Agentforce in Slack: Powering Engineering Excellence at Salesforce

The idea that ‘engineers don't scale’ is false—what doesn't scale is the constant hunt for context. As teams grow, so does the overhead of tracking down historical knowledge, jumping between tools, and answering the same questions over and over again.

At Salesforce, that's exactly what we set out to solve with the Engineering Agent in Slack—deployed right in our channels, where our engineers already collaborate. Built on Agentforce and integrated with Data Cloud, MuleSoft, and Heroku, the Engineering Agent connects directly to the tools we use every day. Whether it's answering technical questions, automating testing tools, or helping onboard various services to engineering systems, it delivers instant, reliable support—freeing up engineers to focus on what really matters.

The results we've seen at Salesforce have been astounding: fewer distractions, faster unblocking, and more time spent on the work that actually delivers impact. In just six short months the Engineering Agent has handled 18,000 support interactions!

Delivering support when and where engineers need it most

As engineers grow in their roles, they often take on ‘on-call’ duties—fielding quick questions, troubleshooting issues, or offering guidance. At Salesforce, the team estimated that this work required 10 minutes per request, eating up hours of senior engineers’ valuable time.

Now with Agentforce in Slack, the Engineering Agent serves as the first point of contact for these enquiries. Engineers in need of information or assistance can interact with the agent in channels or DMs and receive immediate, precise answers and targeted guidance 24/7.

Our team customised our Engineering Agent to act as an ‘agent of agents’ so that when a team member asks a question, the agent will intelligently route the question to the correct specialised agent for the topic at hand. Responses are returned to the requester in Slack in a conversational manner, and include links to relevant resources and citations for the agent's response. This approach makes it easy and intuitive for every member of the engineering team to access highly specific domain knowledge without detracting from other team members’ focus or productivity.

To deliver accurate, relevant responses, the Engineering Agent continuously ingests structured and unstructured data from sources like Slack, Google Docs, Confluence, GitHub, and more. The data sources are refreshed daily, ensuring agents stay current and precise in each response.

Giving engineers their time back

Beyond answering questions, the Engineering Agent takes action and streamlines our teams’ workflows by intelligently orchestrating routine tasks through MuleSoft integrations. It handles everyday processes like creating org requests, managing lifecycle of workspaces—freeing engineers from repetitive tasks and reducing friction across our systems.

The agent also proactively alerts teams to critical system events—like build failures or merge conflicts—helping engineers be aware and respond quickly to minimise disruptions.

Empowering engineers to focus on high-value work

The Engineering Agent has become a vital asset for our Engineering 360 team and wider engineering organisation here at Salesforce.

In just six months, the Engineering Agent has handled over 18,000 conversations across 3,500 users. On average, it answers 130 questions per channel per month, saving ~30 hours per channel. Projecting forward based on March's volume, it could deflect over 50,000 conversations annually, saving 17,000 hours, equivalent to 8+ FTEs and $1.4M.

As we continue to build out our own internal Agentforce, and scale the number of data sources from 18 to 30–40 specialised agents across over 700 Slack channels, we project over 275,000 hours saved annually. This is equivalent to freeing up the capacity of 130+ engineers. As adoption grows across our engineering teams, we project this initiative will save us $20M per year.

More than just efficiency, the Engineering Agent puts knowledge at engineers’ fingertips and frees them to focus on high-impact work. For us, Agentforce is not just a tool—it's an always-on teammate.

With Slack, the Engineering Agent is a powerful digital teammate

As an engineering leader, choosing Agentforce and deploying within Slack—where engineering teams already collaborate—was a strategic decision. It allowed us to customise the agent to our exact needs and maintain control over our data. The Engineering Agent perfectly showcases how integrating AI into Slack can revolutionise support, optimise workflows, and create a more collaborative environment. For any enterprise IT organisation looking to boost engineering productivity and modernise their support, deploying agents within Slack is the clear path forward.

Talk with an expert to learn more about using Agentforce in Slack.