This website requires JavaScript.

New tools and new career paths in customer service, with Sylvain Mlodyszewski

Customer service28 MIN READOct 5, 2023

This podcast episode looks into some of the new opportunities emerging in the customer service career space. 

Sylvain Mlodyszewski is our guest, he harks from Ultimate – a Klaus partner which is a force for good for customer support, helping companies supercharge with AI. Sylvain and Grace talk about how the lack of standardization in AI roles means that there is fertile ground for new job pathways.

Listen in to learn:

  • Why AI might solve the problem of talent retention in customer service. 
  • Why a customer support expert getting involved in AI brings value that will reach every part of the organization. 
  • Which new career prospects Sylvain has observed through businesses employing Ultimate in customer support.

Sylvain Mlodyszewski is Director of Partnerships at Ultimate. He knows a lot about how different companies and people are successfully evolving with these new technologies.

Grace Cartwright is your host. She lives and breathes content at Klaus, and has long been exploring the realm of customer service AI.

You can read the podcast transcript in full below, and we also recommend:

  1. Looking into this LinkedIn course: Career essentials in generative AI.
  2. Following Oliver Molander for AI news.
  3. Checking out the Conversation Design Academy‘s courses. 

Transcript: ​​S3E3 

Grace: Hello, you’re listening to Episode 3, Season 3 of Quality Conversations with Klaus. I’m Grace, your host. 

The topic at hand is AI, a subject that unless you’ve been hiding under a rock, you’ll have heard much about. However, we’re interested in how it will affect customer service careers. This week, our guest Sylvain Mlodyszewski, helps us look at new tools in the AI space and some new career paths that will unfold as a result.

He hails from Ultimate, a company that helps other businesses use AI to enhance and automate their customer service. So he knows a lot about how different companies and people are successfully evolving with these new technologies. Let’s hear what he has to say.

Sylvain: Hi, Grace, and thanks a lot for the invitation. I’m Sylvain Mlodyszewski. I’m Director of Partnerships here at Ultimate. So I’m responsible for our tech partnerships as well as our channel partners. And making sure that we grow like a strong ecosystem of partners that can either complement our offering or build a service offering on top of our platform.

Happy to be here.

Grace: Great to have you because Ultimate is really a company that’s forging quite a new path in customer service and in AI in particular, which is, of course, why we’re talking today. So the first question I have for you: you’ve been at Ultimate for a couple of years now. 

How do you see that the industry has changed over that time?

Sylvain: Yes, I joined Ultimate about two years ago. Truth be told, I was actually new to the customer support industry, so I had to learn everything from scratch. I just started diving into articles like reading blogs, going on forums and Slack communities to really understand what were the challenges in customer support, how it was to like work in customer support. And also obviously understand about Ultimate and what our value proposition was.

I think one thing which is interesting is that Ultimate was founded in 2016. So Ultimate saw the early rise of chatbots, then all the negativity that it brought. Because I think especially before that time you didn’t have proper AI chatbots, like all the chatbots were more like rule-based chatbots.

I think the expectations of the market were almost like where the expectations and technological capabilities are today. So basically the technology couldn’t meet the expectations of the customer. And it created this impression that everyone has. 

When I started my job, I was like, yeah, I’m going to work for a chatbot company. And everyone around me was like, Oh, those chatbots where I never get an answer and they never understand me. 

And, there’s all those misconceptions around chatbots. Then starting to work at Ultimate, I started slowly realizing that it was not necessarily the technology. When I joined in 2018 or 2019, Ultimate was at heart an AI company. We built our own AI model and we tailor it to each company and then that creates possibilities, so that we can understand any conversation in any language. 

So I quickly realized that basically it was not necessarily the technology that was in itself poor, it was more the lack of investment in building nice solutions. And that’s what we’re all about here at Ultimate. 

Ultimate is an AI platform, which means that we enable our clients and partners to build elegant chatbot solutions for their own challenges. But obviously the solution will only be as good as your investment in it will be. And aligned to this, there is also this misconception in AI that it’s something that you turn on and it just works.

Whereas AI will keep on providing value for as long as you feed it data and relevant data. That’s actually one thing now with generative AI that we see more and more coming up. People are starting to understand across the board that it’s not really about the technology. The technology is there. 

It’s about the quality of data that you put into it, and the recurring aspect of nurturing and maintaining your good set of data. Making sure that you make quick adjustments to either the flows or your setup. Making sure you know that your AI setup, whatever it may be, can also not be a chat, but can be in another aspect, area, but that it stays relevant and it stays optimized.

At the end of the day, AI is a lot about optimization. It’s not something that you just flip the switch once and it works.

