Agentforce resolves IT and HR questions 24/7 for Salesforce employees.

IT and HR support delays distracted Salesforce employees from critical deliverables and service.

With customer success at the heart of its mission, Salesforce wants all 76,000 employees to stay focused on building software, running programs, and driving business transformations that help customers succeed with Salesforce technology. But unresolved IT issues or time-sensitive HR questions can easily derail that focus — leading to lost productivity, missed deadlines, and disruptions in service.

Techforce, Salesforce’s internal IT team, plays a key role in keeping employees up and running. It manages 25,000 support tickets per month for everything from resetting access permissions to replacing lost laptops. While the previous Einstein chatbot helped sort incoming tickets and guided employees through basic troubleshooting, people still faced average wait times of 27–30 hours for more complex support. With limited bot capabilities and high ticket volume, Techforce needed a better way to keep up.

Device issues aren’t the only time employees need support. When life events arise — like preparing for parental leave or relocating to a new state — employees often have benefits questions, too. Many turn to BaseCamp, Salesforce’s internal employee portal built on Service Cloud’s HR Service, to search for answers. While 96% of questions were resolved through self-service articles in the past, the remaining inquiries still added up to thousands of HR support tickets each year. And when employees opened tickets, HR reps had to search across systems like Workday and BaseCamp to determine eligibility — a process that took, on average, a day and a half.

Waiting one to two days for support — whether technical or personal — meant employees couldn’t move forward with confidence. And when employees are stuck, they can’t do their best work or deliver the kinds of experiences Salesforce customers count on.

Humans with agents drive employee success, together.

Agentforce, the agentic layer of the Salesforce Platform, is helping employees stay productive and feel supported — both day-to-day and throughout their careers. By deploying digital labor across IT and HR, Salesforce can scale intelligent service around the clock. And because these AI agents work alongside employees right in Slack, they can take action on just about any topic where the right data and knowledge is available. Here are just a few examples of Agentforce, at work across the company: 

Techforce will help employees get support quickly by autonomously resolving 75% of inquiries with Agentforce.

To help employees get the tech help they need anytime, anywhere, Agentforce, embedded in Slack, answers questions and provides on-demand troubleshooting help directly within the platform employees already use to work and collaborate with. Powered by over 6,000 detailed knowledge articles in BaseCamp, Agentforce now serves as a first line of IT support. 

For example, if an employee is locked out of their email, they can start a request either in BaseCamp or directly in Slack via the #help-techforce channel’s workflow. Either way, the request is automatically routed to Agentforce through an omni-channel flow, which creates a case in Supportforce, Salesforce’s internal instance of Service Cloud, and initiates a chat with the employee. The employee might say something as simple as, “I can’t connect to Org 62 (Salesforce instance),” and, thanks to natural language processing, Agentforce understands this request — even if it has typos or imprecise technical terms. It immediately searches the Techforce knowledge base, identifies relevant troubleshooting steps, and guides the employee through them in real time. Once the issue is resolved, the employee is back to work without missing a beat.

“What we found with bots before is they’re very dependent on keywords or very specific terms to trigger items,” said Neil Cummins, Director of Desktop Support for Techforce. “What we’re seeing with Agentforce is the reasoning across articles so it doesn’t just send back a wall of text, but really walks users through it.”

So far, Agentforce has autonomously resolved more than 9,500 support requests in Slack, saving Salesforce nearly $57,000 in support costs. By the end of 2025, Techforce expects that number to grow significantly — with 75% of cases resolved autonomously and projected savings of $1.4 million. 

From reimbursement to leave, Agentforce simplifies HR support and saves employees time.

Salesforce is also deploying Agentforce for HR use cases to help employees get the answers, resources, and insights they need within Slack. 

“By adding Agentforce to Slack, we’re removing friction from this process,” said Zach Gray, Salesforce’s Senior Director of Data Intelligence and Design. “It’s fast, it’s convenient, and it’s freeing people up to focus on strategic, high-impact work.”

The HR team started by using Agentforce to answer questions about a common topic: wellbeing reimbursement for physical and mental wellbeing expenses. Employees regularly submit thousands of support tickets to HR to ask if specific items are eligible for reimbursement — so many that wellbeing is in the top five of case topics received by HR. 

Now, employees can access the Wellbeing Agent via Slack or from the BaseCamp knowledge article about the wellbeing reimbursement program. For example, an employee can ask, “Are cooking classes eligible for reimbursement?” and immediately get an answer letting them know that yes, cooking classes are eligible for reimbursement and how to file a claim. 

Within the first seven business days, the Wellbeing Agent successfully completed over 1,000 conversations, reducing response time from days to seconds. 

In the future, employees will be able to use Agentforce in Slack to submit reimbursement claims on their behalf. They can also ask broader questions about other company policies and get answers grounded in knowledge articles on BaseCamp, saving them time and providing much-needed clarity for complex topics.

HR Service augments specialists with Agentforce-powered insights and summaries to respond to and resolve cases faster.

Wellbeing reimbursement is just the beginning. Because Agentforce is deeply integrated with Service Cloud’s HR Service — the platform HR operations uses to manage support cases — it can automatically summarize relevant case history and identify trends across incoming questions. This is especially valuable during times of policy change, when inquiries spike. These insights help the team recommend next steps based on patterns from similar cases. For example, if an employee asks about extending parental leave, an HR specialist can use Agentforce to quickly summarize current eligibility criteria, understand common follow-up questions, and anticipate nuances based on past cases. From there, the specialist can deliver a comprehensive, conversational response that clearly explains eligibility and next steps. By streamlining routine responses, Agentforce speeds up resolution times and frees HR specialists to focus on more complex, one-on-one support.

