Salesforce Unveils AgentForce: A Bold Move in the AI Revolution

Dreamforce 24 - Salesforce COO Brian Millham on getting Agentforce into the hands of the customer

The impression coming out of Dreamforce this year is that Salesforce has assembled a powerful team of AI engineering talent who have delivered what CEO Marc Benioff hailed as amazing technology in the form of the Agentforce platform. That's all to the good. But as Benioff went on to point out: We believe that our computer scientists have delivered something that's extraordinary, but it doesn't matter until customers use it and get value from it. It just doesn't matter. The challenge now then is getting customers to test and experiment with what's being pitched as a new approach to AI at a time when others - hi Microsoft, hi Oracle, hi OpenAI! - are promoting a different vision. At Dreamforce itself, that's taken the form of the Launch Pad space, where delegates can set up their first agents. It's a model that will be replicated around the world with a series of events in 28 locations. But in order to 'land the message' about Agentforce's different approach, there will need to be more to the go-to-market/sales strategy as not everyone is going to make it to one of those events. That's the scenario I put to Brian Millham, Salesforce's Chief Operating Officer. He argues that among the C-suite, the AI hype cycle is moving into a new phase: Several of the CEOs [I met yesterday] said it feels like the hype cycle is now becoming the value cycle...we're now in a value world. I believe that if you're going to have staying power with your customers and going to drive success for your customers long term, you better be delivering value and efficiencies and productivity to your customers. That's what we think we're driving with Agentforce. This is AI in the flow of work. This is going to make your sales team better and more productive. This will make your service organization more productive and more efficient. It'll make your marketing campaigns more personalized and more targeted. We have these [customer] relationships today, they exist. We know what our customers want. We think we're highly-differentiated. Yes, it's competitive - everybody's talking about AI - but based on the feedback from customers and the value they've derived in a very short amount of time, we feel very good about differentiation. It's a case of keep on keeping on. Benioff made the point several times yesterday that his keynote slide deck would be very familiar to long-term Salesforce watchers. The issues of 'DIY AI' that he has railed against in recent weeks are essentially a problem that has haunted tech stacks for the past 25 years or more. So Millham, who's been at Salesforce for all of those 25 years, says: We're going to continue to rely on the things that have driven our success over time...When you anchor your business from day one on trust, it is a relationship with customers that is unique and different. We want to make sure that we continue to deliver on that, that everything we build is built with that first and foremost. We built in a Trust Layer before anything else. That was the first thing that we did to make sure that we were delivering on that. He adds: We've been anchored in trust from the very beginning. We do feel like customers see it as something that is important to them, and they want to rely on us as a partner with them in the trust world. That's the relationship we want to have with our customers, that's a differentiator for us. That's incredible, and that's what we want to have. I'm not sure that with some of the smaller players, even in the Microsoft world, with copilots, the results they're getting are not only inaccurate, but are toxic. There's hallucinations and other things that are happening there. So we want to get oriented to delivering value for our customers, and that's what we're going to get doing in a very trusted way. But to return to the question of how Salesforce counters the model-centric messaging of rivals, what's the strategy here? Millham says: Many [organizations] will build models, and will continue to build models out there. Our focus is going to be on what value our customers are going to get to in the next couple weeks, the next couple months. We think speed of value is an advantage for us. You see it at this conference with everybody building agents, our ability to get hands on keyboards and then deliver real value for our customers. This is where I think the notion and the promise of what AI will deliver for our customers is really hitting the wall. With copilots, they're just not seeing the benefits. With the work we've done with early customers, the most exciting part of the story of OpenTable and Fossil is that they're already seeing value. They've already seen adoption of this technology as an accelerator on the business. For 25 years, this is what we've focused on - first trust and the customer success, which ultimately means adoption value derived from the parts that we've delivered for our customers. That's what we're going to focus on. And the sales teams are ready to meet the challenge, he argues: My Dreamforce started on Sunday, just to make a long week longer. Starting on Sunday, we brought in more than 2000 of our sales engineers to really get ramped up. Our sales teams are going to go. We've actually activated them for all of these [post-Dreamforce global] tours as well. So they're being enabled. It's a challenge, certainly for any company of our size to get people on message and understanding exactly the capabilities and technology underpinning Agentforce. But that's my job, and that's what we have to go do with the sales teams. So, the AI 'wars' open up on a new front. It's hard to argue with Millham's rallying cri du couer here: At some point you have to deliver for your customers. And if it's six months, 12 months, two years, whatever that happens to be, if you're not delivering for your customers and Salesforce can show up and do it in a timeframe that is much shorter and delivers clear value in the flow of work, that's what we want - to go short. So, sure Microsoft is talking a lot about copilots right now. We believe agents are the future, and that's what's going to deliver value for our customers. The game's afoot. Onwards!

