The entry into force of Spanish Law 10/2025 on Customer Service represents a significant change in how organizations must scale and manage their customer service channels. The conversation is no longer solely about providing a good customer experience. It is now also necessary to guarantee certain levels of availability, response times, access to personalized service, and the ability to demonstrate that these commitments are being met.
AI-powered intelligent agents can play a significant role, not as indiscriminate replacements for human teams, but as an additional layer of capacity that allows service to remain operational when demand increases, extends operating hours, or handles certain processes immediately.
The question, therefore, should not be whether a company must choose between human agents or artificial intelligence. The real challenge lies in designing a hybrid model capable of maintaining service continuity, scaling up when necessary, and respecting at all times the rights established by the new regulation.
Service continuity becomes a matter of operational capacity
The SAC Law stipulates that customer service hours must generally align with the company’s business hours. However, it introduces an additional requirement for certain basic services of general interest that are provided on a continuous basis: the service must be available 24 hours a day, every day of the year, to report incidents related to service continuity.
In addition, inquiries or incidents regarding outages or service suspensions must generally be addressed within a maximum of two hours, providing available information on the causes and an estimated time for restoration.
This obligation poses a primarily operational challenge. The demand for support is rarely constant. A breakdown, a service interruption, a telecommunications issue, or any extraordinary event can multiply the usual volume of calls within minutes.
Permanently sizing a staffing structure to accommodate the highest possible demand is difficult and, in many cases, inefficient. However, lacking sufficient capacity precisely when an incident occurs can compromise both the customer experience and compliance with service levels.
This is where incorporating scalable conversational capabilities makes sense. An intelligent agent can handle some of the initial interactions, identify the reason for the contact, retrieve information available in corporate systems, provide information about a known issue, log a new issue, or collect data before a potential transfer.
Technology thus ceases to be merely a mechanism for reducing costs and becomes part of a customer service resilience strategy.
Automation does not mean blocking access to a human Agent
This point is especially important. The SAC Law expressly permits the use of conversational bots and other automated means, but sets clear limits. Automated systems cannot become the sole means of customer service.
When the interaction takes place by phone or electronically, the customer must be able to request personalized assistance from a human agent through the options available in the main menu and during the interaction. Furthermore, 95% of requests for personalized assistance must be handled, on average, in less than three minutes from the time the customer makes the request.
Therefore, using intelligent agents to ensure continuity does not mean erecting a technological barrier against human agents. The model should work exactly the opposite way. Automation can handle those interactions where it is appropriate, but it must be able to recognize when it is necessary to transfer the conversation. This may occur because the customer requests it, because the nature of the issue requires human intervention, or because the system detects that it does not have enough information to resolve the case correctly.
In these situations, the handoff should preserve the context. The human agent should not simply receive a call but should have access, for example, to the reason for the contact, the information already provided by the customer, the verifications performed, the actions taken, and a summary of the previous conversation.
This combination reduces repetition and allows automation to truly act as an extension of operational capacity, not as an isolated channel.
From call peaks to an elastic service model
One of the main challenges for any contact center is variability. Traditional planning attempts to anticipate the number of calls and adjust available resources, but certain events can throw any forecast off course.
A telecommunications company may receive thousands of additional inquiries in the event of a network outage. An energy company may experience a sharp increase in calls following a power outage. A company with a digital platform may see a large portion of its traffic concentrated there after a service interruption.
Intelligent agents introduce a concept that is common in other areas of technology but still underutilized in customer service: elasticity. Service capacity can be temporarily scaled up without the need to proportionally expand the human workforce. This allows for different levels of intervention:
- Automatic resolution of repetitive and clearly identifiable inquiries
- Providing up-to-date information on known incidents
- Initial data collection and classification of the reason for contact
- Opening and logging of incidents
- Prioritization of specific customers or situations
- Transfer to human teams when necessary
Intelligent agent solutions such as Agentia365, integrated into the Recordia ecosystem, enable the development of these types of hybrid customer service models. The key factor, however, is not merely having an AI agent, but connecting it to the appropriate processes, corporate systems, and escalation mechanisms.
Continuity must also be demonstrable
There is also a second dimension that is sometimes overlooked in automation projects: evidence. It is not enough to have sufficient capacity to handle inquiries; organizations need to know what actually happened. How many interactions took place during an incident? How many were resolved automatically by the system? How many customers requested a human agent? How long did it take them to be connected to one? Which transfers failed? Was the available information provided correctly? Was the incident logged?
The SAC Law specifically reinforces this aspect by requiring affected companies to implement and document service quality assessment systems and, as a general rule, subject them to periodic external audits. The documentation must allow for verification of the reliability and accuracy of the measurements.
For this reason, a business continuity architecture should incorporate evidence collection from the outset. Conversational intelligence platforms can automatically analyze interactions and link them to operational indicators. This allows us to move beyond simply tracking aggregate metrics (number of calls, duration, or abandoned calls) to analyzing what actually happened within each conversation. For example, it is possible to verify whether the customer requested human assistance, whether the call was transferred, whether the customer received specific information, or whether the incident was correctly logged.
This combination of automation and audit capabilities is particularly significant: one technology handles customer service, while the other makes it possible to verify how that service was provided.
From the traditional Contact Center to a continuity architecture
The SAC Law may accelerate a change that was already taking place in customer service centers. The traditional model is based on a relatively fixed number of human agents and uses IVR and automation to route calls. The new scenario allows for a more dynamic architecture, in which different layers work in a coordinated manner.
Intelligent agents can handle certain conversations and absorb call spikes. Human agents step in when the user requests it or when the complexity of the issue requires it. Corporate systems provide information and execute processes. And conversational intelligence records and analyzes what happened to generate evidence.
The goal should not be to automate the highest possible percentage of calls. It should be to ensure that every interaction reaches the appropriate resource and can be resolved within established service levels.
This also requires the design of contingency mechanisms. If an integration fails, if the intelligent agent lacks up-to-date information, or if the conversation takes an unexpected turn, the system must know how to respond and where to escalate the issue.
That is one of the differences between simply implementing a bot and building a true AI-based customer service architecture.
AI as an additional capability, not a replacement
Law 10/2025 does not pit automation against human customer service. In fact, it recognizes the possibility of using artificial intelligence systems to facilitate communications, while also protecting access to personalized service.
For companies, this opens up an opportunity to rethink how customer service operates. Intelligent agents can ensure availability, handle fluctuations in demand, automate processes, and expand the contact center’s capacity. Human agents can focus on cases that require interpretation, negotiation, empathy, or decision-making.
But for the model to work, both worlds must be connected. True evolution does not consist of replacing human agents with artificial intelligence, but rather of building customer service systems that are more resilient, scalable, measurable, and auditable.
In this context, service continuity no longer depends exclusively on how many people are available at any given moment. It depends on the ability of the entire architecture (people, intelligent agents, processes, data, and systems) to respond when the customer truly needs it.
Learn more about how platforms like Recordia can help you by clicking here.