Agentforce Adoption Struggles: Data Readiness, Not AI, Is the Barrier

What happened

Analysts at KeyBanc Capital Markets have published a cautious assessment of Salesforce's Agentforce AI platform, concluding that the product is struggling to convert enterprise interest into committed deployments. According to reporting by The Register, the investment bank's research found that prospective clients are being held back by two compounding problems: the quality and readiness of their own customer data, and a product that KeyBanc characterised as not yet sufficiently mature for reliable enterprise use.

Salesforce has pushed back firmly on that characterisation. The company maintains that Agentforce is the fastest-growing product in its history, a claim that frames the KeyBanc note as out of step with what Salesforce says it is seeing in its own pipeline and early deployments.

The tension between the two positions reflects a broader pattern playing out across enterprise AI: vendors reporting strong momentum at the top of the funnel while analysts and customers describe friction — and sometimes stalled projects — further down the adoption curve.

Why it matters

For CX and service-design practitioners, the KeyBanc findings surface a problem that sits well upstream of any AI tool: the state of a company's underlying customer data. Agentforce, like most agentic AI platforms, depends on clean, well-structured, accessible data to reason across customer journeys and take autonomous action. When that foundation is missing — inconsistent records, siloed systems, poor data governance — even a technically capable agent will produce unreliable or irrelevant outputs. The implication is that organisations rushing to deploy AI-powered service agents without first auditing their data architecture are likely to encounter exactly the kind of disappointment KeyBanc is describing.

From a behavioural-economics perspective, there is also a credibility and trust dynamic at play. Early adopters who experience poor agent performance are likely to develop lasting scepticism — a form of availability bias that can slow adoption across entire organisations long after the underlying product has improved. The reputational stakes for Agentforce, and for agentic CX more broadly, are therefore higher than a single product cycle might suggest.

By the numbers

• Fastest-growing product in Salesforce's history — the company's own characterisation of Agentforce's trajectory, cited in response to the KeyBanc analysis.

The Renascence take

The instinct in most CX conversations right now is to treat AI-agent adoption as primarily a technology problem. The KeyBanc note suggests it is primarily a data-readiness and change-management problem — which is a meaningfully different diagnosis, and one that most operators are not yet structured to solve.

The organisations most likely to succeed with agentic AI are not those with the most sophisticated models — they are those that have already done the unglamorous work of unifying customer records, defining clear service taxonomies, and establishing data governance. Salesforce can improve Agentforce; it cannot fix a client's fragmented CRM history for them. Customer-obsessed operators should treat this moment as a forcing function: before evaluating any AI agent platform, conduct an honest audit of your data estate. If your human agents cannot answer a customer question without toggling between five systems, your AI agent will not be able to either. The technology is not the bottleneck — your data architecture is.