Salesforce Dreamforce 2024: CEO Benioff Attacks Microsoft Copilot Models, Touts Agents As Better

'Many customers are reporting the OpenAI models are not delivering high levels of accuracy,' Salesforce CEO Marc Benioff said. Salesforce CEO and co-founder Marc Benioff railed against copilot artificial intelligence products offered by Microsoft and other vendors, touting his company's agentic approach as one that delivers better accuracy while maintaining the importance of the vendor's services partners. In response to a question from CRN during a press and analyst event, the CEO of the San Francisco-based enterprise software said of Salesforce's 12,000-strong partner ecosystem in delivering AI that "it's critical that we enable and train them and give them this extension." "A lot of them are already next generation Salesforce platform users," Benioff said. "These are the people using the product. And it (new AI features) just appear inside the platform. It's not some new thing that they're going to buy or add on or plug in or whatever. It's going to appear from within." [RELATED: Salesforce Dreamforce 2024: The Biggest News] Benioff's view is that providing Salesforce customers a full AI platform without adding on or plugging in AI products will result in better accuracy and lower hallucination. "We believe that this is the key to making it work," he said. "This is what AI is meant to be." Microsoft has been most visible with its copilot brand of AI virtual assistants, but other vendors have used the term as well, including Salesforce itself with Einstein Copilot. Hewlett Packard Enterprise, Spot AI and SymphonyAI. CRN has reached out to Microsoft and OpenAI for comment. Benioff pointed out that during a virtual Microsoft event Monday on the next wave of the tech giant's AI updates, Microsoft CEO Satya Nadella said AI was becoming "more capable and even agentic" and models were becoming "more of a commodity"--observations Benioff has also made. Benioff directly attacked Microsoft and ChatGPT maker OpenAI, which has received billions of dollars in investment from Microsoft and helped power some of Microsoft's AI offerings. OpenAI CEO Sam Altman notably spoke at Salesforce Dreamforce 2023. "Many customers are reporting the OpenAI models are not delivering high levels of accuracy and resolving even basic customer service issues for them," Benioff said. "The lack of grounding, the lack of access to the metadata, the data itself, the sharing model, all of the components of a platform that are then needed to be able to achieve this kind of level of accuracy." He also said that Salesforce's Agentforce AI offering "is outperforming OpenAI on Azure in cost, time to value and accuracy." Benioff attacked Microsoft's business practices, commenting on the tech giant's treatment of Slack, which Salesforce bought in 2021. "The European Union wrote this interesting statement about how they run their business, about Slack and some of the things that they did inside their company when these two entrepreneurs were trying to build this company and what actions they took and how they went after them," Benioff said. "You might want to read it. It's very interesting. A lot of insights. It reminded me also, on Netscape was another one. They had a very interesting document about that. ... It's a very interesting business philosophy there" at Microsoft. Benioff also expressed some concerns about the perils of AI. "I hope we're on the right side of history here," he said. "It's a very high-wire act. . ... I am sure that there's going to be good stories and bad stories. I'm sure that some of this is going to work, and some of it is going to go horribly wrong. I hope the horribly wrong stories are not as bad as I have them in my mind that they could be. Because they could be horrible. But they also could be magical. And I'm not sure what's going to happen." Here's more of what Benioff had to say during Dreamforce across his keynote and during a press event.

