Sponsored by: Salesforce | How Salesforce Gets More Done Daily with Slack and Slackbot

Summary

Salesforce uses Slack as its internal operating system for all work, deploying Agentforce-powered Slack agents that automate routine requests and address the 5 weeks per year knowledge workers spend on copying and pasting between systems.
TechForce, Salesforce's IT helpdesk automation built on Slack agents, achieved 70% ticket deflection and reduced resolution time from 27 days to 24 hours by handling routine requests automatically without human intervention.
Slackbot is deployed as a company-wide AI agent with zero change management overhead because employees already use Slack daily, and the long-term vision described in this video is a universal agent orchestration layer that routes any question to the right agent automatically.

Chapters

FAQs

How much time do knowledge workers lose to manual copy-paste work?

According to this video, knowledge workers lose approximately 5 weeks per year copying and pasting between systems, navigating tools, and searching for information. Salesforce uses this as the foundational justification for automating workflows through Slack agents rather than accepting manual coordination as the standard way of working.

What is TechForce and what results has it achieved at Salesforce?

TechForce is Salesforce's IT helpdesk automation built using Agentforce Slack agents to handle routine support requests automatically. It achieved 70% ticket deflection — meaning most routine IT requests are resolved without human involvement — and reduced resolution time for remaining requests from 27 days down to 24 hours.

How does Slackbot work as a company-wide AI agent?

Slackbot is deployed as a universal AI agent accessible to all Salesforce employees directly within Slack, handling routine questions and requests across IT, sales, and other business functions. Its key advantage is zero change management: because employees already use Slack daily, adopting Slackbot requires no new tools, training sessions, or workflow changes.

What is Salesforce's philosophy for using Slack as an AI platform?

Salesforce's strategy positions Slack as the operating system for all work, built around three principles: moving from copying and pasting across apps to conversation-driven interaction, replacing manual workflows with automated agents, and breaking down data silos with real-time intelligence. The future vision described in this video is a universal agent orchestration layer that automatically routes any employee question to the right specialized agent.

