For many contact center agents, the rise of AI has felt more like a threat than a helping hand.
The narrative around automation, job displacement, and ‘human-free’ customer service models is understandably worrying for some agents, especially those on the front line.
But the real issue isn’t AI itself; it’s the wrong AI.
Think clunky integrations, inaccurate summaries, tools that slow agents down rather than speed them up, and worst of all: systems that override human judgment instead of supporting it.
This is what can happen when companies choose AI vendors that provide generic AI.
These solutions are not built for purpose, so they struggle to handle the complexities of contact centers and sometimes unintentionally work against humans rather than for them.
What agents really need isn’t AI that takes over; it’s intelligent assistance that stays in the background, listens, offers guidance when needed, and handles some of the more mundane and repetitive tasks.
That’s precisely where solutions like Diabolocom’s Agent Assist come in.
From Call Chaos to Copilot Calm
Launched in late 2024, Agent Assist is built on a straightforward principle: agents aren’t the problem; their tools are.
The product is designed to act as a live, AI-powered copilot that helps agents transcribe calls, surface relevant information, and summarize interactions quickly and accurately.
“The idea is to support the agent during and after the call,” explains Rémi Guinier, Lead AI Product Manager at Diabolocom.
“We’re not trying to do the job for agents. They still need to think, engage customers, and make decisions. But we give them the tools to be faster, more accurate, and less stressed.”
At its core, Agent Assist listens, learns, and lends a hand in real time.
It transcribes calls live, identifies key customer data points, and delivers in-the-moment suggestions based on what the customer is saying.
It can autofill forms, generate summaries in seconds, and retrieve knowledge articles from the firm’s knowledge base – all without the agent needing to search or ask.
Agent Assist is the perfect example of what a copilot should be: AI sitting alongside, not in the driver’s seat.
Performance Without the Friction
While tools like Diabolocom’s Agent Assist offer impressive features, many customers are still failing to maximize the potential of their AI-powered offerings.
Indeed, one of the biggest challenges in deploying AI is usability.
If the tool doesn’t integrate cleanly with existing systems, it creates more problems than it solves.
In order to combat this issue, Diabolocom prioritizes seamless integration and configurability.
Agent Assist’s prompts and outputs are fully customizable. Contact centers can decide exactly how summaries should look, which data gets extracted, and how knowledge base content is presented.
Guinier explained how other tools often force customers into a fixed format, whereas Agent Assist “lets the customer decide.”
“Bullet points? Paragraphs? Custom forms? You choose. The system adapts to agents, not the other way around.”
As an example of this adaptability, Guinier discussed contact center scenarios in multilingual regions like Switzerland, where agents frequently switch languages, explaining how the tool still performs reliably in these areas.
Real Time, Real ROI
If usability and deployment are the biggest AI stumbling blocks from a customer perspective, from the business end, it is undoubtedly proving ROI.
Guinier describes the productivity impact of Agent Assist as “immediate”.
Wrap-up times – often a major source of friction – drop dramatically. Instead of spending two to five minutes typing notes and ticking boxes post-call, agents can complete the process in as little as 15 seconds.
“It’s around one minute saved per call,” said Guinier.
“That time goes straight back into the queue. You reduce wait times, handle more calls, and improve the overall customer experience.”
‘Time is money’ might be an old adage, but it’s still an accurate one.
By significantly reducing the time it takes to satisfy each query, businesses can run more efficiently and cut costs.
Agent Assist also delivers on the traditional contact center metrics, with average handle time (AHT), first contact resolution (FCR), and customer satisfaction (CSAT) all experiencing measurable gains.
Looking beyond these core metrics, Diabolocom is expanding Agent Assist’s analytics capabilities to help teams continuously improve.
The enhancements will allow managers to track prompt usage, monitor token consumption, and analyze the most accessed knowledge content.
This opens the door to better content management, smarter prompt design, and improved coaching – all without overwhelming dashboards or workflows.
And from an executive standpoint, the value is obvious.
“If you save one minute per call and handle 10,000 calls a day, that’s 10,000 minutes saved,” Guinier said. “The numbers speak for themselves.”
What’s Next for Agent Assist?
The roadmap for Agent Assist continues to evolve.
Diabolocom is exploring retrieval-augmented generation (RAG) to make knowledge base results even more relevant and reliable.
Yet, despite these exciting developments, Guinier was quick to emphasize that Diabolocom’s core philosophy remains unchanged: this isn’t about replacing humans; it’s about giving them the best possible backup.
“We want AI to be useful, not magical. Agents shouldn’t be afraid of AI; they should feel supported by it.”
For more information on Diabolocom’s approach to AI, check out this article about real solutions for contact center problems.
You can also learn more about the company by watching interviews with Frédéric Durand (Founder and CEO) and Collin Ehret (Senior Account Executive).