Dreamforce 24 - Marc Benioff on why he tore up his keynote in order to break the "hypnosis" of "science project" AI

With Dreamforce as the single most important sales opportunity of the year in the Salesforce calendar, an enormous amount of prep goes into the opening keynote session. That was the case this year as usual with focus groups, rehearsals, dry runs etc. Then something unprecedented happened, according to CEO Marc Benioff: About three weeks ago, I made the decision to throw the whole keynote away. Usually, by the time I give the keynote, I have already practiced it three dozen times, and have delivered it to 10,000 people all over the world. The keynote just got finished last night. I had never done it before. I didn't even practice it. I was not even in a dry run. Given that this year's gathering in San Francisco was pitched as the most important Dreamforce ever, with its introduction of the Agentforce platform front and center, that's a pretty radical move to make, which begs the question, why? According to Benioff, the decision stemmed from a growing awareness of customers problematic reaction to AI: I realized that customers have this unbelievable, deep frustration around AI and a huge level of excitement - and they hold both in their hands. They're very excited because they realize what the potential is...they get what AI is. It's not that hard - we watch the movies.But then they're like, 'Why don't we have more of this happening in our business right now? I made this investment, I bought this, I licensed this, I have this, and I'm not totally getting the result I want'. A lot of that is down to the messaging that's being put across by the likes of Microsoft and others, he suggests: [Vendors] will come in and say, 'You have to DIY this cloud. You have to DIY the mobile, the data, the social and now this part ,the AI and the agents. We say, 'No, no, you don't have to spend that money. This is a science project. They're selling you science projects. You need to move away from it and we can prove it to you'. And over and over and over again, with our best customers all over the world, we have shown them that our approach is better....It is crazy that people are feeling like, 'I have to go hire some somebody to do all this'. Are we still in this moment in our industry? We're in a DIY moment, and our job is to break the DIY, and the only way we're going to do that is to make all that happen. And so it was that Benioff gathered his team around him, told them the company was going to assemble 4,000 people to get Dreamforce attendees using agent technology, and show them a better way in order to break the "hypnosis that's coming out of not just one, but many vendors". While Microsoft has been in the crosshairs for many months now, it's interesting that Sam Altman's OpenAI also came under fire as Benioff observed: Many customers are reporting the OpenAI models are not delivering high levels of accuracy and resolving even basic customer service issues for them. (Altman was a top-billed guest at last year's Dreamforce. He's nowhere to be seen this week.) But models in general are not the right approach, argues Benioff: They're just commodities. They don't do anything for you. The value is in the data and the metadata...I was like,'What am I doing to help my customers realize that?'...I really think that the copilot is the next [Microsoft] Clippy - it hasn't really delivered for customers what they intended. It's cute, it's fun, does some things and then you're not really using it. It doesn't have the adoption rates [it had] a few weeks ago or months ago. With all that in mind, Benioff determined that the way to challenge the current mindset is to get agent tech into the hands of customers and let them experiment and play with it themselves. This will go one of two ways, he admits: Either, the customer is going to be crying, then it works, and it's going to be a very emotional or they unleash their agent and it completely screws up, and their whole job is gone. It's one of those two things. I don't know. I hope we're on the right side of history here. It's a very high wire act. We realize we're dealing with the most avant garde technology, the most exciting stuff that everybody wants to talk about and try out. But how many of us really have the ability to get our hands in the soil? Usually, we have to rely on what everybody's saying, right? How many times do you have an actual customer in hand that saya, 'Yeah, I resolved 80% of all my customer service issues, and I increased my revenue by 20% Oh, and employee satisfaction also went up by 30% because I am resolving more issues for them'. I have 10 of those stories because we've been in the generative AI world now for three years. But I think those stories are not quite as prolific as they are intended to be. Our job is, like we have for 25 years, to remove the veil and to say, 'Actually, this is the right approach, a completely different approach'. The target audience for this is clearly identified, he adds: These are people who are doing the day-to-day work inside companies. They are not AI experts, they are not necessarily cloud experts, they are not mobile experts, they are not social experts. This is never what they were supposed to be...What I'm excited is these are people who know nothing about agents, very little about AI, who are going to go back to their companies and say, 'Let me tell you about how an agent can change our business'. That's our goal. Our goal is to empower them so that when they come in on Monday morning and they'll say to their manager, 'I'm going to show you how we are going to become a lot more productive, how we are going to increase our revenues, augment our employees, increase our margins and deliver the next generation of AI'. It may be coming from a very unexpected source, but it's backed up with a level of computer science and technology that will be astonishing. And if they're going to benchmark this against other projects going on in the company, we have a high level of confidence that we are going to win because we are winning now in those in those benchmarks. The overall premise here is simple as Benioff states: We believe that our computer scientists have delivered something that's extraordinary, but it doesn't matter until customers use it and get value from it. It just doesn't matter. I'm sure that there's going to be good stories and bad stories. I'm sure that some of this is going to work and some of it is going to go horribly wrong. I hope the horribly wrong stories are not as bad as I have them in my mind, as they could be, because they could be horrible, but they also could be magical. I'm not sure what's going to happen. A typically bold roll of the dice from Benioff. It's interesting messaging, high risk in its own way. But the keynote that we saw on Tuesday morning was one running on conviction and righteous fury at the toxic hypnosis that the CEO sees as holding back the potential of this third wave of AI. Benioff has surrounded himself with some extraordinary AI engineering talent and those people have delivered for him in terms of the tech. Now it's a case of getting across the story of a better way ahead. The stated objective is to have "a billion consumers interacting with agents globally" by the time of Dreamforce 2025. Start the clock!

