Why AI projects fail in companies, according to Vodafone

Artificial intelligence is no longer a promise of the future in the telecommunications sector, but a tool that is changing how companies serve their customers and organize their internal processes. However, many companies, both large and small, find themselves trapped in a phenomenon that is starting to have its own name in the industry: pilotitis.

Pablo Soto Rodríguez, Director of AI and Innovation at Vodafone, addressed this problem in an interview prior to his participation in the IV Andalusia AI Congress, held in Granada under the theme Technology that serves. People who lead. Soto is speaking this Tuesday at 5:25 PM in the Manuel de Falla Hall, as part of the panel discussion The AI Revolution in the Telco Sector, alongside representatives from Orange, Telefónica, and T-Systems.

His explanations come at a time when a large portion of Spanish companies, from major corporations to freelancers, are considering how to incorporate AI without wasting time or money on experiments that lead nowhere.

What changes for the customer with AI in telecommunications

According to Soto, the emergence of artificial intelligence in the telco sector is supported by three specific pillars. The first is employee productivity, allowing people to focus on tasks that truly add value and delegate repetitive work to AI. The second is the automation of business processes. The third, perhaps the most visible to the end user, is the ability to serve the customer in a more personalized way, available 24 hours a day, seven days a week.

The result that any user should notice, the executive summarizes, is more agile support, improved service, and products better tailored to each customer's needs.

Pilotitis: why so many AI projects never reach production

One of the central points of the interview is the diagnosis of a problem that affects numerous organizations: testing AI on a small scale with good results, but never managing to bring it into the company's daily operations.

For Soto, the root of this blockage usually lies in a poor initial definition of the project. He cites the example of trying to automate a process with AI that was created almost thirty years ago as completely analog, without first adapting it to the real possibilities of current technology. In such cases, failure is almost guaranteed if the process is not redesigned or if the specific points where automation makes sense are not selected with criteria.

Why a successful pilot guarantees nothing

The Vodafone executive also explains why a pilot project can be misleading. During the pilot phase, AI usually works well because it is tested on the simplest, general case. The problem appears when the system moves to real production, at which point a multitude of exceptions and particular cases arise that had not been considered, which are what prevent the project from scaling.

The solution, according to Soto, involves dedicating sufficient time to a prior discovery phase, that is, deeply understanding the problem before starting to build any technological solution. That initial phase, he insists, is what makes the difference between a project that works and one that remains a demonstration without continuity.

Advice for freelancers and SMEs looking to start with AI

The market already offers a large number of artificial intelligence tools, but Soto warns that the main criterion when choosing should not be functionality alone, but regulatory compliance. Using free versions of assistants like ChatGPT to manage customer data can lead to problems related to the General Data Protection Regulation (GDPR) and European artificial intelligence regulations.

For a small business that does not know where to start, the executive suggests focusing on two main paths:

  • Implementing a virtual assistant capable of resolving frequent customer questions via phone, web, social media, and WhatsApp, as well as managing appointments following predefined rules. As an example, he cites Vodafone's own IAgente empresas service.
  • Using AI to generate audiovisual content and manage social media, with the goal of improving the sales funnel and increasing the conversion possibilities of potential customers.

Leading AI instead of being dragged by it

Faced with a technology that is advancing at high speed, Soto argues that companies must act as leaders and not just as followers of trends. This means setting their own strategy, deciding with criteria in which areas of the organization AI provides real value, and executing projects always with regulatory compliance and the goal of generating business.

The executive recommends avoiding erratic shifts and hasty decisions motivated by passing fads, opting instead to move forward with a clear and sustained direction over time.

The underlying message for those attending the congress

Soto summarizes his central idea in two points. First, that artificial intelligence is here to stay and no company can afford to be left behind. Second, that the priority should be training, relying on real experts from companies with a proven track record.

The executive also issues a direct warning about certain voices proliferating on social media presenting themselves as AI experts. According to his assessment, many of those profiles live off misinformation and offer solutions that promise unrealistic results, in addition to often not complying with current regulations.

Frequently Asked Questions

What is pilotitis in artificial intelligence projects?

It is the phenomenon where a company tests AI in a pilot project with good results, but never manages to implement it in its daily operations because the system does not scale when faced with the reality of day-to-day life.

Who is Pablo Soto Rodríguez?

He is the Director of AI and Innovation at Vodafone and is a speaker at the IV Andalusia AI Congress, held in Granada.

Why do many AI projects fail in companies?

According to Soto, the main reason is a poor initial definition of the project, without previously adapting old processes or dedicating enough time to deeply understand the problem before developing the technological solution.

What should an SME keep in mind before using AI tools?

It must prioritize that the tool complies with the General Data Protection Regulation and European artificial intelligence regulations, avoiding the use of free versions of generic assistants to manage customer data.

What is Vodafone’s IAgente empresas?

It is a virtual assistant mentioned by Pablo Soto as an example of a tool aimed at answering frequent customer questions and managing appointments via phone, web, social media, and WhatsApp.

Where is the IV Andalusia AI Congress held?

It is held in Granada, in the Manuel de Falla Hall, under the theme Technology that serves. People who lead, with the participation of representatives from companies such as Vodafone, Orange, Telefónica, and T-Systems.

Comparte este contenido:

Deja un comentario

🤖 IA

×
Hola. ¿Qué duda o consulta tienes sobre este contenido?