While the potential of artificial intelligence continues to capture the imagination of the business world, a significant number of enterprises are finding their AI initiatives stalled in the pilot phase, unable to transition to full-scale production. This growing chasm between consumer-driven AI expectations and lagging enterprise capabilities has created what Salesforce UKI CEO Zahra Bahrololoumi describes as an “agentic divide.” During the Agentforce World Tour London, the company detailed a comprehensive strategy aimed squarely at dismantling the two primary obstacles hindering widespread adoption. By targeting the pervasive issues of poor data quality and the financial uncertainty of prevailing pricing models, Salesforce is positioning itself to bridge this critical gap and unlock the true potential of the agentic enterprise. The strategy hinges on a dual approach: fundamentally strengthening its data infrastructure and revolutionizing its commercial agreements to foster innovation rather than inhibit it.
Confronting the Foundational Barriers to AI Adoption
A significant culprit behind the high failure rate of corporate AI projects is the foundational data problem. Many organizations pursuing a do-it-yourself strategy discover that their underlying technology stack is a patchwork of systems “bolted on, not built in,” leading to fragmented and unreliable data sources. This lack of a cohesive, high-quality data layer prevents AI models from accessing the unified, trustworthy information they require to generate accurate and contextually relevant outputs. Consequently, these systems are prone to producing erroneous results or “hallucinations,” which severely undermines business confidence and makes leaders hesitant to entrust AI with critical, customer-facing operations. Without a solid, integrated data foundation that can feed AI agents with clean and comprehensive information, the promise of intelligent automation remains just out of reach, trapping promising initiatives in a perpetual state of experimentation without ever delivering tangible business value.
Compounding the technical challenges is a significant economic hurdle rooted in the prevalent pay-per-token pricing model. Bill Patterson, EVP of Corporate Strategy at Salesforce, explains that this consumption-based structure, common among hyperscalers, is ill-suited for the enterprise environment. Business leaders find it nearly impossible to forecast the “token cost” of a standard process, such as qualifying a sales lead or resolving a customer service ticket. This inherent unpredictability creates immense financial uncertainty, discouraging the very experimentation and scaling that AI adoption requires. The model also creates a fundamental misalignment of interests; while vendors profit from sheer usage, regardless of the project’s success, customers are left with potentially spiraling costs for initiatives that may still be in development or testing. This economic friction acts as a powerful deterrent, forcing many businesses to keep their AI ambitions limited in scope to avoid the risk of runaway expenses that are untethered to successful business outcomes.
Forging a Path with a Unified Data Strategy
In response to the data crisis, Salesforce is aggressively advancing its vision of the “agentic enterprise,” a framework built upon its core AI services, Agentforce 360, and a robust data foundation called Data 360. The cornerstone of this strategy is the recent, approximately $8 billion acquisition of Informatica, a move designed to dramatically expand its data capabilities. Patterson clarified the complementary roles of the two platforms, noting that while Data 360 excels at activating customer-centric data, Informatica is a powerful engine for activating the much broader universe of enterprise data. This strategic integration effectively provides Salesforce with a far larger pool of information to “taxonomize” and make accessible to its AI agents. The move bridges the critical gap between customer-facing information and siloed internal enterprise systems, from finance to supply chain management. This synergy was validated by pre-existing market demand, with clients reportedly requesting compatibility between the two platforms even before the acquisition was formally announced.
The strategic value of this enhanced data foundation extends directly to improving AI reliability and trust. On a recent earnings call, CEO Marc Benioff emphasized that the acquisition significantly strengthens Salesforce’s “Data Foundation,” a business segment projected to reach $10 billion. He explained that the combined power of Informatica, Data 360, and MuleSoft facilitates essential processes like data harmonization, integration, and federation. This unified approach is crucial for providing AI models with the accuracy, reliability, and deep contextual understanding they need to function effectively. By creating a single, comprehensive source of truth that spans the entire enterprise, Salesforce aims to “wipe out the hallucinations” that plague AI systems running on fragmented or poor-quality data. This focus on data integrity is not merely a technical upgrade but a foundational requirement for building the trust necessary for businesses to confidently deploy AI at an enterprise-wide scale for mission-critical tasks.
A New Commercial Framework for AI Scalability
Recognizing that unpredictable costs are a major deterrent to AI adoption, Salesforce has undertaken a radical overhaul of its pricing philosophy, moving decisively away from consumption-based models like tokenization. The company determined that such models were “not appropriate” for its business-oriented platform after an initial experiment with a flat fee per interaction proved flawed. This early attempt revealed that a “one size did not fit all” approach was inequitable, as it priced a simple, single-response query the same as a long, complex, multi-turn conversation. This valuable lesson prompted a fundamental shift in the company’s commercial strategy. The goal became to create a new paradigm that replaces financial anxiety with predictability, thereby encouraging customers to innovate and scale their AI deployments without the constant fear of incurring unforeseen expenses that could derail their projects and budgets. This represents a move toward a partnership model focused on enabling success.
This new philosophy materialized in a portfolio of flexible, outcome-oriented pricing structures designed to offer predictability and align costs directly with business value. The new options include an “Action Model,” which shifts the focus from paying per conversation to “paying for work that is done by the platform,” tying expenses to tangible outcomes. Another alternative is the “Flex Credit Model,” which caters to customers with more variable usage patterns. However, the flagship offering is the Agentic Enterprise Licensing Agreement (AELA). This comprehensive agreement was designed to completely “remove the anxiety” around AI scaling by providing customers with unlimited use of both the Agentforce AI platform and the Data Cloud. In doing so, Salesforce provided a clear and predictable path for enterprises to confidently experiment, innovate, and deploy AI technologies across their entire organization, finally empowering them to close the agentic divide.