Highwire act: Salesforce's big bet on Agentforce spotlights volatile nature of the AI world - SiliconANGLE

Highwire act: Salesforce's big bet on Agentforce spotlights volatile nature of the AI world During his keynote remarks at Salesforce Inc.'s Dreamforce gathering in San Francisco on Tuesday, chairman and chief executive Marc Benioff (pictured) decided to put a local spin on the company's introduction of its artificial intelligence platform Agentforce. Making a comparison between the announcement of his company's ambitious enterprise AI technology and the Waymo LLC autonomous cars currently rolling on the city streets outside, Benioff called the deployment of Agentforce his company's "Waymo moment." The comparison captured both the goal of the firm's signature Dreamforce event, with a plan for attendees to build and deploy 1,200 agents during the week, and the risks associated with trusting AI to safely deliver results. Much as it takes a leap of faith to climb into a car with no human driver for a journey across the busy streets of a major U.S. city, Salesforce customers will now be launching autonomous agents to run key aspects of their business without a long track record of proven success behind it. "What you are going to see at this show is technology like you've never seen it before," said Benioff, during his keynote appearance. "You're going to have that moment when you're saying 'I'm going to deploy agents in my company and it's working'...hopefully." The "on-the-fly" nature of Salesforce's deployment of its Agentforce initiative this week was further underscored in a meeting with the media immediately following Benioff's keynote address. The company's CEO revealed that although his Dreamforce presentation was finished, audience-tested and ready to go three weeks ago, he decided to re-write the entire keynote based on customer reaction to Agentforce during his recent customer tour in late August and early September. It also led to a decision to fly-in 4,000 Salesforce engineers to San Francisco who would help customers use the AI platform onsite for building and deploying new agents, according to Benioff. "I hope we're on the right side of history here," Benioff told the assembled press. "It's a highwire act. We're dealing with the most avant-garde technology." Salesforce unveiled its Agentforce offering last week as a platform for augmenting work across sales, service, marketing, and commerce functions. Users can build customizable AI agents to analyze data, understand customer needs, render decisions, and take action. Out-of-the-box agents can be customized in low-code or no-code environments, driven by a Salesforce reasoning engine called Atlas. In his session with the media, Benioff cited fashion brand Gucci as an example of how the integration of Agentforce has translated into results for the company. Gucci implemented agents in its call center and experienced a 30% increase in sales through improved customer service interactions, according to Benioff. "This was an unexpected secondary gain, they did not reduce any heads" Benioff said. "They ended up with a more valuable business unit. I think this is the right direction for AI. There's no question there will be transformational aspects." In Salesforce's view of the world, those transformational aspects will involve a shift away from enterprises who believe that AI can best be delivered through internal development. This "do-it-yourself" approach may work fine in some areas of business, but the company made it clear this week that it will differentiate itself from vendors seeking to sell tools for building AI in-house. "They're selling you science projects," said Benioff during his session with the media. "Our job is to break the DIY. We have shown them that our approach is better." A key element behind Salesforce's AI agent-based strategy is its existing platform. The company has built Agentforce as an over-arching rainbow across its current solutions that encompass Data Cloud and Customer 360. Its message is that this platform advantage will differentiate it in the increasingly more crowded enterprise AI marketplace. "We're abstracting that complexity out," said Matt DeTroia, Senior Vice President of Alliances and Channel Revenue at Salesforce, in an interview with SiliconANGLE for this story. "Your differentiation is the connection to the entire Customer 360. Now we are traversing customer enterprises well beyond the CRM." DeTroia offered evidence that this differentiation may translate into new market opportunities for the CRM giant. As part of his role in forging partner relationships for Salesforce, DeTroia has noticed interest in AI agents among large business consulting organizations that had previously leveraged his company's resources differently. "We are seeing a degree of interest that I've never seen before from new types of partners," DeTroia said. "Three to five years ago they weren't engaging with us the way they are now. That's very encouraging for me. They are a bellwether for where customers are going." Despite these positive signals, questions remain around how much impact Salesforce's big bet on Agentforce will have on the company's overall competitive picture. How data is used and leveraged in the context of various AI initiatives has often dictated market advantage, and this is viewed by some analysts as a key driver in the Salesforce strategy. "I believe this is an advantage inside the Salesforce ecosystem and product set, I am less sure that this is an [external] advantage nor do I believe it is intended to be a generic AI play," said Rob Strechay, Managing Director and Principal Analyst at SiliconANGLE's theCUBE Research. "Salesforce needs to hold serve and keep the data on their platform. With companies like Oracle and Databricks, to name a few, offering to suck data out of Salesforce and combine it with the data in their platforms to service AI, they need to be better with AI than the competition is with the data that is domain to them." After launching a projected 1,200 agents during Dreamforce in San Francisco this week, the company hopes to have one billion users interacting with Salesforce agents by this time next year, according to Benioff. While that large number may not be as unlikely as it sounds, the uncertainty and novelty of the agentic AI journey, much like riding in a driverless car, may indeed lead Salesforce and its customers on a wild ride. "Customers have this deep frustration and excitement around AI, and they hold both in their hands," said Benioff during his session with the media on Tuesday. "We want the market and the technology to come together. This is a marquee moment for us."

