The Future of AI in Customer Service

We sit with CTO and co-founder of Parloa, Alexander Matthey, where he's building the AI agents that promise to revolutionize how businesses interact with their customers. Unlike the doomsday prophets predicting mass unemployment, Matthey paints a more nuanced picture of what's coming.

In our latest podcast episode, we dive deep into the transformation of customer service, the survival strategies for BPOs, and why building the "Ferrari" of AI platforms matters more than ever.

The Speed Game Has Changed Forever

Starting a company in AI isn't like starting one in fintech or beauty tech. Matthey learned this the hard way after his experiences at Adyen and Glossybox.

"I think the biggest differentiator is, I think I was always obsessed with speed and I was also obsessed with focus. However, currently the industry is moving very fast as well and also there's so much noise and so much things that can disturb you so much more than in FinTech and so much more than in Beauty."

The challenge is of course maintaining focus while the entire industry shifts beneath your feet. Where founders once competed against each other's clock speed, they now race against the market's relentless pace of innovation.

The Radical Vision: People Will Actually Want to Talk to Customer Service

Here's where Matthey drops his most controversial take. Parloa is automating customer service and potentially reimagining it entirely.

"We believe that in the future people will want to talk to somebody at customers. And that's what Parloa is building towards... People will not just reach out when they have a problem like it is with customer service right now, but they will prefer to talk to their personalized AI agent going forward."

Opening hours, language barriers - all moot points. Instead, you have a personal AI agent that knows your entire history and can handle complex requests across multiple channels and languages.

Customer Service Agents: Evolution, Not Extinction

But what happens to the millions of customer service representatives worldwide? Matthey doesn't sugarcoat the transformation, but he also doesn't predict their demise.

"I think the job of a customer service agent will not completely go away, but it will become much more specialized on very complex use cases, on very specific use cases. So therefore it will also be more fun."

The mundane password resets and tracking inquiries? Those are gone. What remains are the complex negotiations, the emotional support calls, and the edge cases that require human judgment and empathy. It's a fundamental shift in the role, not an elimination of it.

Why Enterprises Shouldn't Build Their Own AI Agents

With AI democratizing software development, why wouldn't large enterprises just build their own customer service agents? Matthey has a clear answer, drawn from his Adyen playbook.

"I think you want to control what's really important for your business. And I don't believe that the development resources of many large customers, many large enterprises are best used to build an agentic OS, to think about how to build simulations, evaluations for agents, how to build guardrails across different regions."

The complexity goes far beyond connecting to an LLM - there’s versioning agents, managing hierarchical tenant systems, ensuring compliance across regions, and building evaluation frameworks. These aren't core competencies for airlines, insurers, or retailers - and they really shouldn't be.

The Ferrari Strategy: Why Execution Beats Everything

In a world where everyone claims to have the same features, how do you differentiate? Matthey's answer cuts through the noise with a perfect analogy.

"It's the same with, I don't know, like a Ferrari. Can somebody else build a Ferrari? Yes. It's also only steel and the motor and the engine and some leather and stitching. Is somebody else capable of doing it? Well, that's the problem. So you need to get the right team together and you need to focus on really building that Ferrari."

The Jerry Maguire Moment: Building Modular from Day One

When pressed on how someone could beat Parloa at their own game, Matthey reveals a crucial insight that many AI startups miss.

"There is a strong lever into the industry, especially the enterprise industry, if you from the very start think about how to make it a modular system, to be honest. So that you do not only rely on one very specific agent SDK, one specific LLM, one specific STT or TTS provider."

Large enterprises have existing contracts, approved vendors, and specific compliance requirements. A one-size-fits-all solution won't work. Building modularity from the start by allowing customers to plug in their preferred LLMs or use their existing cloud credits. It will become a massive competitive advantage.

The Future of AI Agents: More Than Just Chat

Looking ahead, Matthey sees AI agents handling increasingly complex tasks, moving far beyond simple query resolution.

"What you see right now is that, I don't know, like the first use cases are intent recognition, or the first use cases are, you know, take over a few of the higher volume use cases, where I think that will grow over time and over time towards, you know, also being able to receive payments or also being able to not just help cancel a booking or refund the booking, but actually plan a whole trip."

The enterprise systems that need to expose their capabilities through APIs are often to blame for bad connectivity. As these barriers fall, agents will orchestrate complex, multi-step processes that today require multiple human touchpoints.

BPOs at a Crossroads: Adapt or Perish

For Business Process Outsourcers, the message is stark but not hopeless.

"I think the BPO's that actually start driving this change will have a chance, a good chance of being successful in the future as well... However, if they don't change, they will be out of business relatively soon."

The opportunity lies in becoming the bridge between AI capabilities and non-digital native industries. Airlines, insurers, and energy companies need partners who understand both their legacy systems and the new AI paradigm. BPOs that position themselves as transformation partners rather than just labor arbitrage will thrive.

The U.S. Expansion: Bringing European Execution to American Speed

Parloa's U.S. expansion follows a familiar playbook—one Matthey helped write at Adyen.

"We want to be a US company, there is some value in being global as well. If you're looking at the largest companies in the world, they are all global as well, and they will want to run their agents globally."

But this isn't just about planting a flag. It's about absorbing the American approach to innovation—the willingness to think "out of the box" that Matthey admires from the Ford vs. Ferrari story. By combining European execution discipline with American innovation speed, Parloa aims to serve global enterprises that need both.

The Path Forward: Value-Based Pricing and Outcome Focus

Perhaps the most fundamental shift Matthey predicts is in how AI services are priced and valued. And it’s a take we love at Paid.

"Going forward, yeah, I think it would be nice to be able to pay agents for the tasks that they have successfully achieved and actually deliver value. So that is our long-term goal."

Moving from time-based or consumption-based pricing to outcome-based models aligns incentives between providers and customers. It's a shift that mirrors the broader transformation in customer service - from measuring minutes to measuring satisfaction and resolution.

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.