*Thisarticle is based on a YouTube video by Sophie Hundertmark, an expert in the use of artificial intelligence with a focus on chatbots and strategic AI applications in companies and public institutions. Sophie is a researcher and lecturer at the Lucerne University of Applied Sciences and Arts and is doing her doctorate in Conversational AI at the University of Fribourg. The blog text was created using a custom GPT model that was trained on Sophie’s video content, language style and expertise. The result is well-founded, up-to-date articles based on Sophie Hundertmark‘s own expertise.
You can find the link to the video at the end of this article.
Responding to reservation requests, handling emails, switching between different systems, and being there for guests on-site all at the same time: Everyday life in the restaurant industry has long since been about more than just cooking and serving. In the background, there are numerous administrative tasks that take up time and distract employees from their actual work.
This is exactly where artificial intelligence can come into play.
On my podcast, “Sophie Next AI Talk,” I spoke with Selim Akyol of Molinoteq GmbH about how AI can automate communication and administrative processes in restaurants. What I find particularly exciting about this is the approach of not simply introducing AI as just another tool. Instead, it is intended to connect existing systems and relieve employees of as much unnecessary administrative work as possible.
Why AI Is Particularly Interesting in the Restaurant Industry
The restaurant industry is a sector in which people continue to play a central role. Guests want personalized attention; employees must respond to individual situations, and human contact is especially important in the service industry.
At the same time, there are numerous processes running in the background for which this personal contact is less critical.
Reservations are a good example. If one location is fully booked but tables are still available at another location, changing the reservation can quickly trigger a series of communication steps. People talk on the phone, message each other on WhatsApp, or communicate via email. At the same time, reservation systems must be checked and information transferred.
It is precisely these kinds of recurring processes that are well-suited for AI support.
The challenge: Restaurants use many different systems
According to Selim, a key challenge in automation is that restaurants use different systems.
Reservations are managed in one system, work schedules may be managed in another, and tasks are organized using yet another application. In addition, information is received via email, phone, or other communication channels.
So far, it has often been necessary for a person to understand this information, switch between the various systems, and carry out the necessary steps.
Molinoteq’s approach, therefore, is to use AI precisely between these systems. For example, when a reservation request comes in via email, the AI can identify relevant information from the message, cross-reference it with the reservation system, and then draft an appropriate response.
As a result, AI not only handles individual text-based tasks but also serves as a bridge between communication and existing enterprise systems.
AI can tell what an email is about
Another important use case is the automatic classification of incoming messages.
For example, an AI can determine whether an email is:
- Application
- Supplier Invoice
- Reservation Request
- Group Request
deals with.
Depending on the content, different processes can then be initiated.
For example, when a reservation request is made, the AI could check whether a table is available and prepare an appropriate response. For group inquiries, we can compile information and, for example, provide appropriate group menus.
This highlights an important difference between simple automation and an AI-powered solution: The AI can first interpret the content of a message and, based on that, determine which process is relevant.
Employees first review the AI’s suggestions
One interesting point that came up in our conversation concerns trust in AI.
Selim describes the first month of implementation as a learning and training phase. It’s not just the AI that needs to learn. Employees must also familiarize themselves with the new system.
Especially at the beginning, employees often still want to check every suggested answer. They want to know: Would I send this message the same way?
Over time, this process changes. When the AI consistently provides relevant suggestions, trust grows. Responses can be approved more quickly, and the administrative burden is gradually being reduced.
According to Selim, there are already significant time savings after this adjustment period. This allows employees to focus more on the guests on-site, rather than spending a large portion of their time on emails and administrative tasks.
Ease of use is crucial
When it comes to AI projects, we often talk about models, data, and technical capabilities. In day-to-day work, however, there is another factor that determines whether a solution is actually used: the user interface.
Not every restaurant employee is deeply involved with new technologies. At the same time, an application must work quickly and easily in a hectic work environment.
That is why the interface was deliberately designed to be simple.
On their smartphones, employees see various tiles for the features that are relevant to them. For example, when a reservation comes in, they can open the relevant section, review the AI’s suggestion, and confirm it.
Ideally, this process takes only a few seconds. The guest will then receive a response or reservation confirmation right away.
