IT Leaders at TDX 2025: Don’t Get Left Behind in AI

IT leaders recently convened at TDX, Salesforce’s annual developer conference, to discuss the impact of agentic AI on their businesses.

Why it matters: Just 11% of CIOs say they’ve fully implemented AI — 18 to 38 percentage points less than their line of-business counterparts. The discussion revealed both the challenges and opportunities of AI adoption, and emphasized the need for speed alongside data quality, trust, and a unified platform approach.

Driving the news: The panel featured executives from OpenTable, Copado, and Saltbox, who are each using Agentforce as their platform for building and deploying agentic AI. Together with Salesforce technical leaders, the panel discussed how AI agents are improving productivity, efficiency, and innovation.

What they’re saying:

• “AI isn’t a passing trend — it’s here to stay. The real question is, how fast can you move?” said Shayne Smyth, CTO at Saltbox.
• “AI doesn’t belong everywhere. The key is finding bottlenecks — where time is wasted. It could be streamlining discovery, refining user stories, or even automating regression testing,” said Gloria Ramchandani, SVP, Copado.
• “When you build an API that instantly answers questions and integrates seamlessly‌ — ‌whether it’s OpenTable or another system‌ — ‌that’s where the real value of AI shines,” said George Pokorny, SVP, OpenTable.

Salesforce Chief Engineering Officer Srini Tallapragada added additional POV. AI outputs and actions, he explained, are only as good as the data it’s being fed. And for most companies, data is siloed in a way that makes it nearly inaccessible.

• “The problem isn’t a lack of data — it’s that most of it is trapped. APIs are locked away, and data lakes are inaccessible to the people who need them most. Right now, only data analysts can unlock insights, but for business users, it’s like searching in the dark,” said Tallapragada.

Between the lines: Data silos, point-solution apps, and too many systems are like disjointed parts that companies have to self-assemble, with escalating security risks and costs. Businesses need a deeply unified platform that combines apps, data, AI models, and agents into a cohesive system. 

Salesforce execs also highlighted their own internal deployment of Agentforce, explaining how the company treated the help.salesforce.com channel as a real-world pilot for the technology.

• “We started off by defining the five steps to building an agentic system by bringing in actions and data. There is a reverse angle to it … If you turn the problem upside down and you think of all the data and the action space available to an agentic system, the agentics will come and tell you in the builder what’s possible,” said Jayesh Govindarajan, Salesforce SVP of AI.

If you turn the problem upside down and you think of all the data and the action space available to an agentic system, the agentics will come and tell you in the builder what’s possible.

Jayesh Govindarajan, Salesforce SVP of AI

What’s next: Salesforce used TDX as a moment to announce upgrades to Agentforce and its ecosystem.

• Agentforce 2dx: The newest version of Agentforce expands beyond the reactive, user-initiated world of chat interfaces, enabling proactive AI agents to work behind the scenes, without constant human oversight, to unlock new customer and employee workflows of any kind. 
• AgentExhange: Launched with more than 200 initial partners and developers, plus ready-made actions, topics, and template, this marketplace and community allows Agentblazers to participate in the rapidly expanding $6 trillion digital labor market.  

Go deeper:

• Read about the $6T digital labor opportunity — and what it will take to capitalize on it
• Watch CIOs go in-depth on AI in CIO Corner episodes featuring executives from PWC, Medtronic, and more
• Read how CIOs are helping employees adopt agentic AI at work
• Find out how digital labor will reshape the enterprise