Salesforce’s Smarter AI Approach: You Can’t Just Throw an LLM at a Problem

TL;DR: At Dreamforce 2025, Salesforce emphasised that deploying enterprise AI requires strategic integration and governance—not just throwing LLMs at problems. With 95% of AI projects failing (MIT study), the company’s “agentic enterprise” approach grounds AI in proper oversight, using internal deployment as proof-of-concept (1.8M conversations handled in one year).

Mike Moore reports from Dreamforce 2025 where Salesforce CTO Muralidhar Krishnaprasad outlined the company’s vision for responsible AI adoption that balances innovation with realistic implementation expectations.

Context and Background

AI has become commonplace in enterprises worldwide, transforming workflows and bringing productivity gains. However, successful implementation requires more than simply deploying models—it demands strategic thinking about how AI integrates with existing systems and governance frameworks.

Dreamforce 2025’s theme centred on the “agentic enterprise” concept, with Salesforce CEO Marc Benioff declaring that merely deploying enterprise AI models is insufficient unless properly integrated and grounded in governance.

Muralidhar Krishnaprasad, President & CTO of Engineering at Salesforce, explained the expansion strategy: “Previously we were always just relegated to saying you are only doing sales, or service marketing—but now we are going beyond that, we’re managing customers, we’re managing employees, and we’re going to be managing operations…and most excitingly, we’re managing agents as well! All managed by our unified platform.”

The Agentic Enterprise Philosophy

Salesforce positions itself as “customer zero” for its AI tools, testing internally before wider deployment. Krishnaprasad highlighted their Agentforce launch experience: building and launching support agents in just weeks, which subsequently handled 1.8 million conversations in one year, freeing human workers for higher-priority cases.

This internal deployment demonstrates practical commitment to the approach Benioff described: being “on the journey” with customers toward the agentic enterprise age.

The problem with rushed AI adoption: An MIT study cited by Krishnaprasad reveals 95% of AI projects fail. The primary reason? “You can’t just take an LLM and throw it at a problem,” he states.

Strategic AI Implementation Principles

Krishnaprasad draws historical parallels to address AI adoption fears: “When the dotcom came we all feared for our jobs, when Tesla introduced auto driving, we all wondered what would happen—but the reality is—humans are the best at adapting—over millennia, we have adapted to so many different things.”

He positions AI as transformational technology comparable to electricity, fundamentally changing how work happens rather than eliminating human contribution.

Developer productivity example: Research shows 40% of developer time is spent maintaining code rather than creating innovation. “If AI can really help solve a whole bunch of issues there, imagine what new things we can create in just 20 years,” Krishnaprasad notes. “This acceleration is happening because of technology, and with AI, it’s just going to go even faster.”

Human-AI Collaboration Vision

Salesforce’s optimistic vision sees AI agents and humans working alongside each other, maximizing respective strengths. By automating maintenance drudgery, AI enables focus on innovation and new interaction methods.

“We will be able to leave out the drudgeries of maintenance, instead focusing on innovation, new ways of interacting, new ways of helping our human race together,” Krishnaprasad explains.

For SaaS companies like Salesforce, the approach mirrors how they “made the dotcom useful for the enterprise”—applying similar principles to make AI genuinely productive for enterprise deployment.

Looking Forward

The agentic enterprise concept reframes AI from tool to integrated system requiring governance, strategic deployment, and realistic expectations. Salesforce’s internal-first approach demonstrates commitment to proving concepts before wider customer deployment.

As enterprises navigate AI adoption, the lesson is clear: success requires more than deploying the latest LLM—it demands thoughtful integration, proper governance, and understanding where AI genuinely adds value versus where it creates unnecessary complexity.

Source Attribution:

Share this article