because it can be a huge opportunity, like I'm saying, really level up our product, or, if we don't execute upon it, it can also be a huge risk. And so how we handled it ChatGPT was released early December 22. And so the leaders and the R&D leaders we all went away over Christmas break and played around with the technology and we were like we're all going to play around with it, we're going to figure out, we're going to think about what we can be using this for. We come back in January and the only thing that we know and we know two things actually Firstly, that it's going to completely change the game and, secondly, that we have no idea what we're going to use it for. We just know that we need to start building with it. That's how disruptive it was. It's not like a single- use case. It has to be. It's going to change so many different areas of the platform. So what we needed to do was get our business innovating on it. We needed to get the business shipping. It's raising, with customers shipping again, ship feedback, build more, and so on and so forth. So what we believe what we're talking about internally is the compounding effects. We need to move fast, so we need to launch quickly and then we need to learn fast than anybody else and then in 12, 18 months, we'll have built those compounding effects and reestablish market leadership. Right, because I said it was a leveler playing field, and that's what we did. We killed our product roadmap at the start of the year, we completely bent it. We killed the concept of company goals and we shifted completely into an iterative, like two week or one month learning cycle. And we are just now just actually three days ago reintroducing the concept of product roadmap back into the business because we've learned enough. We released ultimate GPT in April of this year our LLM capabilities and six months on now we have over 100 paying customers integrating LLMs into their workflows for support and, like I said, they're seeing automation rates up by 50%. So it's a really exciting time. So, bringing it back to your initial question of how can startups compete, I'd say two things. Firstly, it's a sprint, so you have to move fast. Moving fast isn't about moving fast to acquire market share. I've heard that said a lot on a lot of different stages. It's not about market share. If you have a better product in a year, you will win that market share from the person that failed to execute. It's about moving fast in terms of R&D, development, shipping and learning the compounding effect. And the second thing that you have to do is you also have to realize it's a marathon, so you need to be able you put your business in a situation where it's resilient enough to last, because we are only at the very beginning and you need to know that you're gonna survive 12, 18, 24 months. It's not a great funding market out there, and so I don't think the big enterprises are gonna win this at all. I think that, logically, they should. They have the resources and everything that they need. I just think that information moves too slowly and they're too risk averse. The super, super small companies that are just founding now. I think that they can move very fast, but again, they need to realize it's a marathon, right. So I mean, obviously, from where I'm sitting and seeing what we're doing at Ultima, I'm most excited about what we're doing, but I actually think it's the growth companies that were already in AI that are using it to replace parts of their architecture and move so much faster. That's where I'm most excited.

Joeri:

Yeah, a little bit like agility move fast having results. You mentioned it is not an ideal funding market, no, so did you have there any issues, challenges to get funding? Or is it just because you have such a good story that it works?

Sarah:

So I mean we were lucky that at the end of the last year, when the market, the tech market, crashed, we had enough cash in the bank to know that we were gonna very comfortably be able to like ride through the next few years. I shouldn't say very comfortably, I should say that we were gonna focus on efficiency, right, and we were gonna focus on efficient growth, which, as COO, I love. I hated the fact that we were like hiring like crazy beforehand because everybody was doing it and salaries were growing like too much inflation in salaries, especially in certain markets in Europe for certain R&D talent. So I love this efficient growth like manifesto, but so that's the first thing we were able to know that we wouldn't have to fundraise. This year has been such an amazing year for AI and such an amazing year for Ultimates that, let's put it this way, when ChatGPT was released, what it did was completely renew trust in AI. When I founded Ultimates, bots were so on call and now they are so cool and so our contact sales inbound demand. So a contact sales submission is when someone comes on our website and books a demo, right, they want to speak to a salesperson. That grew 18 times between January and June of this year, right. So when suddenly you're running a business that has so much more inbound demand it's so much more and you haven't needed to hire to do that, you haven't needed to increase ad spend all of your efficiency metrics start to move into a much, much, much healthier place and you're able to start making very strategic. So we're still not hiring like crazy, but like very strategic investments into key roles, especially like an R&D, to ensure that we can execute on yeah, on the opportunity in the market. So I know everyone's position will be different, but that was the position that we found ourselves in.