Grace: Absolutely. I love that description because it really puts AI as a limb that’s working for you. It’s not something that’s outside of the business or even outside of the team. It is something that is 100% part of it – and an evolving part too. 

You talk about people using it to work for them, but it needs continuous investment. So how do you see customer service careers changing alongside that?

Sylvain: Customer service carriers have always evolved. Maybe now we see more and more changes because the pace of changes has gone faster. I also think the benefits that you can get from automation … basically the sky’s the limit. 

You can say tomorrow the chatbot will take 100% of my incoming messages, but maybe that’s not like the smartest choice to make. Every time that you see those new technologies, you have to basically do a balancing act of figuring out, okay, what is the right choice for me? How much do I want to automate so that it makes sense for my customers? So that it makes sense for my queries? Then create those synergies between the customer service organization and the automation solutions that are here today.

I think what has changed is just the access to technology. Recently we were seeing that large enterprises had access to very advanced technologies like natural language processing technologies for quite a while. Actually, since 2018 or 2019, when you had the release of the lapse model from Google.

Those large enterprises have invested for a long time, but in very custom solutions. I think what we have seen more recently in the recent years – and what we’re playing a role in – is trying to level the playing field. So that enterprise companies but also smaller companies can reappropriate those technologies and with the help of platforms like Ultimate. They can keep the expertise & knowledge. Build & maintain those solutions without the need of having a machine learning engineer or analysts that go into the details of the data and maintain those very custom solutions, but have a no code platform, where they can do everything and keep the knowledge in house. In the sense of, for example, within the customer support organization.

So I think that’s a big change that operated in the past years, which opens a lot of new opportunities for these organizations.

Grace: Yes, I think with the rapid change of pace just in the last year, I would say, there’s this fear. If you don’t have machine learning engineers on-site, if you don’t have a fully-fledged data team, does this mean that you will miss out on something that is certainly becoming quite essential for ambitious teams across the board?

Yet what you’re saying is actually the democratization of this with the right tools means that it is open for all, right? You don’t have to have AI at your core in order for it to work for you.

Sylvain: Exactly. And I would even argue that certain companies that do not have a background in AI may now be at a stage where they want to double down, and use some of their internal resources to try and build solutions themselves. Obviously I’m a bit biased because we provide a platform that does this, but I think it’s a very interesting time. 

We see some companies that decide to take a leap of faith and try to build it themselves because, somehow, they have technical resources internally.

We see some other tech companies, that would be in like financial services or other areas, but have already the required capabilities in house. But obviously that will still be a very small amount of companies out there. So I think for most companies, these doors would be either closed or would require a tremendous investment into recruiting the right people, and so on.

So I think obviously having a dedicated platform, that basically does the hard job of translating those technological capabilities into features, that you can just drag and drop or toggle in and off and move around and select like this – I think it opens a lot of opportunities.

Another part, which is interesting, I think we may come to see is that there is still no organization of the market. As we see with our users, there’s not a standard term for what the job description or job title should be. 

We have more and more users and we see that they are getting fast-tracked in their careers. But depending on the industry they’re in, depending on their customer support organization, they may not have the same titles.

That’s really exciting because if you compare that to a more established function – sales is a good example – sales has been perfected and really turned around in all possible senses. But today everyone has an understanding of what an account executive is, what kind of role an account executive does. There are standards that are understood everywhere about these roles. 

We see that for AI related roles, it’s still a bubbling space that still needs to be standardized.

Grace: Absolutely. It’s something that over the next few years probably we’ll see much more. We should see many more solid jobs emerge that would be more dedicated to that kind of role, right? 

What new career paths do you see starting to emerge already?

Sylvain: We do see a lot of career paths. It’s important to mention that, historically, customer support of organizations have had a hard time retaining their best talents, and offering them a logical next step.

And I think that’s the main thing now with platforms like Ultimate, we create a space where customer support organization leaders can promote their best performers into automation managers, conversation designer, chatbot manager, and so on.

It fast tracks those individuals into successful positions. 

The advantage for these positions, and we come across new ones every day, first of all, is that those positions are very strategic. You need to not only work with customer support, but be a bit cross functional. Maybe work with sales if you have proactive messaging and support path that would lead into an upsell or keeping a customer. So you would need to be in good touch with sales. You would also need to be in touch with marketing to make sure your automation dialogues and chatbot experience remains on brand with your company guidelines.

We see that those positions are very exciting and just bubbling. On our side, depending on the customer size, we may have entire organizations that have a team of five people that are dedicated solely to the management of the automation, but at an international level. You have all the geographical breakdowns of regions or even types of chatbots. So it opens a new field. 

