When it comes to innovation, Cloud for Good has always been on the front lines. We’ve been a part of Salesforce’s Nonprofit and Education Cloud Design Partner group and are among the first to see the newest products, features, and enhancements in action. We’re consistently looking at how the latest and greatest innovations can be used to help enhance the impact sector’s technology.
Agentforce, Salesforce’s newest suite of autonomous, more advanced, accurate, and intelligent AI agents, is the latest evolution of AI business tools from Salesforce. It represents significant innovation of the Salesforce Platform and is slated to help augment tasks across any industry and drive customer success, efficiency, and increased productivity. The impact and potential that Agentforce has for nonprofits and higher education institutions have generated great interest and discussions at Cloud for Good. I recently sat down with Kestryl Lowery, Managing Director of Technology, and Ryan Blake, Solutions Engineer Team Lead. Here are our first impressions and thoughts about Agentforce, how it works, and its potential use cases in the nonprofit and higher education industries.
How Agentforce Works
Humans+AI+Data+Actions
Agentforce leverages generative AI and data access to provide an AI agent system with specific roles and proactive, actionable capabilities built upon Salesforce’s trusted security layer, which can function autonomously or alongside human agents.
“The whole value of Agentforce is that someone can be able to execute sometimes complex tasks by just typing instructions in plain language instead of needing to know where to navigate, where to click, how to apply specific logic, so it can make it a much smoother experience for someone to use, especially if they’re not familiar with the system,” Lowrey explained.
The AI of Agentforce is the Atlas Reasoning Engine, a large language model (LLM) designed to simulate how humans think and plan. This LLM powers Agentforce’s autonomous reasoning, decision-making, and task-completion abilities while also delivering results and actions that are accurate, precise, and meaningful. The Einstein Trust Layer, embedded natively in Salesforce, safeguards the customer data from Salesforce CRMs or data sources used for AI.
The data behind Agentforce is Data Cloud, which takes and harmonizes data from an organization’s internal and external sources in real time to feed an Agentforce agent with the right data to work with and operate with context and precision. This process helps Agentforce agents be intelligent and dynamic about an organization’s business processes and customers and deliver seamless AI experiences across customer touchpoints and systems.
Unlike a typical virtual agent, Agentforce agents can act across any system and connect data to action through metadata, defined workflows, and AI automation. This ability makes Agentforce agents capable of completing high-value, meaningful tasks that drive efficiency and productivity, including those taken on a human employee’s behalf.
Automation and Guardrails
Agentforce’s automation capabilities are made possible by its seamless integration with Salesforce’s existing automation capabilities and methods like MuleSoft, Salesforce Flow, and Apex. With these existing automation methods and actions, users can extend Agentforce’s functionality easily and build and execute new Agentforce automation with little effort. However, it is essential to remember that even though Agentforce’s agents have powerful, more precise, more intelligent automation capabilities that enable them to act on behalf of a human, there should still be a human handoff at some point in the process. Our experts noted that an agent still needs parameters and guidelines for what to do if and when an agent doesn’t have an answer to a prompt, gets confused, and needs more information to clarify a prompt.
“That’s where I’d be cautious about using an agent wholly autonomously at this point,” explained Lowrey, “but Agentforce is set up to have Guardrails, and they can be built into the agent, ‘Here’s when you need to hand off to a human.’”
“You can build on Guardrails where you can explicitly tell an agent what it can and can’t do,” said Blake, “And that goes to the extent of interacting with your Salesforce database, such as being able to trigger an action like an Apex trigger, a Flow, and then handing off to another live person or agent.”
Using Agentforce: Use Cases and Agent Types for Nonprofits and Education
Agentforce’s agents are out-of-the-box, scalable, and do not require any coding to customize or deploy. Through Agentforce’s Agent Builder, only a few clicks are needed to customize or set up an agent for any channel easily. Multiple pre-built agents are available in Agentforce; of these agents, here are the ones that we believe would be the most useful to our customers in the nonprofit and higher education spaces:
Fundraising + Advancement
For people who want to donate to a fund, an Agentforce agent can be programmed to help suggest relevant funds or fundraisers to a donor based on their interests. They could type in a question such as, “I’m having a hard time finding a fund I’m passionate about. Can you tell me what fund makes the most sense to me?” The agent could then ask questions like “Tell me about your interests.” It could take those interests, align them to a record in Salesforce, and identify matches between what the person said and what Salesforce records have identified and tagged appropriately. Then it will say, “Here’s a list of funds you might want to consider donating to within our organization or school and start to build that relationship with it.” This information can also be added to a record in Salesforce to give Gift Officers more insight into their donors.
Major Gifts and Grantmaking
One Agentforce agent, Sales Coach, uses Salesforce data and generative AI to provide personalized, realistic sales roleplays tailored to each deal, helping sales teams practice pitching, responding to objections, and negotiating. Although this agent is designed for sales professionals, employees who work in major gifts and grants could also benefit from using it.
It could benefit major gift officers by practicing proposals that will be successful with donors. The agent can look at the information you have on that donor and what’s been successful with them before, and essentially operate like a sales coach and give the gift officers feedback on their pitch. This could help them accelerate their major giving strategy and help with optimization questions like, “What should we emphasize in our grant application for this foundation?” by looking at data from past grants that they won and have not won.
Admissions + Program Management
Let’s say you are navigating a website to learn about all the majors available at your institution or programs available at a nonprofit. This scenario could be a great potential use case for Agentforce’s Service Agent, which replaces traditional chatbots with AI to engage with customers autonomously and resolve customer inquiries 24/7 in natural language rather than preprogrammed scenarios or if-then logic. Users can define clear guardrails for Agentforce Service Agents and when to escalate a customer inquiry to human employees. For an admissions website, a Service Agent could be customized and programmed to respond to a user asking what the best major program is for them by setting up a prompt asking the user what their interests are. In response to what the user provides, the Service Agent could try to align the best, most relevant major programs to the user’s interests. The same goes for programs at a nonprofit. If a constituent enters their needs, the Service Agent could potentially help them find the program that best fits what they are looking for.
Want to discuss your specific AI use cases with a Cloud for Good expert, receive the latest AI content and events, and take an AI Readiness Assessment?
Overall, these tools have a lot of potential for those in the impact space. Boards and donors have challenged many CIOs to embrace technology. Still, it comes back to identifying where this tool can accelerate mundane and repeatable tasks for staff so that they can spend their time doing the things they bring their intelligence and creativity to. Agentforce isn’t a tool that will put anyone out of a job—it will change the jobs people are doing so that they can be more impactful. Current and potential clients thinking about it that way will be well-positioned for success with AI.
If your nonprofit or institution wants to discuss how Agentforce can benefit and impact your institution or nonprofit, contact Cloud for Good today and contact one of our experts.
