Salesforce: Customer-Driven AI for Innovation

Salesforce Relies on Its Customers to Guide Its AI Strategy

In a world where artificial intelligence is evolving at breakneck speed, companies must innovate quickly to avoid being outpaced by more agile competitors. Salesforce, the giant in customer relationship management software, has adopted a unique approach to stay in the race: directly involving its customers in defining its AI roadmap.

While many companies seek to collaborate closely with their customers to improve their products, Salesforce's approach stands out due to the breadth of its network and the frequency of its interactions. The company does not settle for annual or quarterly meetings; it meets with some of its customers as often as once a week to gather their feedback and adjust its products accordingly.

Jayesh Govindarajan, Executive Vice President of Salesforce AI, recently emphasized the importance of this approach in an interview. According to him, Salesforce's 18,000 customers represent an invaluable source of information that enables the company to succeed. He explains that the technology stack developed by Salesforce has resonated well with these customers, and that the continuous improvement of large language models (LLMs) and autonomous agent systems is a long-term innovation path in which the company is investing.

The Rise of AI Agents at Salesforce

Salesforce was one of the first companies to launch AI agent management software at the end of 2024, anticipating the rise of agent-based AI. Since then, the company has ramped up its efforts and continues to roll out new products, including for voice AI and Slack, at a steady pace.

The company attributes the speed of its product launches to the involvement of its customers in the development process. By allowing them to guide the creation of the AI product roadmap, Salesforce can respond quickly to technological advancements.

Muralidhar Krishnaprasad, President and Chief Technology Officer of Engineering at Salesforce, explained that when large language models were introduced, companies were eager to adopt them but lacked the necessary technology to fully leverage them. This gap led Salesforce to launch its agent management platform, Agentforce.

A Customer-Centric Strategy

The need for last-mile technology has driven Salesforce to adopt a bottom-up strategy focused on themes such as agent context, observability, and deterministic controls, rather than specific product timelines. This approach relies on direct feedback from rotating customer groups to develop products tailored to similar needs in other businesses.

Customers as Drivers of Innovation

Jayesh Govindarajan highlighted that Salesforce's innovation is directly linked to its collaboration with a wide range of customers and the classification of the problems they encounter in the real world. The company breaks down these issues to determine which can be solved at the LLM level and which require additional components for agent-based systems.

This close collaboration with customer engineering teams allows Salesforce to quickly resolve issues before technology surpasses them. Muralidhar Krishnaprasad explained that the company cannot afford to wait several months for feedback and to resolve problems. It must respond continuously, week after week, to adapt to the rapidly changing technological environment.

The Engine travel management platform is an example of this customer feedback dynamic. Elia Wallen, founder and CEO of Engine, explained that the company's operations team meets with Salesforce every week, allowing Engine to access AI tools before their official release. This collaboration gives Engine a competitive edge by enabling it to derive more value from AI tools.

The relationship between Engine and Salesforce is reciprocal. Wallen noted that feedback from Engine has been integrated into Salesforce's tools. For example, after noticing that the interaction of an AI voice agent seemed artificial, Wallen shared his observations with Salesforce, which quickly adjusted the agent, thereby improving A/B testing results.

"If someone is willing to help create and build products we need, they can better assist us and truly understand our problem and how they can solve it," Wallen said. This strategy also allows Salesforce to deploy user-built solutions and workflows to its entire customer base.

A Successful Collaboration with PenFed

The credit union PenFed has also benefited from this collaborative approach. Shree Reddy, Chief Innovation Officer and Executive Vice President of the company, explained that PenFed was able to reduce its technology stack by working closely with Salesforce.

Reddy stated that PenFed developed an IT service management (ITSM) workflow using existing tools and agents in Agentforce, which worked well for the company. Salesforce was able to see this success and deploy the tool on its broader platform for other businesses.

However, this approach presents challenges. It relies on the idea that the customer is always right, which is not always the case. Salesforce hopes that its customers are correct, although many companies are still trying to determine the role AI will play in their business. Some have yet to find value in the technology, which could limit their ability to influence long-term product development.

Moreover, being willing to test and preview technology in beta does not guarantee long-term usage habits or future software contracts.

Internal Innovation at Salesforce

Salesforce also applies this bottom-up approach internally. Jayesh Govindarajan stated that Salesforce employees are the biggest users of its AI tools.

The company reallocated staff and resources at the beginning of the AI boom. When ChatGPT was launched, Salesforce reorganized its teams and resources to create a new AI team. This strategy has proven effective during previous waves of innovation, according to Muralidhar Krishnaprasad.

"As technology evolves, we never know what will come out in a month," he said. "We will adapt. And that's what we've done throughout the last year. If you think about it, agents weren't even in the vocabulary a year and a half ago. And then we had to react. We had to respond to all the advancements, and we had to respond to our customers."