Full transcript

[00:08] Hi, my name is Camilla Gaidrov. I am one of the product marketing team at Slack. And then we're going to get slides up on the screen. It's going to be great. Normally, I do this session with my fellow co-worker from our business technology team or IT team. He
[00:22] unfortunately has the flu today, so I apologize. I will have to try and I don't know figure out how to do a two-person banter situation with just one human. Maybe a you know a little bit of role-playing. We'll see what how that goes. Um but today we're going to talk
[00:34] about how Salesforce gets more done with Slack and Slackbot daily. Um All right. You guys have seen the slide before. I'm not going to go through the slide. You know, promises that I make up here probably don't mean anything. So
[00:47] don't keep me don't hold me to them. But I love starting with this stat. Um on a daily basis, how many of you are in Slack in email in Salesforce all day?
[00:59] Really? For the people not raising your hands, I want to know what you're doing. Because what I find fascinating about this stat is that for all of us knowledge workers, if you think about all the time that we spend copying and pasting between systems, navigating,
[01:11] trying to find things, etc., that actually ends up adding up to a 5 weeks a year. So if my company wants to pay me for 5 weeks of copying and pasting, I would rather be on a beach with a mai tai.
[01:23] Alas, I am still chained to my desk and have to do work. But if that is really the cost of doing your daily lives, there's got to be a better way to tackle that. So when we were thinking about all of
[01:36] this at at Salesforce, um we kind of thought about agent AI is coming. Uh we've started experimenting with agents. We've started thinking about what are the new processes that we're going to be building, how are we going to be
[01:48] improving how we all work together. And we really came the philosophy came down to three things. From apps and tabs, aka from copying and pasting, to conversation. Instead of trying to just enter in data, we wanted to make sure that that data
[02:01] was part of the conversation. The conversation was the data itself. Instead of manual workflows, let's automate everything. Let's give it to the agents. Let's take those onerous manual tasks off of our plates. And instead of having silos of data, which
[02:14] haha, hello, we're at a data conference. Instead of having silos of data and having that live in a lot of different places, let's make sure that we can drive real-time intelligence by connecting all of the pieces. So,
[02:26] Andy White, who is one of the heads of our business technology organizations back in 2020, 2019, said, "Look, I think we should get Slack. I think we should use this internally. It's going to be a great product for our team."
[02:39] So, Salesforce went, "We're going to do one better than signing an enterprise license agreement. We're going to buy Slack." And we did that. And one fine day, BT, our IT depart- our IT organization, they turned off Google
[02:51] Chat, and they turned on Slack, and none of us knew what to do with it. But, the funny thing was really quickly, we figured out that it was a completely new way of working, and it was going to drive change in how we all work together
[03:03] as individuals, as teams, as organizations, and as a company. So, that's how we kind of think about it, right? Slack impacts Salesforce across the board from each individual employee all the way up to the company culture.
[03:16] Because what we realized is that it was only possible with Slack to achieve some of the things that we wanted to do. And Slack was going to set us up for a long-term improvement in how our
[03:28] processes worked internally. Why? Because first, it's designed as a conversational platform. It's not designed with conversation bolted on to records, or conversation kind of as like
[03:40] a weird little bubble thread in a comment. It's designed conversation first. The work is the conversation. Two, it's a perfect place. If your all your people are there already, and now all your data is there. That's where all
[03:53] your agents should be, too, right? It's a platform that is designed to incorporate that digital labor easily. And it's designed, like I said, to expand past the the the conversations and the and the agents into the rest of
[04:06] your tech stack. What else can you bring into Slack to make sure that it is the front end to everything that you're doing? So that at the end of the day, instead of your, you know, your your employees logging into 10 different systems and spending 5 weeks 5 weeks a
[04:20] year copying and pasting information, they're actually doing everything in one place, so all of that information is in one place. Um look, I know that the numbers look a little hinky because everything is at
[04:32] 30%, but it's interesting that every single line of business that uses Slack uses it in a slightly different way, right? Us marketers, we plan events, we bring our vendors and our and our contractors into
[04:45] channels, we we work with them, and we can put on an event faster. We can uh you know, increase employee productivity, build content faster, get approvals faster. Um in our sales organization,
[04:58] their win rates have gone up. Why? Because we're taking away some of that those onerous workflows of this is how you get legal approval, this is how you get finance to get involved. Um speaking of finance, they get faster decision-making because M&A is in Slack
[05:12] and quickly churning through all of their decisions. Across the board, every single business unit, every single organization has seen marked improvements by using Slack. And then, of course, we layered Agent Force on top of that. We talked about Slack being the
[05:25] natural home for digital labor. We leaned into that. We built Agent Force and we wanted to see how it could work for a company. And then, we went a little um crazy with the agents because we thought, "Let's build an agent for
[05:38] everything." Turns out you don't always need an agent for everything because this is not even the entire Rolodex of agents that we have built. We thought, all right, across every organization, every um type
[05:52] of of employee, we're going to give them dedicated agents. And what people really wanted was I just want one place to go and ask a question, and I don't care what agent is tackling that. I just want to do the
[06:04] thing and make it easy, not click into 17 other agents. Um so, we're working on that. We'll talk about that in a sec. So, let's talk about company productivity. Um like I said, we have a lot of agents. This is just a selection
[06:17] of some of my favorites, but the where I want to particularly um Does this do a little laser pointery thing? Am I going to blind anyone? I hope not. Uh where I want to focus on is
[06:29] TechForce. So, TechForce is our internal help IT help desk. Um I'm sure all of all of our IT teams across all of our companies hate us as employees, right? Why? Because they want to build
