Salesforce Exec: How CIOs Can Overcome Hurdles Around Agentic AI

Reflecting on industry conversations this year, Salesforce President and Chief Engineering & Customer Success Officer Srini Tallapragada shared that CIOs are eager to embrace agentic AI, saying, “Everybody wants to go in on an AI strategy.” In fact, CIOs are bullish on the technology, with 66% reporting confidence that AI will drive ROI. But they still have questions and concerns.

From vague definitions of agents to integration challenges, CIOs want to be certain this is the high-impact technology for their business before committing to implementation.  

Why it matters: As AI startups and solutions flood the enterprise, CIOs need robust solutions to provide the quality and productivity that AI promises, not just cool demos. With the deeply unified Salesforce Platform and Agentforce, its agentic AI layer, Salesforce addresses these real-world needs with a comprehensive solution for deploying trusted AI agents at scale. 

What they said: Tallapragada’s conversations with CIOs across the globe reveal key pain points:

• Distinguishing real from false agents: Agentforce’s access to enterprise data sets it apart from other “agents” on the market, Tallapragada said. While other companies may allow you to actively curate data inputs for your model, Salesforce goes further to let users decide how deeply they want their agent to think through the question before responding. The powerful combination of data and metadata in Data Cloud, ‌ Salesforce’s hyperscale data engine, along with the Atlas Reasoning Engine in Agentforce, gives agents the ability to query and reason, making their outputs smarter and more trustworthy.
• Making data strategy investments: Tallapragada noted CIOs are usually reluctant to change established data strategies, which typically involve significant investments in data lakes, warehouses, and governance frameworks. They’re also frequently skeptical of any new AI platform that requires a fundamental overhaul of their existing data infrastructure. Data Cloud helps ease this concern by seamlessly integrating structured and unstructured data, data lakes, warehouses, and customer relationship management (CRM) data. An open and extensible platform, it’s also integrated with Salesforce’s Zero Copy Partner Network ecosystem, so data can be actioned in Agentforce and across the Salesforce Platform.
• Implementing agents across the enterprise: CIOs have also told Tallapragada they need AI that works across the enterprise, not just within customer-facing functions. CIOs are responsible for the overarching technology strategy of the entire enterprise, which extends far beyond CRM. When it comes to agents making an impact at the broader enterprise-level, only Agentforce offers a complete AI system that connects data, AI, automation, and humans to deploy AI agents for concrete business outcomes.   
• Addressing compliance concerns: CIOs in regulated industries — from finance to healthcare and government — face unique pressures to ensure responsible AI use, Tallapragada said. They use must provide robust assurances and meet the most stringent compliance requirements. The Salesforce Platform addresses these challenges by providing tools for innovation, security, and operational optimization, with Agentforce ensuring secure, scalable AI. 
• Managing vendor lock-in fears: Tallapragada said CIOs are often wary of becoming overly dependent on a single vendor, fearing vendor lock-in and limitations on their flexibility and choice of technologies. Salesforce’s open and extensible platform allows customers to integrate with their existing data lakes and various LLMs, he said.

Between the lines:

Tallapragada shared that scaling to an enterprise-grade AI platform requires Agentforce, which offers robust development, operation, and governance. 

• “We allow you to link … it’s not just integrated but also it’s open,” he emphasized. “This ensures you can use existing data and even bring your own AI models, avoiding vendor lock-in.”

Mike Hite, Chief Technology Officer at Saks, agreed, saying on CIO Corner:

• “Being able to leverage something like Agentforce, to really leverage that data and action, that data is what’s going to be a differentiator for us. It allows us to meet the customer where they want to be.”

Being able to leverage something like Agentforce, to really leverage that data and action, that data is what’s going to be a differentiator for us.

Mike Hite, Chief Technology Officer at Saks

Prakash Kota, CIO at Autodesk, had a similar take.

• “We’re building around Agentforce. We’re implementing agents inside the company for multiple applications, whether it’s customer support or customer success,” Kota shared on CIO Corner. “And we’re starting to see how our employees are taking advantage of these capabilities much more effectively.”

Go deeper:

• Read how digital labor will reshape the enterprise
• Discover key considerations for integrating AI agents responsibly
• Delve into why metadata is so critical for building an AI agent ecosystem
• Find out how Agentforce works
• Read how Agentforce is inventing the future of agents