Salesforce CEO: Get ready for agents - the third wave of AI

Marc Benioff outlines Salesforce's AI future - with agents taking the fore The future of AI is set to be powered by agents able to provide a customizable and personalized experience like never before, Salesforce's CEO has declared. "This is the third wave of AI - agents," Marc Benioff said in his opening keynote at the company's Dreamforce 2024 event in San Francisco, championing its new Agentforce platform as the next major leap for the technology. "This is what AI was meant to be," he declared. This year's Dreamforce keynote was heavily focused towards AI, with Agentforce taking center stage during the event. Announced a few days before the event, Agentforce is designed to improve productivity and efficiency across sales, marketing, commerce and customer service. Benioff was keen to emphasize that this was a new era for AI - one that does away with "DIY AI" and doesn't require customers to do the heavy lifting when it comes to implementing AI. "Agentforce has to be the biggest breakthrough we've ever had on technology, and I think it's the biggest breakthrough I've seen in a long time in artifical intelligence," Benioff declared. "I don't think you can DIY this - you want a single, professionally managed, secure, reliable, available platform. You want the ability to deploy this Agentforce capability across all of these people that are so important for your company." Agentforce bots will be able to carry out tasks independently, reducing the need for human supervision and improving overall business output. "We are all going to be using agents, but we're going to do it all within our Salesforce platform," said Benioff. "This platform is the key strategic motion of Salesforce, and now this platform has the best AI in the world," he added, declaring this year's event "the most important Dreamforce ever." "We have the data and the metadata and the workflow and the business process and the security model and the sharing model, and all those things that we love and have used and are deeply wedded to for 25 years -- it turns out those things make a more accurate AI." "It's not just what is possible, but what you're going to make possible," Benioff told the thousands of attendees at Dreamforce. "Our goal is simple, for you to augment your employees, create better experiences, and deliver better business results for your companies."