So AI doesn’t just have to be powerful. It must also be integrated into day-to-day work in such a way that employees can use it without creating unnecessary additional complexity.
Why Employees Should Be Involved From the Start
I found Selim’s answer to the question of why the solution has been well received by employees particularly interesting.
A key factor for success is involving future users from day one.
This means that employees are involved not only when the finished AI solution is rolled out. You can contribute your experience right from the initial brainstorming session, through the prototype phase, and into the further development process.
This results in a solution that more closely reflects actual day-to-day work.
At the same time, acceptance is growing. Employees aren’t suddenly presented with a finished system; rather, they were an integral part of the development process themselves.
In my view, this approach can be applied to many AI projects. If you want to automate processes, you should first talk to the people who carry out those processes every day.
AI is intended to ease the workload on employees, not to replace interaction with guests
When introducing AI, the question of how employees will react quickly arises. When it comes to automation in particular, there is often concern that technology could replace jobs.
In the use case described, the situation is different.
AI primarily takes over administrative tasks that employees find unappealing anyway. Instead of spending time on repetitive emails and switching between systems, they can focus more on taking care of their guests.
To me, that is precisely where AI is used in a particularly meaningful way: not necessarily to remove people from a process, but to relieve them of tasks that add little value.
This can be particularly crucial in the restaurant industry. While administrative communication can be automated, providing personalized service to guests remains one of the most important tasks for employees.
Not everything is working perfectly yet: Swiss German remains a challenge
However, the interview also shows that not every AI application is yet reliable enough for production use.
One example is communicating over the phone in Swiss German.
While Selim says Standard German is now going well for him, he continues to have difficulty with Swiss dialects. Both the ability to understand different dialects and the output of a natural-sounding Swiss German AI voice do not yet fully meet the desired requirements.
In addition, there is a slight delay during AI phone calls, which can make the conversation sound unnatural.
In tests, therefore, it is clearly communicated that callers are speaking to a synthetic voice. Cloning or mimicking a real voice is also being tested to make communication sound more natural.
This example illustrates well that just because a technology is generally available does not automatically mean that it has already achieved the necessary quality for every use case.
One tool for employees, many systems running in the background
Another important point that came up in our conversation concerns the number of applications that employees have to work with.
The vision is that, in the long term, employees will use only one central interface as much as possible. The various systems can continue to run in the background and be coordinated with one another.
For example, employees will no longer have to switch back and forth between the reservation system, time tracking, and task management.
That is why Molinoteq is working, among other things, on integrating additional processes into the central interface. These include, for example:
- Tracking Work Hours and Expenses
- Allocation of Responsibilities for Individual Locations
- Recurring daily and weekly tasks
- Possible integrations with supplier management
The underlying principle is fascinating: Not every existing system needs to be replaced. Instead, a central interface can simplify access, while the various applications continue to run in the background.
From reservation management to supplier management
The potential for AI in the restaurant industry goes far beyond simply answering emails.
When reservations, communication, staff scheduling, and other data are linked together, new opportunities for automation arise.
In the future, for example, information from reservations could be more closely integrated with other business processes. Supplier management is also a possible next step.
The point is not to equip as many processes as possible with AI simply because it is technically feasible.
The key question is where the effort actually lies and what benefits automation provides.
Which process takes up a lot of your time but yields little results?
At the end of our conversation, Selim asked a question that isn’t just one that restaurants should be asking themselves:
Which process takes up the most of my time in my daily routine while adding little value?
That’s exactly where it’s worth taking a closer look.
Perhaps it involves manually transferring information between two systems. Perhaps it’s recurring emails, sorting through inquiries, or administrative tasks that follow the same pattern every day.
Such processes are often good candidates for AI-driven automation.
This doesn’t just apply to restaurants. This principle applies equally to freelancers, small and medium-sized businesses, and large corporations.
The starting point for a successful AI project, therefore, does not have to be the question: “Where can we use AI?”
A better question would be:
“Which task regularly takes up a lot of our time, even though it adds little value?”
Anyone who identifies this process may have already found a very good initial use case for AI. Support implementation.
Any further questions?
Do you have any questions? I am happy to support you, act as a sparring partner and answer your questions. I am always happy to receive your messages, preferably by WhatsApp message or e-mail.
