“Make gives us so much power”: How Make-powered AI voice agent saved Stellantis &You £200,000 per year

Stellantis &You UK represents multiple vehicle manufacturer brands across around 30 dealership locations and 1,500 employees in the UK. 

For over a decade, any of the many calls a Stellantis &You UK dealership couldn't answer went to a third-party contact centre in Birmingham. The team only worked office hours, and because they weren't dealership staff, the answers varied depending on who picked up. This external support also required significant financial investment. 

Rather than keep paying a lot of money for inconsistent service, the team decided to bring the work in-house and build their own AI voice agent with ElevenLabs. But putting AI in front of customers raised real questions. Would it handle the call volume? Would it do so reliably? 

To make sure it does, Stellantis &You UK decided to connect it to their own data through Make. The AI agent now saves the business £200,000 a year and handles 20,000 calls a month on this use case alone.

"Make gives us the power to learn fast and get solutions to our users in a rapid way. We are not developers, we are not programmers, so we need the solutions to be no-code, low-code, where we can configure things in a visual way. We tried a competitor before, but the clarity and interface of Make are incomparable."

David Male, Head of Transformation at Stellantis &You UK

The challenge: Inconsistent service and rising costs

Paying for the external contact center became a substantial investment. 

But cost wasn't the only problem. Stellantis &You UK’s team had to review hundreds of the contact centre's calls before deciding this was the route to take, and what they heard was a different customer experience every time. Quality depended entirely on which of the external agents picked up and how well they knew the business.

These issues were substantial enough to move the support in-house. 

So, the company ended up creating their own conversational AI agent with ElevenLabs. It worked well, but brought up new concerns. Not only whether it will handle thousands of calls each month, but how it would do so. The range of queries was wide, answers often complex, and their accuracy business-critical. Stellanis &You UK had to ask themselves: will the AI do what it’s supposed to do? Will it do so well enough to actually improve the customer's experience?

"The biggest risk is the unknown with AI. LLMs are constantly changing and developing. How are they going to respond to so many different customer queries? How can you add guardrails to make sure it shouldn't be doing things that it shouldn't be doing?"

Connor Power, Digital Process Improvement Manager at Stellantis &You UK

The vision: Trading a black box AI for a glass one

The ambition was straightforward to state but hard to pull off: use conversational AI to transform how the business talks to customers, without losing control of what that AI actually does.

That control question mattered a lot. An AI agent talking to customers unsupervised is a bigger risk than a slow process, and the team knew it.

Make solved that with governance: every scenario tracks who changed what, rolls back to a previous version in seconds, and surfaces exactly what went wrong when it breaks. 

And so Make became the connective layer between the ElevenLabs AI agent and the business's own data. When a call routes to the agent, Make queries the customer database using the caller's phone number, surfacing who's calling and which vehicle they own before the conversation even starts. That way, Stellantis &You UK can provide fast and reliable service. 

"Make was absolutely the glue of this use case. It's performing vital processes in the call that takes place… This gives us so much power in the conversation right from the off. It really transformed the journey, really reduced the call duration, because you cut out so much time you'd otherwise spend on understanding who's calling and what about."

Connor Power, Digital Process Improvement Manager at Stellantis &You UK 

Another big advantage is that none of the setup required a developer. The visual interface lets the team be fully in control of what and how they build.

"There are lots of off-the-shelf solutions which can certainly do a job. The reason we didn't want to go with them is because we like to tailor and tweak what we build to suit our needs. The only way we can do that is by being in control of our own destiny and building our own solutions."

Connor Power, Digital Process Improvement Manager at Stellantis &You UK 

Make Grid (a map of the entire automation and AI ecosystem that helps visually manage and optimize it) became one of their favorite features, helping the team see an entire process at a glance. Built-in error handling flags problems before customers noticed them, and pre-built ElevenLabs modules means nobody has to handle the API documentation directly.

"It feels satisfying to see a complicated process mapped out in a workflow that looks simple and easy to follow."

Elizabeth Bolton, App and Software Engineer (Digital Team) at Stellantis &You UK 

The impact: 20,000 calls a month, £200,000 saved

After the first month of a single-dealership pilot, the agent had handled around 500 calls at a 95% success rate. It can now tell a customer exactly where their car sits in the workshop queue. Today, the agent handles 20,000 calls a month across all dealerships, saves the business around £200,000 annually on this use case alone, and has freed up the human reception team for other work.

"None of this would be possible without Make. We've built so much structure around it. Ten years ago, we had one scenario. Today, we're running over a thousand of them, and processing around 900,000 operations a day.”

Connor Power, Digital Process Improvement Manager at Stellantis &You UK 

What's next: One use case was never the end goal

With the voice agent proven successful, the team is already building outbound versions of the same idea: an SMS booking agent with ClickSend, and an AI "sales genius" and "product genius" to support both sales advisors and customers directly on the website.‍

"It's evidence that conversational AI is absolutely scalable, absolutely reliable, and we can now look at it for more use cases on our project roadmap.

Connor Power, Digital Process Improvement Manager at Stellantis &You UK

The results haven't stayed contained to one team, either. The wider Stellantis business across Europe is watching what they do, and the team is now actively helping other parts of the organization get started with Make.

‍