To go back to your question, at the base is that those automation strategies – and not just chatbots in general, but AI capabilities – we see that it’s not black or white. It’s not either or. It’s more about thinking how can we use this as an incremental help to augment my productivity or replace repetitive tasks. 

So basically the advantages that we see in customer support can be translated into other fields. 

Interestingly, I saw the news last week that, for instance, the user number of ChatGPT as a public tool had decreased by 10%. The reason for that is because we see ChatGPT as a tool you need to use it as an increment to summarize. If you’re a copywriter, let’s say you, you don’t want to use chat GPT for everything. You want to use ChatGPT to summarize a conversation or isolate bits and texts. Or, give a text and ask ChatGPT to isolate a few terms. You need to have an end in sight. And so I think that’s why you will not completely replace humans.

Maybe I’m digressing now, but we see a huge rise of people understanding that now prompting is actually the key. So basically you need to have a good prompt for the AI in order to get the right results, because the AI is not going to do it for you.

So I think this is a small thing, but that actually underlines the fact that AI in itself is not smart. It’s still at the end of the day the human behind it’s maneuvering that is smart, and can get either good results or bad results. And now, again, this drop of users can also be correlated to the fact people start to realize that the AI do hallucinate, they do provide wrong results and they cannot be blindly trusted. So you need to have those guardrails.

Grace: It’s like we’ve all been on a learning curve through ChatGPT, probably of its capabilities, but also how much better we can be as users when we are very specific. My prompts when I first started using it were like one line long, whereas now if I want it to help me summarize something, I will insert so many adjectives about the tone I wanted to use.

And that is just a small example of how AI tool adoption can work on a much larger basis. It goes into what you were saying about needing people who are going to be feeding the best information or the most information to the chatbots possible for them to work for customers, for them not to be these old-school, more automated bots that are only able to provide very limited answers to questions. It needs to be a much greater understanding of what the company does instead of this universal tool for everyone. It’s very tailored, right?

Sylvain: Absolutely. And furthermore, that obviously with the rise of open AI and generative AI and also those large language models, opening their capabilities through API to pretty much anyone and any company to start building APIs, you start seeing a large increase of new companies being created – surfing the wave.

At first we were looking at this. How is this going to develop in our markets? And right now we actually see that obviously it’s not enough. It’s not just building this API. We’re really happy to have this strong platform in the background, so that we actually are able to use the best of generative AI capabilities, but embedded into our platform. At the end of the day, if your objective is just to turn on generative AI, no company, no organization will turn it on for all of their conversations. That’s not the end objective. 

So you need to have this platform in the background that has like a holistic view of everything that enables you to build integrations, not only use generative AI, but also use NLP.

I think that’s an aspect that’s very interesting because, at the end of the day, you want the hybrid approach. Because with generative AI, no matter what you do – you can do some tweaking, you can change the tone of voice – at the end of the day, the concept behind it is that you don’t control the output.

Even with one same question from one day to another, you don’t know for sure what the outcome will be. So you need to be sure that you want the generative AI to answer only on questions that are not critical, that are not important for your business. 

And for the questions that are important for your business, you actually want to tailor a very well crafted path and have those right integration in the background. And so what the future will look like, and what we offer at Ultimate, is this hybrid model where you’re able to embed parts of the generative AI into our normal flows with our own AI that is trained on the customer’s data, to build this best of both worlds approach.

What we see now is a lot of new companies, obviously that were not created in 2016, that they don’t have a platform to do that. So there’s this glass ceiling that is happening.

But at the same time, I think those companies probably make sense for certain customers that just want the automation and just want something very standardized, where they can dabble into the world of automation.

The fact that you have new incumbents coming in is, overall, very positive for the market. It raises awareness, and despite AI having been around for, technically speaking, 50 years,it’s misused. But what’s real is that there’s still a huge proportion of the market that needs education, that needs to understand better how they can actually leverage those technologies in their best interest.

Grace: Right. Instead of just being a bit overwhelmed with everything and feeling like you’re acting rashly by just needing to have it, it’s coming at it with a much more strategic approach. To make sure that you are striking the right balance with how you are applying it. 

To get into more specifics – what specific skills or knowledge then would you give as advice for people if they want to be able to come at this with a more strategic approach?

Sylvain: I think there are three aspects around AI. You have the brains, the tools and the skills. 

The brains can be related to the AI technology, the AI engine, all the algorithms in the background making the magic happen. 

Then you have the tools – potentially using a platform like Klaus and/or Ultimate to help you achieve your goals. 

