AI in Business
September 30, 2025
10 min read
The initial wave of hype around AI triggered widespread concerns about the potential risks of the technology. However, now, it has become clear that AI can act as a powerful tool capable of transforming business processes.
But how can companies find the right approaches to enterprise transformation? And how can they distinguish true innovation from mediocre solutions? Max Golikov, Sigli’s CBDO, discussed these questions with Thibaud De Keyzer, CEO of Mobilexpense, Board Advisor at TechWolf, and Strategic Advisor at Fortino Capital, in the Innovantage podcast episode.
Thibaud began his career at IBM, and later spent nearly a decade at SAP, Europe’s largest software company, where he held roles across Belgium, France, and the Netherlands. He founded and successfully sold an SAP consulting business before becoming the CEO of Mobilexpense.
Thibaud joined this company three years ago. At that time, the team had a clear goal to make expense management fully touchless.
Founded in 2000, Mobilexpense is headquartered in Belgium but operates as a truly European company. It has a presence in eight countries. Its team consists of 200 employees, who represent 28 nationalities. These are people from different generations and with different backgrounds.
Bridging cultural gaps in such a team is a crucial task. Thibaud admitted that the foundation was already in place when he became the CEO of the company. His role has mainly been to nurture and strengthen it. For Thibaud, this diversity is not just a cultural asset but a business advantage as it offers perspectives from many angles.
Innovation has always been part of Mobilexpense’s principles. For example, the company launched a web-based solution long before mobile technology became mainstream. Many of the processes were embedded in the product years ahead of competitors.
When it comes to AI, for Thibaud, it is not just a new feature but the next fundamental utility. It is comparable to the internet in the 1990s or mobile phones in the 2000s. Artificial intelligence represents the next great wave of transformation.
According to Thibaud, innovation is one of the strongest cultural drivers. And at the same time, it is a leadership responsibility. At Mobilexpense, he encourages curiosity across the team and provides employees with access to the tools, time, and resources needed to experiment. His role as CEO is to connect the dots between these initiatives and ensure they align with the company’s broader vision.
AI is a non-negotiable priority that requires CEO-level ownership. He leads this effort personally and guides teams to focus on solving real business problems instead of pursuing technology for its own sake.
As Thibaud explained, at Mobilexpense, every new idea enters through a discovery tool where employees can contribute suggestions. From there, the most promising concepts are evaluated with input from business stakeholders and the CTO.
Projects that pass this initial filter quickly move into a 90-day pilot phase. If early results are positive, the next step is to test scalability. Once a project proves it can scale, it is moved into production. Here, governance becomes critical. It is very important to track ROI and ensure that the project is secure and ethical.
AI projects are not prioritized just because this is AI. Their value is measured by the tangible improvements they can bring to the business. For example, in customer success, that might mean reducing ticket resolution time; in sales, accelerating deal conversions; and in HR, shortening the time required to qualify candidates.
For Thibaud, it is necessary to make sure that AI adoption delivers compounding value across the entire organization.
Both failed and successful projects can teach us a lot. Thibaud mentioned a couple of examples from his practice.
One early AI experiment involved deploying a sophisticated chatbot in customer success. The team attempted to load the model with a large volume of information to provide comprehensive answers. However, the results fell short. Responses were unclear, unhelpful, and frustrating for customers.
Looking back, Thibaud noted that the project’s biggest mistake was removing the human-in-the-loop too quickly. The team later corrected this. They scaled back the chatbot’s scope and reintroduced experts at critical points in the process. The failure demonstrated that human oversight remains essential in AI-driven customer support.
On the success side, he pointed to an internal project designed to help auditors quickly access compliance information related to expense reports and company policies.
Initially, it was introduced as a small internal efficiency tool. But the project attracted unexpected customer interest and was rolled out externally. It enabled clients to check compliance before submitting expense reports and turned out to be five times more valuable than anticipated.
AI is reshaping not only how companies operate but also how CEOs should lead their teams. Today, CEOs often feel the pressure to innovate, driven by the fear of being outpaced by competitors.
In this context, AI can help a lot. A decade ago, product development cycles required months of effort. Modern AI tools allow companies to test, iterate, and scale at unprecedented speed.
A “10x engineer”, who is 10 times better than a mediocre engineer, can become a “1000x engineer”. A skilled professional who leverages AI effectively can outperform peers not just tenfold but a thousandfold.
Now, for Thibaud, one of the greatest concerns is neither the technology itself nor the competition. It is vital for him to ensure that his team remains curious and engaged with AI. For him, a lack of interest in AI can be a sign of the organization’s weakness in today’s fast-moving landscape.
Thibaud believes Mobilexpense has successfully fostered a culture of innovation.
To make innovation part of everyone’s responsibility, Mobilexpense holds monthly all-hands meetings where teams share outcomes, discuss opportunities, and raise concerns.
He is confident that it is important not to make innovation a purely top-down initiative. It needs to come from the people.
He explained that in his role, he also has strong leadership support from the CPO and the CTO. At the same time, he values continuous employee feedback.
Though he remains quite optimistic about the future of this technology, AI’s rapid evolution at a large scale looks like an “uncontrolled experiment.” Only time will show the real impact on people and humanity.
Thibaud shared that the past 12–15 months have been a period of intense learning for him as he wanted to truly understand what AI is, what it is good at, and what it is not.
