Chances are, your first internal AI agent was something of a jack-of-all-trades. You could ask it questions about your company’s PTO policy, get help provisioning a new laptop, or even task it with putting together a competitive analysis ahead of an important sales call. It did an admirable job at most of these tasks, most of the time. But like any worker stretched too thin, it could sometimes drop the ball.
We’ve all experienced that disappointing moment when an agent fails to execute a task it ought to be capable of, or when it suddenly seems to forget the context of what it’s working on. These shortfalls make it difficult to fully embrace AI agents and pivot away from legacy processes.
At the same time, we’re also learning more about where and how to integrate AI within the flow of work. Many early agents took the form of standalone apps or were baked into software in ways that limited their utility. The good news? It’s time to turn the page on your first AI copilot and say hello to agents tailored for employee use cases (Seriously, say hi – they’re on Slack).
Jack-of-all-trades, master of none
As organizations continue to advance along the agentic maturity curve, it’s becoming clear that specialized, purpose-built agents can deliver better, more reliable results than monolithic copilot deployments.
In part, it’s a simple matter of latency: The more actions an agent has to choose from, the more time it’ll spend evaluating which one is best for the task at hand. By narrowing that decision space, agents can work more efficiently and reliably because they’re not distracted by irrelevant “semantic noise.”
When we launched Agentforce, customers were limited to a single copilot instance across their entire org. In practice, this meant that while the copilot could understand user-level context, it often got drowned out by semantic noise as the agent reasoned across the many actions in its arsenal. We heard your feedback loud and clear on this – that’s why we now offer a specialized agent template for employee use cases, architected from the ground up to support multiple agents that can each be tuned to a specific use case.
We also heard from many of you that agents are most helpful when they’re present in the spaces where employees do most of their work. For this reason, agents built with the employee agent template can more easily be integrated into Slack, where it can take advantage of some of the most valuable enterprise data around – your own Slack conversations.
By deploying purpose-built AI agents with user-level context and native integration on channels like Slack, organizations can more reliably automate internal processes for things like onboarding, help desk support and common CRM tasks.
Let’s take a look at how Agentforce makes this all possible.
From one AI agent, to many
With the employee agent template, you can now deploy multiple tailored employee-facing agents for specific use cases. For example, an HR team might build a specialized agent to automate leave requests and expense reports, while a customer success team can tune an agent to answer nuanced product questions and prepare executive briefs. By calibrating roles and guardrails around a tightly-defined scope of work, Agentforce for employees use cases can generate much more accurate and personalized responses.
With a rapidly expanding Agentforce, security becomes more important than ever. Role-based access controls (RBAC) automatically carry over permissions and tailor agent outputs to the objects and fields the employee has access to. We’re making AI agents more available to everyone, and every interaction more personal and dynamic.
In that same spirit of democratizing access to AI agents, we’re eliminating the need to go through CRM to find Agentforce. Instead, you can simply head to Slack and chat with your new agents directly in the Agentforce tab. This not only brings Agentforce natively to your favorite collaboration space, but also infuses all the rich context from your DMs and channels into every agent interaction, opening up exciting new possibilities. You can ask Agentforce to create an opportunity record based on a DM with a colleague, or kickstart an expense report based on the travel itinerary your manager just added to the team Canvas.
To learn more, check out the demo video below and visit agentforce.com.