And those tools are where you will apply your skills. Your skills can also be divided in a few areas, because you have the functional skills of how to use such a platform, how to build the dialogues, how to figure out your quality control on your automation on your platform. So how to use these platforms can be easily acquired. And obviously we have paths to train our users.

And then you have another aspect, which is more related to copywriting. That’s conversation design. Conversation design is how to build imaginative flows that will first of all be on brand with your company, that will talk to your customer, but at the same time will guide them well.

I remember a customer success manager on our side told me once that ‘a well built flow could sometimes replace an integration.’ Because it means that actually you understand the customer better, and then you are able to guide them either through knowledge base articles, get them to the information fast without needing an integration there.

So the interesting part is that you have an aspect of both functional understanding tools, managing tools, but at the same time a very crafty approach as with copywriting. This is for managing a chatbot.

And finally, a very important aspect is integration – being able to build those integration with other systems. So with a platform like Ultimate, you can do it natively from the platform. And say, if you’re like an e-commerce customer, being able to finish the order number from a Shopify or WooCommerce, that’s super important because then you’re delivering value directly.

To the customer and as a chatbot manager, understanding: okay, in this part of the flow, I need to retrieve that information from that system. And being able to own everything is a tremendous help. Obviously, typically you still have different teams for managing that, but you have one chatbot manager that understands when to position those integrations.

So in a nutshell, I think it’s very interesting because it’s both a very technical role, but also a very crafty, very creative role because you have the copywriting aspect. 

And finally, at the head of that you need to have someone – or it can be the same person depending on the organization – that has a strategic like perspective and is able to liaise with other functions of the company.

Grace: A hundred percent. To really make the most out of any interaction, right? Because then you’re plunging as many resources as possible into it. 

I think that it’s interesting what you said, that it has to be someone quite creative, because I think that’s certainly true. You have to know the customer well enough to understand how to design that flow, but you also have to be aware of the potential that they might not be asking. 

And how to tap into that without overstepping the mark too much,

I think it’s a balance of knowing the tech and knowing the customer, but also, like you said, knowing just the entire company goal as well. And understanding where you want that conversation to go isn’t quite as simple as just designing a simple flow.

There are so many different considerations that are quite difficult to balance, I imagine.

Sylvain: Exactly. And actually my hot take on this is that because those technologies historically had to be owned by IT. departments, and maybe had been built-in (obviously you always find exceptions), that’s one more reason why those projects need to belong and stay within customer support organizations that are close to the customer.

They understand the needs of the customer better than anyone. And I do think some solutions are developed from a technical perspective very well, but we’re not really speaking to the audience, or do not come from a place where that the business actually needed.

And I think bridging that gap, bringing the technologies, and giving access to those technologies, through platforms to customer support organizations is a very important aspect.

Grace: Absolutely. It doesn’t matter how sophisticated that side of it is if it’s not actually serving the customer, which at the end of the day is what the goal is.

And how else do you think that organizations can support teams in this endeavor? Or can even support individuals who want to upscale for this?

Sylvain: We see a lot of trainings being created everywhere. You have some on LinkedIn, I think Microsoft partnered with LinkedIn on releasing general training and certification on generative AI.

We (at Ultimate) obviously offer training and certification for our users and partners. So we’re really proud of that. You see that it’s democratizing quite fast. I think there are open source training programs that are free to use for anyone out there. 

I think actors such as ourselves, we also have a responsibility to educate our customers and users of the platform, not only on how to use the platform but understanding what’s at stake – what are the developments.

It’s also interesting because you see news happening like every second week. You need to be aware of what’s happening, and you need to translate that to your customers and partners.

Grace: It’s continuous really, isn’t it? Especially at the moment, I think. 

We’ve come to what is a more quickfire round. What would you say are your very short actionable tips to someone who’s wanting to upskill right now in customer service alongside AI?

Sylvain: I would say keep up to date with the news, because I think that’s a relatively low effort versus the knowledge that you’re going to get from that. It sounds obvious, but not everyone does it – finding the right sources of information.

There are good newsletters out there that are very useful. There are also a few individuals on LinkedIn, such as Oliver Molander, on top of my head, who’s quite influential on LinkedIn and has very good content on what’s happening on a weekly basis. So keeping up to date with that.

Also, if you want to start looking into this field, I would say try to find a chatbot experience that you like, or that has been particularly well built and try to get in touch with the people who built it. We can obviously facilitate those introductions, but I think that’s great. Just talking to these people and seeing what their journey has been, to understand basically how you can upskill, yourself.

