Salesforce's introduction of Agentforce, with its freemium model for foundational AI tools, marks a significant shift in the company's AI strategy. This approach not only generates excitement but also cultivates an environment of experimentation, feedback, and continuous product improvement. By democratizing AI access within the Salesforce ecosystem, Agentforce is poised to drive widespread adoption and innovation.
The Power of Freemium: Democratizing AI Access
Previously, many Salesforce AI tools, like Prompt Builder, were accessible only through paid licenses. This limited access for many Salesforce professionals, hindering experimentation and exploration of potential applications. Agentforce's freemium model addresses this by providing tools like Agent Builder, Prompt Builder, Model Builder, and Data Cloud to all Enterprise Edition and above customers. This empowers the vast Trailblazer network to directly engage with AI.
This strategic decision has several positive implications:
- Encouraging Experimentation: Freemium access encourages Salesforce professionals to experiment with AI agents and use cases within their organizations, providing invaluable hands-on experience.
- Driving User-Generated Content: Free access will likely trigger a surge of user-generated content (blog posts, videos, etc.) sharing real-world experiences and insights.
- Fostering Innovation: User experimentation will uncover new applications and possibilities, revealing practical use cases that enhance the platform's utility.
- Gathering Direct Feedback: Opening the platform provides Salesforce with valuable feedback from a diverse user base, crucial for refining the product's usability and effectiveness. This aligns with Marc Benioff's emphasis on direct feedback.
- Upskilling the Ecosystem: The freemium model allows Salesforce professionals to upskill and prepare for the AI-driven future, ensuring the ecosystem remains at the forefront of technological advancement.
Understanding Agentforce Pricing: A Consumption-Based Model with Nuances
While foundational Agentforce tools are freemium, the pricing for using AI agents in business applications is more nuanced than initially presented. It follows a consumption-based model with key distinctions between internal and external use cases (i.e conversations):
- External Use Cases: For external deployments (e.g., website agents), the pricing is $2 per conversation. Each interaction with the agent is counted as a conversation.
- Internal Use Cases: For internal use cases (e.g., HR agent, expense management), the $2 per conversation model doesn't apply. Instead, businesses need to purchase Service Cloud seats for every user who interacts with the agent.
- Bundled Usage: Agentforce SKUs often come with "conversations" and "Einstein requests" bundled together, adding complexity to the pricing structure.
Important Considerations:
- Clarity and Transparency: Despite the freemium model, there's still confusion surrounding Agentforce pricing. Clear and comprehensive documentation from Salesforce is crucial to avoid misunderstandings.
- Einstein Requests and Data Cloud Credits: These remain key components of Agentforce pricing, especially for internal use cases. Businesses must understand how these are consumed and the associated costs.
Einstein Requests: A Deeper Dive
Agentforce and other Salesforce generative AI features often consume Einstein Requests. These requests represent a unit of measurement for the computational resources utilized when interacting with large language models (LLMs) through the Einstein platform. Understanding how these requests are calculated and consumed is crucial, as they directly impact the cost of using AI features within Salesforce.
Key Aspects of Einstein Request Consumption:
- API Calls as the Basis: Every time a generative AI feature is used (e.g., an Agentforce agent generates a response or Prompt Builder creates content), an API call consumes Einstein Requests.
- Usage Type Multiplier: The type of LLM used influences the number of Einstein Requests consumed per API call. As of June 2024:
- Starter: For "Bring Your Own Large Language Models" (BYO-LLMs), the multiplier is 7 Einstein Requests.
- Standard: For Salesforce-enabled foundational LLMs, the multiplier is 10 Einstein Requests.
- Varying Consumption: Different AI interactions have varying consumption. Simple queries consume fewer requests than complex conversations. Generating longer content consumes more requests than generating shorter content.
- Cost Management: Understanding this calculation is crucial for cost management, especially with BYO-LLMs. Optimize prompts and responses to minimize word count and reduce consumption. Compare the rate card options for Standard and Starter to determine the best fit for your needs.
- Impact of Usage: Increased reliance on generative AI for tasks like content creation, data analysis, or conversations will increase Einstein Request consumption. Monitor usage patterns and adjust your approach as needed.
- Integration with Other Salesforce Services: Generative AI usage may also consume Data Cloud credits.
Relationship to Agentforce Pricing:
- While a "conversation" is the unit of measurement for Agentforce agent use in real-time, individual API calls within the agent interaction are measured with Einstein Requests.
- A single conversation could involve multiple API calls, each consuming Einstein Requests.
Implementation Costs
Beyond usage, consider implementation costs:
- Partner Network: Salesforce's growing partner network offers pre-built agents, actions, and skills on the AppExchange, allowing customers to leverage third-party tools or build their own.
- Consulting Costs: Salesforce is investing with Systems Integrators (SIs) to assist customers with agent building and deployment. Consulting services can facilitate smooth implementation and employee training.
- Customization: While pre-built options exist, companies may need to customize agents to fit unique business processes, potentially requiring specialized skills.
- Training Costs: Salesforce offers free AI training on Trailhead and instructor-led courses through 2025. However, internal training and process changes will still be necessary.
Impact on the Ecosystem
Agentforce's freemium model and pricing structure will significantly impact the Salesforce ecosystem:
- Increased Adoption: Lower barriers to entry will drive higher AI tool adoption rates.
- Faster Implementations: Consumption-based pricing incentivizes faster implementations.
- Partner Opportunities: The partner network plays a crucial role in customer adoption and customization, creating significant opportunities for co-creation and consulting services.
- Value-Driven Approach: The focus on ROI ensures AI is viewed as a tool for driving business results.
- Shifting Perceptions: Agentforce's positioning as a practical solution to real-world challenges may alleviate concerns around the "AI hype cycle."
Conclusion
Agentforce's freemium model is a strategic move by Salesforce to drive AI adoption and foster ecosystem innovation. By providing free access to foundational AI tools, Salesforce empowers its community to experiment, provide feedback, and create new solutions. The consumption-based pricing model, with its nuances for internal and external use cases, aims to ensure businesses pay only for the value they receive. This approach transforms AI perception and usage, moving the ecosystem toward a future where AI is an accessible, practical tool for driving business results.
