Investment analysts are questioning the maturity of Salesforce’s Agentforce platform, citing weak customer traction and significant implementation hurdles.
KeyBanc Capital Markets released a research note that highlighted a slowdown in adoption for the AI agent platform, attributing the dip to the product itself. According to the report, customer conversations have not been strong, and feedback has been largely negative. The research firm noted that their CIO survey found Salesforce being a standout for the wrong reasons, with more respondents expecting to deprioritize the vendor within their IT budgets over the coming 12 months.
Feedback from Salesforce partner and customer events revealed two primary issues. First, customer data is often not organized enough to support meaningful AI work. Second, the product itself has not reached a stage where users feel confident using it. Conversations with Salesforce partners indicate that proof-of-concept deployments are only now beginning to generate actual pipeline opportunities, suggesting a gap between testing and real-world sales.
Related: Tech leaders losing AI debate to staff
The findings stand in contrast to Salesforce’s sustained push to position Agentforce as its flagship enterprise AI offering. Since introducing the platform nearly two years ago, the company has expanded it with new foundation models, integrations, and deployment options, most recently introducing its Headless 360 strategy to make the platform available beyond conventional CRM workflows. This shift toward a more flexible consumption model has drawn scrutiny, as it complicates budgeting for CIOs.
Data readiness and pricing concerns
Industry analysts agree that pricing is a significant barrier to entry. Three pricing model changes in roughly 18 months have made procurement committees nervous. The latest consumption-based model is harder to budget for than traditional seat-based licensing. Gaurav Parab of NelsonHall noted that because the model charges for AI activity rather than outcomes, enterprises are under pressure to evaluate total cost of ownership and expected return on investment before expanding deployments. Most organizations prefer confidence in measurable results before committing to broader rollouts.
Data modernization has emerged as another major hurdle. Salesforce positions Agentforce alongside Data Cloud as the foundation for enterprise AI, but preparing that data foundation represents a significant part of the implementation effort. Greyhound Research Chief Analyst Sanchit Vir Gogia pointed out that Data Cloud cannot make incoherence disappear, and the platform depends on trusted, unified enterprise data to generate reliable outcomes. For many companies, cleaning up fragmented, duplicated, and inconsistently structured CRM data takes months of work that does not show up in a vendor’s deal count.
Related: Benefits and Risks of Cloud Hosting
Despite these challenges, Parab views the current pace as a timing issue rather than a structural failure. Enterprise AI adoption has consistently followed the maturity of data and governance, he said, suggesting that as organizations strengthen their data foundations, Agentforce adoption will broaden significantly. Chopra offered a more critical view, arguing that Salesforce needs to fix its go-to-market strategy to address the positioning gap. He noted that the complexity of implementation is too high for most mid-market buyers.
For CIOs evaluating the platform, the next six to twelve months should provide a clearer picture of its trajectory. Analysts suggest focusing on tangible indicators of enterprise adoption rather than headline announcements. Parab emphasized the importance of enterprise-scale production deployments, broader adoption across business functions, and stronger customer references. Chopra advised CIOs to measure containment rates in production—the percentage of agent interactions resolved without human escalation—rather than simply tracking token volume or deal counts.