Joeri:

Yeah, you are before the wave yet, ChatGPT but you are able to catch the wave. That's what I always say to people like you. Know new technologies? Okay, you see it coming and you are thinking why should I act now on that? Why should I be innovative? Well, those are the reasons. Is it a good one, or do you prefer another question?

Sarah:

No, it's more that there aren't really ethical considerations.

Joeri:

You should. So we are almost at the end, Sarah, but because I know what webs meet where we are, they are kind of severe. But I'm curious how do you see the future? And maybe, what are you most excited about when it comes to SaaS, AI companies and the things that you are building?

Sarah:

Amazing, yes. So firstly, thank you for having me here having the best time. Great question to end on, I think, to answer that we have to ask ourselves what have we fundamentally done with LLMs, right? What has this technology fundamentally done for the world? And from where I sit, we have democratized human understanding, okay. So what do I mean by that? So if you're building an AI product, all you're trying to do is capture human understanding. So, if it's an LP product, you're trying to understand the intent behind an utterance, right, the meaning, the understanding behind an utterance. What is the subject of the utterance, what is the goal, what is the sentiment? And to do that you're building these kind of intent models to understand what is going on in the data, and it's just a lot of work to be able to do, right? What chat GPT has allowed us to do is it is so context rich that it understands you immediately. So bear with me. You've played around with chat GPT I think most people have by this point and you can tell that when you write to it, it understands your meaning, right, it might not give you a perfect answer or even the right answer, but it understands what you're trying to say. And, like these models GPT-4, for example. It's already multimodal, okay, so it can watch you through your webcam. It can see that you're present, it can see the facial expression that you're pulling. You can use your iPhone camera or whatever camera I'm not advertising for the iPhone. You can use your phone camera and you can show it any scene you know People playing football, a crowd, a bike, an animal and it understands what it is seeing, and so the possibilities that you can build on top of that are endless. Right, this is when machines can interpret the world the same way that we can through texts and images. It is gonna be this massive unlocking for SAS. So what I believe is that every application in the future is going to be an AI application. We just need to see how we're gonna be applying the technologies, and yeah, that's what I'm really, really excited. I mean, I know exactly how Ultimate is gonna be applying to solve the customer support problem, but I wanna see what every other innovator is gonna do.

Joeri:

Amazing. I am also excited about all these possibilities and everything that is changing so fast. Everything you see some new possibility like, even if you only look at chat GPT. Sarah, this was so insightful. Thank you for being a guest on my show.

Sarah:

Thank you so much for having me. I had so much fun, honestly.

Joeri:

Guys, this was actually the last episode that I recorded at the WebS ummit with Sarah Al Hussaini and, yeah, if you liked this episode, I'm sure you like it and you think this can be really useful for people around you other entrepreneurs, startups, maybe friends or family. Be sure to share this episode with them. If you're not yet subscribed to our podcast, this is a really good moment to do it. Absolutely, and of course, I would like to see you back next time. Bye, bye.

How do you envision AI influencing customer service, particularly in terms of personalization and efficiency, given the current advancements in AI technology?
Is current AI in customer service functioning with 100% accuracy, even in challenging scenarios like being in a desert and speaking in a specific language?
How does AI contribute to people's job satisfaction by making their tasks more efficient and reducing stress in the workplace?
What are your thoughts on how emerging technologies like GenAI and ChatGPT offer opportunities for smaller startups to compete with larger companies, and do you have any advice for them on leveraging AI for growth and competitiveness?
Did you face challenges securing funding for your project in a less-than-ideal funding market, or has your compelling story played a significant role in overcoming potential obstacles?
How do you envision the future for SaaS and AI companies, and what aspects of the innovations you're working on are you most excited about?