Salesforce Agentforce, the Ally for Scaling Services

Salesforce held Agentforce World Tour Madrid yesterday, an event to showcase the capabilities of its autonomous AI agents, which have already been deployed in more than 60 of the vendor’s customer projects. It is a mature technology that promises to revolutionise the way companies serve both their employees and their customers internally.

In front of more than 3,000 professionals, Enrique Polo de Lara, Senior VP and Country Manager of Salesforce Spain, explained the potential of Agentforce to scale the services of practically any organisation. He also introduced several use case presentations, such as the one on Caixabank, which we will summarise below.

Before doing so, it is worth explaining what Salesforce Agentforce represents and what it consists of. It is a platform that, despite being only a few months old (it was launched on the market in October 2024), has become a key player in the offering of the leading cloud software manufacturer.

Its architecture, natively connected to Customer 360 and Data Cloud, combines artificial intelligence with organisational data and task automation. In doing so, it enables the generation of autonomous AI agents that perform repetitive functions and assist human workers so they can focus on other tasks that provide more value.

These agents are capable of connecting to data sources, reasoning about that data and executing actions without human intervention.

But, according to the company, it is not about eliminating jobs, it is about boosting and scaling the services of any organisation to meet the demands of its customers. Even more so at a time when it is not easy to find qualified professionals with the right digital skills.

“Agentforce is the driver of a new way of approaching work in companies. More than 41% of the tasks performed by workers are repetitive or of little added value. Agentforce frees professionals from these tasks and increases efficiency, profitability and agility in organisations,” said Polo de Lara during the event.

Thanks to its construction on Customer 360 and Data Cloud, the deployment of these agents can be carried out transversally, but it is also possible to connect them so that they can ‘talk’ and understand each other.

In this way, companies can deploy agents in their sales, customer service or marketing areas independently, but also connect them if they wish, which represents a huge potential and an answer to the traditional silos (or poor communication) that used to exist between different departments. The executive assured that these capabilities will soon be extended to other areas such as finance, human resources and logistics.

CaixaBank, one of the pioneers with Agentforce

CaixaBank executives took to the stage to explain in detail how they have implemented their first agents.

The Spanish financial group, which integrates banking, insurance and investments, was one of the first organisations to work with Salesforce technology. They have been working together for 7 years, mainly to change the way they relate to their customers.

In their strategic plan to 2027 they have taken into account Agentforce’s cognitive capabilities to work more efficiently and in a more personalised way with them. Their spokespeople emphasised a very important detail: “We have mixed AI and Data Cloud capabilities with other technologies that CaixaBank already has, all in a joint way. Having technology partners is the only way to survive and be competitive”.

Indeed, Salesforce technology is open and allows the use of existing third-party developments and reduces time-to-market: it is not necessary to use Agentforce exclusively. For example, with Zero Copy it is possible to access data without copying it while maintaining complete privacy, which for a bank is of vital importance.

CaixaBank’s adoption of AI agents is a response to ‘increasingly autonomous customers, so it will be very challenging to meet their needs without the use of Agentforce’, as spokespeople said.

Caixabank is already employing different agents in customer service. He exemplified the case of a user who is interested in signing a mortgage. In a first stage online, they can upload the relevant documentation, such as payslips, tax returns and relevant data. This is where the first AI agent comes into play to confirm receipt of all these documents, process them and validate them. If any errors are found, it is able to autonomously draft an email to let the customer know and request the necessary changes.

Once everything is correct, the AI agent notifies the (human) manager and makes available all the information received, with the possibility of including other user data such as the places he/she has visited, his/her interests, etc.

At this point, the manager asks the agent to elaborate three mortgage proposals based on the features offered by the bank and their adaptation to the client’s needs. The results may include a success rate for each of the proposals. In other words, the same process that until now was carried out manually by personal agents can be achieved in much less time and, according to the bank, with the same quality of service.

The next process is to arrange an appointment between the user and the agent to discuss details, doubts and the signing of the mortgage, something that the agent also takes care of by providing different times when the agent is free according to his or her schedule. The AI agent (who may be another agent specialised in managing appointments) is able to access it and propose different options for the user to choose the day and time that suits him/her best.

Moreover, throughout this process, the bank can activate more AI agents to offer the customer any other product they may be interested in based on all the information collected. It is about closing a complex cycle in which everything is automated and which allows personal managers to focus on tasks that provide more value and enhance the customer experience.

According to Caixabank, which claims to have the largest customer service in Spain, the implementation of Agentforce allows it to solve a large number of administrative tasks, resolve cross-selling processes, identify opportunities that the personal manager did not have or prepare customer meetings with data. All with very reliable results, in his words.