[06:42] interesting, fantastic, strategic projects, and we're like, "Hi, my laptop is broken for the 40th time." And if they have to ask us one more time, "Have we turned it off and turn it back on again?" we will be drop-kicked out of a window. Um TechForce
[06:56] actually solved a lot of those challenges because it started as a channel, just a regular channel, help TechForce. And instead of filing a ticket, we could go to help TechForce, and there were, you know, FAQs, and they
[07:08] said, "All right, have you read this first? Can you attest to the fact that you've read it first?" Oh my goodness, all of a sudden, ticket numbers ticket numbers were dropping. And then they expanded that, and they said, "All right, let's do a workflow.
[07:20] Here are the next steps." And all of a sudden, ticket numbers dropped. As we've improved TechForce, our resolution time, A, 70% of tickets have now officially gotten deflected either by bots or these
[07:32] days the TechForce agent. And B, of the 30% of tickets that are left, uh the resolution time has dropped down from 27 days to No, 24 27 hours to 22
[07:45] minutes. And the average time to response is less than 30 minutes because we're meeting people where they are. And that's really the whole philosophy of Slack, right? Give our teams the tools, give our teams the functionality that
[07:58] they need to do their work and do that in one place so that you don't have to go hunting for it. So, let's talk about the business units. All right. Um how many of you guys are coders?
[08:10] Great. I am a sociology major and I have never coded anything in my life, which is why things like this are perfect for me. Why? Because with AI and automation in Slack, actually 80% of all of the
[08:23] workflows that we build internally are built by marketers and French majors like myself. Um we don't have to code anything, but we know, "Hey, our process can get better, and I don't want to
[08:35] bother IT to do that." The tools in Slack allow me to take care of my own processes and find improvements. And this just keeps getting better, right? Slack bots out there now, Slack bot skills are out there. I can bring in
[08:48] Claude code if I want into Slack. All of that in one place again. Hey, business users can Well, we still don't remember how to turn our laptops on and off, but you know, other than that, we can actually
[09:00] take some of that work off of IT's team. And instead, they can focus on building things like approvals in Slack. This is single-handedly my my usual partner in crime likes to talk about this as the
[09:12] the thing that he is most beloved for in his beat in his IT career. Why? So, who here's a frontline manager? You have a team, you have Great. Do you like approving expenses?
[09:24] No. Does anyone like approving expenses? Never. Because you have to go to a different system and find the information and and then expect prove it. And it's onerous. You're like, "I got to find time in my day to do
[09:36] something else somewhere else. Instead, what they did is they used all of the APIs from all of these systems that we use that generate an approval process, and they funneled it all into Slack. So, as a manager, I can approve timecards, I
[09:51] can approve expenses, I can um I can approve pretty much anything in one single place. So, approval times have gone down from 2 days to 2 hours. Our credit card fees for late fees have
[10:05] it's been slashed by millions of dollars of spend. It's been pretty great. Um and then there's individual productivity, right? Like it's one thing to transform your company, but if you're just doing tops down and not bottoms up, you're not really going to get a lot of
[10:18] uh the behavioral changes that you want to see. So, when we acquired Slack, they had built this product called Midas for their own sales organization. What Midas did is it
[10:30] pulled from Slack metadata and the the you know, the database of customer information, and you would say, "Hey, I'm going to go talk to Bob on Tuesday." And Midas would say, "Great, here is the deck that you should use with Bob about
[10:42] how Bob is and his company are using Slack." So, we evolved that. We started with Midas, then we took it into the global sales insights app. Again, letting our sellers go to one place to prep for any
[10:55] kind of customer meeting and get a really deep insight, really deep view on their customer because they're not going to be doing the 13-plus hours of research for one meeting for a 30-minute meeting. But if you give them all of the information they need in 5 minutes,
[11:09] they're going to go ahead and read through that, and they're going to go to that meeting more prepared than ever before. And then that global sales insights app turned into our sales agent, the single point that every seller in Salesforce can go to to get information about how to do their job,
[11:22] how to talk to their customers, what to say, etc. It's It's It's an onboarding tool. It's a customer research tool. It's a here's the process tool. It covers everything a seller needs to
[11:34] know. And then there's Slackbot. How many of you guys are using Slackbot? So, the cool thing about Slackbot is um and granted I am biased because I work on Slack every day, but the cool thing
[11:47] about Slackbot is there's zero change management required for this. We, unlike when we turned Slack on for everyone and there was some learning curve, we turned Slackbot on and within 1 week we had 73%
[12:01] organic adoption of Slackbot. It just makes sense. Why? Because it is a personal AI agent where you're working.
[12:13] I can't find the deck from 3 weeks ago. Who remembers the movie Overboard with Goldie Hawn and Kurt Russell in the '80s? There was a boat. No, no one? Oh goodness. All right, guys, you got to go check this out. It's delightful. But
[12:26] there's a scene where the butler is like Rupert, I've lost my ruby earrings somewhere between 83rd and 87th 2 days ago. Go find it. That is Slackbot. Slackbot is that butler. Whatever you
[12:39] need in Slack, and not just in Slack, but from all the information that's being brought into Slack, so anything that you've integrated into Slack, Slackbot can search. Slackbot can put together a PTO doc for you for all of
[12:52] your coverage. Um it can tell you what your spirit animal is based on the context of your conversations that you're having with your teams and with your um with your customers. Basically, any
[13:04] conversation that you're having, it can pull from that context. That's the important part. Slackbot, unlike other you know, tools that you may have in a different tab, in a different window, in a different system, Slackbot's pulling
[13:17] from the context that it has in Slack. That means everything that you know, it knows. It knows more than you know, let's be honest. It knows more than we've forgotten in the day. Um it gets smarter with every interaction and we keep on
