5 Takeaways from Our Agentforce Implementation on Salesforce Help

Everyone wants an easy, frictionless experience when they need help. Nobody enjoys being on hold or hunting through discussion boards and articles for answers. But creating that seamless experience is no easy task. 

Scaling human support is costly, and there’s a point where self-service becomes overwhelming for the user. Customers want something better — the ease of asking a human their question combined with the speed of thousands of knowledge articles at their fingertips. AI agents are bringing those two things together.

At Salesforce, we’re bringing AI agents into how we work and connect with customers. We see a future where people and AI agents team up to make every customer interaction smoother and more meaningful. Today, we’ve already taken the first steps with our recent launch of Agentforce➚ on the Salesforce Help➚ site—a full-screen experience that makes getting support simpler and more intuitive.

Salesforce is Customer Zero for Agentforce

With over 60 million visits each year, Salesforce Help offers a vast range of product content through organized directories, search, and direct support. However, finding the exact answer isn’t always seamless. Customers still need to filter through thousands of search results to find the right answer. Our new AI service agent on Help, powered by Agentforce, changes this experience. We’ve made the process faster and more intuitive for our customers. 

How did we do it? We focused on key use cases, prepared our data, and rigorously tested to validate responses. Within a week, our first test agent was live, meeting real customer needs and transforming the experience. In the first month, Agentforce handled 81% of inquiries, doubling the performance of our previous bot. Implementing Agentforce taught us valuable lessons that any enterprise can apply to make AI work effectively. Here’s what stood out:

1. Agents need content

AI agents are only as good as the content they can access. Imagine expecting perfect answers from an outdated encyclopedia. Pretty useless, right? That’s why we’ve designed our AI agent to draw from a rich base of over 740,000 trusted, relevant pieces of content, powered by Data Cloud. This way, you get accurate, timely responses every time — making it easier to find the answers you need, faster.

2. Agents need all types of data

AI agents rely on both structured and unstructured data to deliver effective responses. Unstructured data, such as knowledge articles, provides context for complex questions. Structured data, like determining whether you are a Premier or Signature Success Plan customer, enables personalized recommendations. By using Data Cloud and our Zero Copy➚ network, we seamlessly connect different data sources, revealing that much of the context needed for self-service lies within unstructured data. This combination of both data types ensures that agents can deliver accurate, relevant responses tailored to your needs.

3. Always back agents with humans

AI is incredible, but sometimes you just need to talk to a person. Customers want to know that if their issue gets too complex or personal, there’s a human on standby. Agentforce handles the repetitive stuff, but we’ve made sure there’s a quick, smooth handoff to human support when it’s needed most. It’s all about finding that perfect balance.

4. Incremental deflection is gold

Self-service has been around for decades, but it typically requires users to search, browse through links, and read lengthy content — a process that can be effective but isn’t always simple. Agentforce streamlines this experience by making it easier and more intuitive. Not only do they enhance the customer experience, but they also drive incremental deflection, effectively reducing the volume of support requests. And that’s where the real gold lies: Agentforce helps us scale by keeping more inquiries within the conversational experience while maintaining quality and boosting satisfaction.

5. One experience across many agents

Getting the first agent up and running is exciting, but the real vision goes further. Imagine multiple agents acting as concierges across all digital properties. The key to success, however, isn’t the number of agents; it’s creating one seamless experience. Customers shouldn’t have to think, “Which agent do I need today?” The goal is to build a unified experience across all AI agents and channels, making interactions simpler and more satisfying.

Our journey with Agentforce has shown us that AI’s true potential lies in making customer experiences more fluid, intuitive, and meaningful. By combining AI’s speed and precision with human support, we’ve been able to transform how we connect with our customers‌ — ‌and we’re just getting started. As we continue to learn, adapt, and refine Agentforce, our commitment remains the same: to lead by example and pave the way for businesses everywhere to create smarter, more impactful interactions.

We invite you to join us on this transformative journey. Let’s shape the future of customer success together‌ — ‌where every interaction is fast, easy, and designed to meet real needs.