And finally on the conversation design piece, we partner for instance with the Conversation Design Institute, which is a foundation, and they actually provide specific and dedicated training on how to build meaningful conversations.

There are a number of organizations and a number of ways that you can actually upskill in this world. And I think, now you have a lot of sales accreditation, you can say Oh yeah, I’m a salesperson. I’m accredited and I’m very proficient in the medic method. I think like all those concepts and all those standards are going to come for AI and automation.

Looking forward to seeing that happening.

Grace: I know exactly what you mean. It’s hopefully going to get easier and easier for people to be able to access materials that make these technologies much more digestible.

And for the final question, which is controversial.

But do you think AI is coming for our jobs?

Sylvain: I think that’s a good question, which is always a good way and not a good way to finish an interview because we can ramble for hours on this.

But I think, very quickly, I don’t think so. Otherwise it would have happened already.

You see basically how Ultimate GPT is being used. And there was like a moment of hype and now it’s like more stable again. I think it’s all about those incremental productivity advantages that we will find through that. I was very surprised – I was also asking some of our customers that have automated a large amount of their queries and actually found out there were literally thousands of people answering the same question, day-in & day-out.

80% of their questions were the same and they were manually typing the same answer. And so then you need to ask yourself: okay, that’s a position, but does it make sense?

Grace: And is it interesting? Is it fulfilling?

Sylvain: Is it fulfilling? The answer is no. And I think the answer would be no for anyone. 

The question is more that there will be attrition in jobs. But at the same time, and hopefully only for positions where we want the individuals to have better perspectives. 

At the same time, those automation possibilities enable people to actually enhance their work, because then they have the possibility to get macros to summarize, their answers get suggestions on what kind of answers they would get. They would get automated feedback on what sentiments the customer has, like these sentiment analysis features. So in a way it enables them to do their jobs better. 

Be it an automation platform or agent-enhancing functionalities automation, I think it comes together to create better synergies. At the end, when in a support organization, you grow and evolve into a position where you manage automation projects and become like chatbot program managers, or you remain an agent and you just use AI-powered features. I think, no matter what, you’re going to benefit from it. 

The answer is that it’s never going to replace everyone. Would a company automate a hundred percent of their queries? They can, they could do it already before ChatGPT.

The answer is, does it make sense?

Grace: There was something in the news yesterday about a CEO who fired 90% of his workforce because he figured that ChatGPT could carry it out. And the response was: why on earth would you do this? Because of course, without anybody in place, you’re gonna drive it to the ground. It’s not there yet, and it might never be there.

Sylvain: I think it’s never going to be there. Cause now if you see again the rise in Oh, now prompts are like the new future. Like we need to get upskilled in prompts. We need to be able to understand how to create good prompts. That’s not rocket science. It was there before, technology is only enabled by human craft, human actions and human thinking behind it on how to use it. 

And that’s the same exact example for prompts and the same way that you describe that – your own example is like you started using touch with one sentence. Cause you don’t really know what you can ask? How can you ask it? And then you start upskilling yourself on how to ask it, feeling more confident with the tool. You see that you actually get a much better output from it. 

I think that’s what’s interesting. And I think we’re going to see this across the board.

Grace: Thank you very much. That was a wonderful answer. I think there’s a lot there to unpack in terms of how that relates to customer service, just as a way of making teams hopefully happier. With customer service in particular there’s perennially been the issue that employee retention is low.

Whereas this is hopefully something that, if service is a much more rewarding department to work in, there is going to be a lot less of that. It will be much more of an asset to the team to have people who work for longer, who are smarter with the capabilities, who are more ingratiated into the company as a whole. Because they are able to be the designer of this success. Really facilitating that relationship on a higher level rather than the person who’s sitting answering questions that could easily be redirected into FAQs.

Sylvain: Absolutely. And hopefully it can also help in the direction that customer support leaders have been shouting for years to see customer support organizations as a source of revenue and not as a cost center.

Grace: A hundred percent.

Sylvain: I think this is a saying that has been going around, probably not just in customer support, but in a lot of organizations you could argue the same. Hopefully now with those new positions – like chatbot program managers, for instance – where you need to be cross-functional, you need to work with sales, you need to work with marketing.

That’s actually a good cue to push organizations to see customer support organizations as a source of revenue and not just a cost center. So elevating the conversation for customer support organizations to be really driving value and driving the brand forward for a company.

Grace: Exactly. That’s probably a good place to say thank you very much. This has been super fruitful. I’m really glad we had you on.

Sylvain: Thank you, Grace. Thanks for the invite.

Never miss an update

By subscribing you agree to Klaus' Privacy Policy and would like to get educational content to your email.