[13:30] driving more features, more functionality. So, that's Slackbot's becoming more and more like the front end to Slack and Slack is the front end to your entire tech stack. Um, like I said,
[13:42] 80% of us use it every day, highest adopted, most beloved adopted feature, and again, zero change management. This is the kind of functionality within Slack that made it a no-brainer for us
[13:54] as an organization at Salesforce to adopt it across the board, right? And to really lean in with making Slack the operating system for our for our company. Was it always smooth sailing? It was
[14:08] not. Is there ever smooth sailing with technology? There is not. But, there are things that we learned that we would like to impart to you today so that it's a do as we say, not as we do kind of situation. Number one,
[14:21] a lot of times it's really daunting. Hey, there's this new piece of technology and I don't know how to incorporate it into my tech stack or into my process, so I'm not going to. I could do anything with it and now I'm
[14:34] too scared to know where to start, so I'm not going to. And then you have shelfware. It's great, we all love that. But, if you start small, if you pick one use case, and quickly iterate on that, that's where you're going to see results. So,
[14:46] like I said, with Help Tech Force, we started with a channel, we ended with an agent. It was just about taking a use case, taking a team, taking their needs, and quickly figuring that out, quickly trying it out, and getting something out
[14:58] into the open. Iteration is the name of the game, test, tune, repeat. If you're not making changes to the way that you're working in Slack, and if you're not making changes to your process overall, you're not using it to the best of its ability,
[15:10] the best of its capacity. And it all comes down to data. Is the data telling you that it's a good idea or a bad idea? If it's a bad idea, why are we still doing it? If the process doesn't work, let's change the process. You have to be able to track and your
[15:23] data and then also centralize it so you've got accuracy and adaptability. You also need to be able to measure that value. Like how many times have you bought some technology and then at the end of the
[15:36] day the CFO, the CTO is like, "Look, yeah, you've been trying this out for a year, but can you tell me whether or not it's done anything for the company?" And if you haven't been measuring it, no, you can't. You might love it, but if you can't prove that it's worthwhile, if you can't prove that it's making
[15:49] a material impact to the business, it's not worth keeping around. And finally, it's really important to embrace new ways of working. I keep looking at Rachel. I'm like, "Do you remember when we used to do X?" Right? Like when we had emails. I mean, these
[16:02] days, I don't know. My emails is basically just Google Gemini notes and calendar invites. I'm never in there. Our entire culture at Salesforce has changed.
[16:14] All of our processes have changed because we said, "Look, technology is going to improve things, but it's never going to be a magic pill." Right? You can't say, "Oh, okay. I'm doing things this way and now I've got this new piece
[16:26] of technology and I'm just going to map that process exactly into it and hope for a better outcome." No, that's the definition of insanity. The point is if you're in adopting new technology, it's time to
[16:38] embrace that new way of working and really say, "Does this still serve my team, my organization, my company?" If it doesn't, let's figure out the best way. This is a clean slate. This is our opportunity to really understand how we
[16:52] can improve as as an organization. And that's why I love talking about this because we ourselves, we we we drink this Kool-Aid. We drink this champagne, right?
[17:04] We consider ourselves customer zero. So, anything that we're rolling out, we're going to be testing on ourselves first and we're going to be telling you about Did it work? Did it not work? How well did it work? And frankly, you're not going to see it until it works for us
[17:17] because if we can't figure out how to make something work and make something impactful, then it's not worthwhile for our customers to see that either. All right. Uh we're just getting started. That's what the titles of the slide says. But we are. We are just
[17:30] getting started. Again, that iterative process makes it really interesting and really easy for us to dig in and say, "What's the next thing that we can do better?" We started with all of the agents. All all of the
[17:42] agents. Too many agents. We started with that. We started with builder enablement. We started simple. Um now we are looking at high priority personas. Who are the heaviest users of Slack? Who are the heaviest users of certain pieces
[17:55] of our tech stack? Let's try and prioritize and tailor the content and the and the processes to them. Um let's make sure that everyone feels comfortable on building their own agents, building their own skills. Let's
[18:08] make it easy again for IT to focus on strategic projects and for the business users to do what business users need to do. And then in the future, how do we bring everything back together?
[18:20] Like I said, we had a very large Rolodex of agents. We don't need that. Let's get rid of some of them or let's bundle them together. So, for example, we had like an HR agent and an onboarding agent and a this agent and a
[18:32] that agent. Now we just have an employee agent. One single employee agent. You don't need to know, as an employee, where to go. It's pulling from our back-end HR and finance and IT systems. It will do the heavy lifting. You just
[18:45] need to ask it the question. So, expanding on topics, expanding that capability, making sure that the agents can orchestrate across more, making sure that Slackbot can orchestrate across all of those agents. So, in fact, you don't
[18:57] even need to know I mean, I am a manager, which means I can use both the manager and the employee agent. I don't want to know which one is being used. I just want to ask a Slackbot a question, and it will orchestrate across all of
[19:10] our agents to figure out where to route the question, how to tackle the workflow itself. And we're just going to continue retooling, um, tuning, refining, etc. and making sure
[19:22] that all of our employees feel comfortable making Slack their own because that at the end of the day, that ability to tailor, that ability to expand the platform, that ability to really, um, make the experience unique to your
[19:35] needs and your line of business, that's what's made Slack impactful to us as an organization. Um, and with that, please complete your surveys. They're in your apps. They're easy to get to. Thank you so much for joining me today.
[19:47] Have a great rest of the time.

Learn more about the Databricks Data and AI platform.

The information provided herein is for general informational purposes only and may not reflect the most current product capabilities or configurations.