Salesforce hands its AI roadmap to customers, speeding up Agentforce and new feature releases

Salesforce is changing how it ships AI products: instead of a long-term plan, the company builds its roadmap together with customers and rewrites priorities almost in real time. The approach is meant to help it adapt Agentforce, voice AI and Slack integrations to the real-world tasks of large companies faster.

Customers instead of a plan

The pace of AI development is forcing enterprise vendors to release features much faster than before. Salesforce concluded that in this environment, a classic roadmap that looks quarters ahead moves too slowly. So the company relies less on fixed deadlines and more on a constant stream of feedback from customers. According to Salesforce AI executives, some customer meetings happen every week, not once a quarter.

For Salesforce, this is not just customer development in the usual sense. The company believes that if one large organization hits a specific problem while adopting AI, others will soon feel the same pain. Internally, this works as an early-warning system for needs: the team watches which problems recur and turns them into product themes for the whole platform.

How the loop works

Agentforce was launched because of a simple problem: once large language models appeared, companies lacked the "last mile" needed to turn AI into a working enterprise scenario. The model alone is not enough. It takes context, control, observability and clear rules so that an agent does not behave unpredictably. So Salesforce is building its AI stack from the bottom up, and plans directions of development rather than specific release dates.

• context for AI agents and access to company data
• observability: the ability to track what an agent is doing
• deterministic controls: constraints and predictable rules
• early beta tests and "gates" before a broad release
• a fast code release cycle measured in weeks and months, not half-years

"We are literally reacting week by week, month by month," is how

Muralidhar Krishnaprasad, president and CTO of Salesforce's engineering division, described the new working rhythm.

The next step is to break each customer request into parts: what the LLM layer can solve on its own, and what requires additional agent infrastructure around the model. This matters in enterprise settings, where a polished demo feature is not enough. Companies need a product they can control, verify and embed in their existing stack without constant manual oversight.

First results and risks

One participant in this loop is the travel platform Engine. Its team meets with Salesforce weekly and gets access to AI tools before their public launch. According to Engine's CEO, this helps the company test new capabilities ahead of competitors and genuinely influence the product. One example is a voice agent for hotel bookings: after feedback that the conversation sounded unnatural, Salesforce refined the agent's behavior, and A/B tests showed a better result.

PenFed has a similar story. The credit union built its own ITSM workflow on top of existing tools and Agentforce agents. When the solution proved itself in real use, Salesforce moved that scenario into the platform as a more general-purpose tool for other customers. In other words, customers here act not only as testers but as co-authors of finished enterprise features.

This model has a downside. It assumes the customer clearly knows what AI it will need a year from now, even though many companies are still searching for practical returns from such systems. A willingness to sit in a beta and discuss the product every week also does not guarantee a long-term contract or mass adoption. Salesforce reduces this risk by actively using its own AI tools internally and constantly reallocating teams to match new technology shifts.

What this means

Salesforce shows what enterprise AI development can look like when the market changes faster than quarterly planning. The winners here will not be those with the prettiest roadmap slide, but those who turn recurring customer pain points into working products for the whole customer base.