Vidiv: when artificial intelligence begins to really talk

There are projects that are not born from a closed idea, but from a search process. Vidiv is an example of this. Its trajectory is built on trial, error and continuous learning to identify a specific problem that many organizations share today. As explained by the startup, the starting point was to detect “the saturation of sales and service teams, and the enormous number of repetitive conversations that consume time and energy in companies”.

The first version of Vidiv emerged in 2020, in the midst of a pandemic, as a solution aimed at online events. It was a period of intense experimentation, but also of the realization that the initial idea did not quite fit as expected. The real turning point came in November 2024, when the team launched what today constitutes the core of the project. “There we were clear about the problem we wanted to solve”, they point out from Vidiv, in reference to the development of artificial intelligence agents aimed at conversational attention.

Although the commercial area remains one of its main use cases and the one that attracts the most customers, working with different scenarios over the past year has broadened its vision. “We have worked with many different cases and that has led us to better define where we want to go”, they explain. Today, the goal is clear: “Automate any response or process that requires conversational attention, whether by phone, web or other channels, without losing naturalness or quality in the experience”, they aim.

Victoria, Vidiv's AI agent

That vision is embodied in Victoria, its artificial intelligence agent. The usual reaction when someone interacts with her for the first time is surprise. “Approximately 30% of people don't realize they're talking to an AI agent”, they explain from Vidiv, “the experience is very similar to that of a human conversation. The customer calls or writes with a real problem and receives a clear, contextual and direct answer”.

The difference is in the pace. “There are no waiting, no awkward silences or looping music while someone is looking for information”, they point out. Victoria responds practically instantly, shares links, locates information and guides the conversation in a fluid way. For the user, this eliminates one of the great frictions of traditional attention and makes the interaction “more agile, more comfortable and, paradoxically, more human”.

Behind that experience there are very precise technical decisions. For Vidiv, one of the key pillars is response time. “The naturalness of a conversation is directly dependent on latency”, they explain, and that's why they work to keep it under 0.5—0.6 seconds, adjusting the model according to the use case. Added to this is total availability. “Victoria is operational 24/7, doesn't get tired, doesn't get saturated and always maintains the same level of care”, they explain.

Another differentiating aspect is its ability to scale conversations. “You can manage anything from a single interaction to tens or hundreds of conversations at the same time, without loss of quality”, stand out from the team. This technical capacity is complemented by something less visible, but just as relevant: measurement. “As an AI agent, we can analyze practically everything”, they stress, “from the degree of satisfaction or the number of queries resolved to the average duration of calls, peaks of activity or even questions that were not initially planned and that appear in real use”.

Real impact on organizations

The results observed in projects where Victoria is already operational reinforce that perception of value. Beyond specific metrics, there is an effect that is constantly repeated. “The wow effect”, they point out from Vidiv. Many customers arrive after previous experiences with Chatbots very limited and they discover that artificial intelligence can really be useful when it is well applied.

This discovery also has a significant internal impact. “We have seen how a very common barrier, resistance to change, is broken”, they explain. In several projects, initially reticent teams have ended up asking for the agent to be active around the clock because “frees them from an enormous repetitive workload”. In terms of impact, the clearest results are reflected in “the liberation of the human team”, greater efficiency in low-value tasks and an increase in the resolution rate of frequent queries. In some cases, they add, improvements have been observed in conversion that “they reach 20—30% in forms, payments or guided processes”.

The collaboration with Santalucía Impulsa

This tour explains the recognition obtained by Vidiv and its choice as the winning startup of the Santalucía Impulsa's acceleration program in 2025. A process that has meant, in addition to visibility, a demanding environment that is key to its growth. “Working with SMEs or medium-sized companies is not the same as with a large corporation”, they explain from Vidiv, “needs, processes and product design change completely”.

From his experience, collaboration with ecosystems such as Santalucía Impulsa is decisive. “It allows us to grow faster and access levels of demand, scale and rigor that are not common in smaller projects. At the same time, it is a constant source of learning. This contrast forces us to improve both the product and the business vision and accelerates our growth in a way that would be very difficult to achieve alone”.

Looking to 2026 and Beyond

Looking ahead to 2026, Victoria's evolution is closely linked to the pace at which technology is advancing. “In addition to voice and text, we are exploring new formats and contexts”, explain, as exclusively textual experiences for certain use cases. In parallel, they continue to work to further reduce latencies and improve voice quality. They are also moving forward, “with special care and always with consent”, in the possibility of outgoing calls, a step that opens up completely new scenarios.

The approach is clear. “Building a living product”, underline from Vidiv. “Today we're talking about text and voice; tomorrow it could be video or another format that isn't even on the table right now”. The key, they insist, is “adapt quickly and meaningfully to what really provides value”.

Vidiv is, in this sense, a clear example of how artificial intelligence is beginning to abandon the terrain of promise to settle in that of real, useful and measurable conversation. And also how innovation, when put to the test in demanding environments, finds its definitive value.