How Should You Get Started on Agentforce
Most people we talk to share a common challenge: they recognize the potential benefits of Agentforce and are eager to explore but are uncertain about how to proceed. Many struggle to bridge the gap between Salesforce’s AI vision and the tangible value it can deliver to their organization.
These conversations often highlight recurring themes: determining the most effective Agentforce use case for the business, assessing the effort required to bring it to market, and understanding the financial implications.
Effective Agentforce use cases
Salesforce’s Agentforce offers businesses customizable AI agents that automate routine tasks such as order tracking, customer support, and troubleshooting. These agents integrate seamlessly into various communication channels, improving operational efficiency and delivering personalized service.
As AI handles simple tasks, it frees human agents to tackle more complex issues. The platform also uses AI for knowledge assistance and can escalate complicated queries to human agents when necessary, ensuring that businesses can provide the right support to their customers. For more details, check out Salesforce’s Agentforce Use Cases.
At N28, we have successfully assisted clients in a variety of impactful use cases to get started on Agentforce. Here are some examples:
- Sales and marketing use cases
- Generate quotes: Where our AI agents interact with Sales staff to produce customized quote PDF documents for specific customers.
- AI Account manager: Recommend products based on insights from current orders and market data, optimizing sales strategies.
- Operational use cases
- Validate purchase orders: Automating tasks like price checks, identifying complex offer pricing, and initiating rebate calculations
- Admin AI: Automates routine administrative tasks within Salesforce orgs, saving time and reducing manual effort.
Technical framework behind Agentforce
Agentforce’s architecture is built on four fundamental components that work in harmony to deliver intelligent, secure, and effective agent-based solutions:
- Data for AI: Foundation begins with comprehensive data integration capabilities, supporting both structured and unstructured data from internal and external sources. This robust knowledge base serves as the fundamental building block that powers agents’ decision-making and operational capabilities.
- Large Language Models (LLMs): Leverages state-of-the-art LLMs to transform raw data and analytics into natural, conversational interactions. This technology enables fluid communication between users and agents, making complex data accessible and actionable through intuitive dialogue.
- Agent Core: A sophisticated suite of tools for designing customized agent experiences. This framework enables organizations to:
- Define Topics: Create precise agent parameters through detailed instructions, descriptions, and clearly defined operational scope
- Configure Actions: Establish specific, executable tasks and operations within each defined topic area
- Trust framework: A comprehensive security and governance layer that ensures controlled access management, secure channel operations, compliant agent interactions and protected data handling.
This integrated architecture ensures that Agentforce delivers not just powerful AI capabilities, but does so within a secure, controlled, and trustworthy environment.
Other common questions
- Data Cloud: Agentforce leverages the power of Data Cloud while maintaining implementation simplicity. While Data Cloud activation is a prerequisite for Agentforce deployment, organizations can begin operations using solely their CRM data—eliminating the need for extensive Data Cloud configuration at launch.
Key architecture points:
- Data Cloud houses the Vector Database infrastructure required for RAG (Retrieval Augmented Generation) implementations
- Initial deployment can operate efficiently with existing CRM data
- Flexible architecture allows for gradual expansion of Data Cloud functionality
This approach enables rapid deployment while preserving the option for enhanced Data Cloud integration as organizational needs evolve.
LLM options:
Out-of-box support:
- Pre-configured integration with select OpenAI models
- Pre-configured integration with select Anthropic models
Flexible deployment:
- Support for custom LLM integration (“bring your own model”)
- Maintain control over your preferred large language models
This flexible architecture allows organizations to leverage either pre-integrated models or implement their preferred LLM solutions based on specific requirements.
Financial overview
- License cost: Unlike traditional user licensing, Agentforce employs a conversation-based pricing structure, charging $2 per agent interaction, with potential volume discounts. This model shifts costs from upfront licensing to usage-based fees, potentially reducing initial expenses but increasing costs with higher usage. There might be additional costs associated with Data Cloud credit usages and LLM model usages. With this on-demand pricing model, organizations must evaluate the trade-off between automating tasks through Agentforce and the associated per-interaction fees to determine the most cost-effective approach.
- N28 implementation cost: Staged implementation approach. Understanding that AI adoption requires careful consideration, we’ve designed our AI services to support a measured, step-by-step implementation journey. Click the link below to review our Agentforce offerings.
