Not only do AI agents improve employee productivity, but they also enable organizations to challenge themselves
Why AI agents don’t spell the end of SaaS, but are transforming the role of enterprise software. Bob Vanstraelen, CEO, EMEA North, Salesforce, shares his experience after creating his own team of AI agents.
“Dashboards, forecasts, customer insights, pipeline development—I’ve been a loyal user of our own CRM platform for years. However, I wanted to use my time more efficiently and leverage insights more intelligently to prepare for meetings, assess progress across different countries, and set priorities. So I decided to create a team of AI agents—not to replace the system itself, but to make it easier for me to access and use the information it contains.
“In fact, my use of the platform hasn’t decreased. On the contrary, I now consult business information about ten times more often than before. The platform provides me, more than ever, with information that’s essential for making critical decisions.
“And that’s an interesting paradox. It’s often assumed that AI will relegate enterprise software as we know it to the background—that there will be a ‘SaaSpocalypse.’ My experience proves exactly the opposite: AI makes these systems more accessible and ensures that the information is used more often in decision-making.
Five agents, one team
“Agent-based AI doesn’t spell the end of SaaS, but rather drives a real transformation of CRM. How have these agents improved my work? They combine data and speed up access to information. Meetings are better prepared, and I can more easily determine where to focus my attention for the greatest impact. Be aware, however, this is not the work of a single agent, but of a group of assistants who collaborate and each have a role to play.
“So what exactly do these agents do? My first digital assistant reviews revenue trends like a sales director, asking critical questions about the current quarter. The second agent monitors staff capacity and turnover. The third looks ahead and analyzes the upcoming pipeline. The fourth acts as the digital chief of staff and helps prepare for and follow up on meetings. Finally, the last agent always takes the management perspective: where can we make the biggest difference?
“Although this setup may resemble a digital management team, the agents don’t make any decisions; they gather information, identify patterns, prepare options, and list possible actions. It’s always up to me to make the decisions and take responsibility for them.
Much More Than Just Answering a Question
“For me, the greatest value of AI agents doesn’t lie in a single spectacular application, but in dozens of small moments when I can access the context I need more quickly. They don’t work in isolation; they’re an integral part of how I lead. For example, when I need to determine which issues and initiatives will have the greatest impact, draft a briefing for a client meeting, conduct an updated forecast analysis, or document decisions and next steps after a meeting.
“They not only save me time but also help me make better decisions. Moreover, a good AI agent doesn’t just answer questions. It also highlights connections, anomalies, or risks that I might not have noticed—without me even having to ask. As a result, I’m better prepared for meetings and consult company information much more often than before. This broadens my perspective when I make decisions—without abandoning existing systems. It’s the interface that’s changing, not the foundation. I do less research, but I use the underlying systems more often.
The latest AI model isn’t the key to progress
“I also don’t think organizations stand out by opting for the latest AI model. This technology is becoming more widespread at an ever-faster pace. The real competitive advantage lies in the quality of the data, the clarity of the processes, and the context you can provide to your AI. Is the customer data complete and up to date? Are definitions consistent across the organization? Do we know who is authorized to access which information, and can we verify the basis for a response?
“Without this foundation, an agent that may seem convincing can draw the wrong conclusion. In such a case, it’s not intelligence that’s automated, but confusion. This may be the most important lesson to learn from AI agents. Not only do they improve employee productivity, but they also enable organizations to challenge their own assumptions. Inconsistent data, fragmented processes, and unclear responsibilities become immediately apparent.
“Anyone who wants to start using AI seriously should therefore not begin by choosing a specific model, but by asking whether the organization is ready to provide the AI with reliable information. It’s less spectacular than implementing the very latest AI technology, but that’s precisely what makes the difference.
Start with a real decision
“My advice to other executives is not to wait for someone to bring them a complete AI solution. Choose a recurring management issue yourself that requires too much manual work, consultation, or research. Create an initial agent around it and link it to information already considered reliable. Then check not only whether the answer seems plausible, but also where it comes from, who is authorized to use it, and what action it requires.
“Ultimately, you’ll be able to leverage AI more quickly this way than if you wait for the perfect model. The model facilitates thinking and communication, but the quality of the result depends on everything the organization has built: data, context, processes, permissions, and trust.
“My AI agents have changed the way I organize my workday. They’ve made me faster and more accurate, but also more critical. And that may be their main contribution: they provide answers, but they also force the organization to bring order to the information underlying those answers. So the best AI strategy doesn’t start with the smartest agent. It starts by asking whether that agent is built on a solid foundation!”