Soon, Agentforce will also be able to take action on behalf of employees to submit time off, start the leave process, and more. The HR operations team also hopes to use Agentforce to suggest updated content for knowledge articles based on conversations with employees, which will ensure that BaseCamp content always reflects the most helpful guidance. 

Career Agent helps employees grow — and could boost internal mobility by 30%.

Agentforce is also helping Salesforce employees find growth opportunities — and even their next role at the company — by tapping into Career Connect, the company’s custom career development hub.

At the heart of this effort is the Career Agent, an autonomous assistant that helps employees navigate their careers by surfacing personalized growth and learning opportunities. By tapping into Data Cloud, the Career Agent pulls together data from across the employee ecosystem, including demographic and job posting data from external ERP system Workday, and course and certification progress from myTrailhead courses and Astro’s Course Tracker, an internal learning management platform. It also taps into each employee’s Career Connect profile, where they define their current skills, areas for development, and long-term goals through self-assessments and manager reviews.

With this rich data foundation, the Career Agent can suggest next steps in an employee’s development journey — whether that’s enrolling in a course, exploring a new career path, or applying for an internal role. Employees can ask open-ended questions like, “What are some next steps I could take in my career at Salesforce?” and receive dynamic, personalized recommendations. They can also mark preferred roles, flag areas they want to grow in, and receive updated suggestions as their interests evolve.

Career Agent operates within clear guardrails, ensuring suggestions align with employee goals, internal policies, and available opportunities. And because it’s embedded within Career Connect, employees get tailored career guidance directly within the platform they already use — with no need to navigate multiple systems.

Salesforce expects the Career Agent to drive a 30% boost in internal mobility and help retain top talent by making career development feel more intuitive, proactive, and personal.

Manager Agent brings a full-picture view of teams to help leaders better support and grow their people.

Managers juggle a range of responsibilities, from daily tasks to check-ins and training. To help them work more efficiently, HR introduced the Manager Agent in Slack, which surfaces key insights across their organization — no need to dig through the notes on each case and each employee’s profile to identify trends and opportunities for improvement.

Manager Agent is grounded in data from Employee360, a centralized platform that gives managers a holistic view of their team’s priorities, growth areas, and engagement. Powered by Data Cloud, Employee360 pulls in metrics from multiple sources, including performance check-ins and evaluation comments from BaseCamp, Agentblazer and certification status from myTrailhead, and current and desired skills from Career Connect profiles. It also integrates in-office attendance data from badge scans stored in Snowflake, and volunteer activity from Benevity, an external volunteer management platform. 

With this full view, Manager Agent can respond to Slack prompts like, “What are some of the top needs across my organization, and what resources could I provide to support them?” It might recommend relevant courses based on performance feedback, or suggest a volunteer event to strengthen team connection and encourage in-person collaboration.

For example, Agentforce might reveal that several team members want to improve their public speaking skills and recommend a Trailhead course on giving compelling virtual presentations. A manager could then share that course with the team or ask Agentforce for alternatives that better align with other priorities. These personalized suggestions for learning, collaboration, and skill development save managers time and ensure each team’s needs are addressed holistically — freeing managers to focus more on meaningful 1-to-1 conversations and the topics that matter most.

“We believe that Agentforce is not just a technology transformation — it’s a people transformation where we work better alongside AI agents and develop the skills of the future,” said Gray.

This mindset of humans plus agents driving success together has shaped every Agentforce rollout across Salesforce’s internal teams. But while Agentforce is already 75% more accurate than DIY models, just like a new employee, it doesn’t always get things right at first. Each AI agent needs feedback, guidance, and ongoing monitoring to improve over time – and the low-code Salesforce Platform makes this scalable for every team across the company.

With Agent Builder, Salesforce teams can easily test and refine agents using low-code tools. These include out-of-the-box topics, natural language instructions, built-in guardrails, and prompted actions that can securely pull in data from across the company and in adherence with Salesforce’s data governance. Adjusting instructions, connecting missing data, and optimizing guardrails have all helped Agentforce improve week by week across use cases.

Take the Manager Agent, for example. The initial version of this agent was designed to answer questions about compensation planning by reasoning across Salesforce’s compensation and rewards content. Given the sensitivity of the topic, it was strictly limited to answering only from this curated dataset. Guardrails prevented the Manager Agent from making specific compensation recommendations, providing financial or benefits advice, or responding to questions about individuals.

However, the high-level nature of the Manager Agent’s grounded content meant it struggled to address nuanced questions around compensation philosophy, eligibility, and related topics. The team quickly recognized the need for the agent to access more granular and refined information to be truly helpful. They tested hundreds of user queries, continuously evaluating and refining the agent’s responses and the data set. Sometimes, the agent even flagged instances of conflicting information in the source content. Thanks to Data Cloud’s real-time unifying and harmonizing of the different data sources, retesting responses with updated content was a fast and efficient process.

The Manager Agent became a true extension of the HR team, autonomously handling approximately 3,000 manager inquiries during the compensation planning period with 74% accuracy and cutting HR ticket volume by 50% compared to the previous year.

During the Manager Agent’s rollout, HR worked closely with the implementation team to monitor performance by reviewing responses, fine-tuning the conversational design, and analyzing satisfaction ratings. They also refined the agent’s guardrails to ensure it only answered questions based on grounded, up-to-date data. If an employee asked something that required human support, the agent could escalate it right away. For example, it can explain how to enroll in the Employee Stock Purchase Plan — but it won’t offer benefits advice.

Looking ahead, the Manager Agent will expand into new areas such as learning trends, volunteer time, feedback, and job priorities. “HR is leading the charge as an Agentforce-first organization,” said Sadie Koeppel, a Content Manager at Salesforce. “We’re continuously seeking feedback and exploring how to build these tools to help employees do their jobs better.”