He recalled that during his time at IBM (it was around 2007–2010), artificial intelligence was already a buzzword, but there was little clarity on practical application. Today, thanks to breakthroughs in large language models (LLMs), the possibilities have become tangible.
Even though, as a CEO, Thibaud doesn’t need to code or build infrastructure himself, he sees the potential of AI in such tasks. According to him, it is essential to grasp AI at the application level. It is important to find the use cases where AI can give an advantage in business, for example, in identifying patterns, creating content, or running predictive models.
With AI, innovation.has become more accessible than ever before. Young people demonstrate outstanding skills in working with tools like LLMs to explore business problems and potential startup ideas.
Now, you don’t need to have a full-stack developer or a machine learning engineer for every task. Nearly 80% of questions can be answered by an LLM. And these answers will be good enough at least for moving forward with your ideas.
According to Thibaud, AI is unlikely to eliminate roles entirely. However, it will transform how people work. While some jobs may become obsolete, new roles will emerge. And quite often, these new roles will require different skills and approaches.
Curiosity and engagement with AI will be key for employees to thrive. Entry-level roles may be affected first, particularly tasks that are repetitive, as they can be easily automated. However, younger professionals who embrace AI early are well-positioned to adapt. At the same time, more experienced employees can leverage AI tools to work more efficiently and focus on higher-value tasks.
For example, developers using tools like GitHub Copilot can reduce time spent on routine documentation or repetitive coding. It means that they will have more time to concentrate on more meaningful work.
AI is also reshaping learning for the next generation. Students increasingly rely on laptops and digital tools. As a result, the value of traditional skills like handwriting can become rather questionable.
The pace of technological change means today’s students may receive education very differently from previous generations. Curricula must adapt continuously, and the skills taught at the start of a multi-year program may be outdated by graduation.
AI can now draft essays or assist with research far more efficiently than students could on their own. But true learning comes from the process, which includes structuring ideas, conducting research, and developing critical thinking.
Apart from that, Thibaud noted that formal education is no longer the only path to success. Many of his colleagues at Mobilexpense, including some in leadership roles, never completed traditional schooling. But they have strong self-learning skills and solid practical experience.
Moreover, AI can be used as a great research tool. Nevertheless, it shouldn’t be seen as a replacement for critical thinking. It is still very important to know how to ask the right questions.
AI can help with market research or idea generation. But the ultimate intelligence lies in identifying the right problem to solve and understanding the target audience. Human judgment still remains central.
Today, we can observe the emergence of AI-native companies. Such startups are built from the ground up with AI tools integrated into their business models. These can be just five-person teams creating projects that could become unicorns. Such companies can scale efficiently and leverage AI to accelerate research, planning, and product development.
However, established companies like Mobilexpense retain significant advantages. They have extensive transactional data, deep industry insights, and a loyal customer base. With all this, they can leverage AI to enhance existing operations and create new value.
For example, for Mobilexpense, AI presents an opportunity to transform expense management from a routine administrative task into a strategic process. Thanks to the AI capabilities, the company can provide a comprehensive solution that supports better decision-making and creates a more engaging experience for users.
Large language models are just a small subset of AI, which spans a wide range of computing capabilities. For executives and investors, understanding the different subsets and their practical applications is crucial. While AI offers significant opportunities, not every solution represents true innovation.
Many new startups that enter the market appear to leverage AI superficially. For instance, they just package LLMs into niche applications without addressing real business problems. These so-called “ChatGPT wrappers” often fail to deliver long-term value and are likely to be demystified quickly. Investors always need to analyze the actual impact and novelty of AI solutions to distinguish genuine innovation from marketing hype.
Thibaud mentioned that AI is similar to prior waves of tech cycles, where repetitive business tasks were transformed into scalable software solutions. The potential of artificial intelligence is enormous. But not every AI application substantially improves outcomes.
For this reason, investors should focus not only on the initial idea alone, but also on teams and their ability to pivot and execute.
Enterprises have broader expertise and resources than smaller businesses. That’s why they are better positioned to validate emerging AI technologies. Smaller companies or less experienced investors may be more vulnerable to hype and may adopt solutions that lack sustainability.
Thibaud predicted that the current wave of AI hype is likely to settle within the next 12–24 months. Its true winners will be those startups that can integrate AI into enterprise-ready, platform-ready, and scalable solutions.
At the end of their conversation, Max also asked Thibaud to share his thoughts about the areas that are most likely to be transformed by AI.
While it is still difficult to predict the full scope of AI’s impact, Thibaud believes the greatest potential lies in enabling augmented humans. These employees across different departments will leverage AI tools to dramatically enhance productivity and decision-making.
According to Thibaud, in the future, engineers, salespeople, and product managers will be able to operate at a thousand-times capacity. It will be possible not only through technology, but also by using AI to solve real business problems.
The key idea that Thibaud emphasized is that tech transformations have never been about technology but also about business problems. In turn, business is always directly related to people.
Given this, it becomes vital to educate people, motivate, and incentivize them to get on the AI train. As long as organizations stay transparent, they can work together to make this world a better place for everyone.
How technologies are changing the world around us is one of the most widely discussed topics in the Innovantage podcast. If you are curious to learn more, don’t miss our next episodes that will uncover new insights from leading tech and business experts.
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.webp)
.webp)