Dreamforce 24 - how Agentforce can simplify AI development and governance

Salesforce's new Agentforce platform introduces a new programming model for interconnecting and orchestrating workflows across special-purpose agents. To users, it will look like more capable chatbots that can interact with web apps and copilots that listen to calls. But the real magic is on the back end, where it could dramatically simplify application development, improve security controls, and make investigating and fixing problems easier. I sat down with Adam Evans, SVP Product, Salesforce AI Platform, to understand how it works and the various mechanisms the company is using to improve trust and experience design for customers and employees. (Evans came to Salesforce as part of the RelateIQ acquisition in 2014, where he was developing predictive analytics tools long before the generative AI boom.) To appreciate the significance of Agentforce, it's important to dis-ambiguate the UI side of chatbots and copilots and what goes on under the hood. This is important since, on the surface, the Agentforce ecosystem will surface a similar experience, just more capable and trustworthy. For example, early chatbots required a declarative flow to guide users through different kinds of problems and queries. The back end of Agentforce surfaces a new approach for setting up the agent logic - powered by AI in some cases and other more structured logic and API access in others. Evans gives the example of one customer who spent a year setting up a chatbot using the old approach. They were able to roll out their first Agentforce chatbot in a couple of weeks using the new approach. Evans explains: Not only is it an order of magnitude or two less work, but who does it? Instead of having developers do it. They add the subject matter expert to it because they're the ones that understand the use cases of the customers. With regard to copilots, there are two interesting distinctions to clarify. First, one that Evans did not directly touch on is the first-generation copilots, which were all-or-nothing affairs, such as code autocompletion capabilities in developer tools. Chris Middleton recently pointed out some problems that can arise when trusting these suggestions too much. The second generation of copilots, such as those Salesforce has been building, can surface suggestions reactively within the flow of work. For example, they could allow someone to ask questions about how to respond to a case to help a customer or make suggestions for next steps. The key here is to assist the users, but the user must ask since it does not take action by itself. The big leap with Agentforce is that agents can be pro-actively triggered and work in the background. Not just from within the Salesforce screen, but also from answering emails directly or facilitating processes from voice calls to an 800 number. Additionally, they can non-conversationally work across more workflows to drive automation by looking at and analyzing information in the background. For example, they could analyze a call with a customer to suggest next steps for a seller and identify lessons learned. Details about product defects that surfaced in the call could be forwarded to the product team. At the same time, it does need a few extra guardrails since you don't want to spam the product team with irrelevant problems. So, another part of the launch is a guardrail infrastructure that ensures that agents follow the instructions and adhere to business policies. These guardrails also create a foundation to ensure that an agent asks for help from other agents or humans when required. Evans suggests there are multiple levels of guardrails Salesforce has been thinking about: Think of it as like an AI supervisor that's watching the conversation. So, as you're talking to an agent, another agent is looking in the background and making sure that it's following policy. Additionally, we are model agnostics. There are only a couple of models that are capable currently of doing the higher-level reasoning and planning. So we don't just run the OpenAI GPT-4 family. Salesforce recently published a CRM benchmark for assessing how various models perform in responding to multiple types of inquiries. This helps to identify when a particular model is good at a certain kind of reasoning and planning that a customer can trust. Evans says that even though Salesforce has been developing some of its own AI models, they are not in the AI business: We want to be kind of agnostic to this, like Switzerland. We want to build that supervisor above it, the guardrails above it, the orchestration above it, and allow whatever the model is of choice of the day to operate. An essential element is that many types of higher-level policy kinds of agents require a declarative logic that can be audited and vetted by subject matter experts. It's not just for preventing hallucinations but also for ensuring an agent does not do the wrong thing. Evans explains: It's very similar to the kinds of systems we design around temporary staff and whatnot. In a contact center where you've got a seasonal employee, what is on the screen for somebody representing your brand and talking to your customer? Do they have a button that can do everything, or do they have limited access? At what moment do they have to actually ask for a supervisor or after they get certain information? There's a progressive revealing of what the next thing is that they can do, and so on and so forth. It's exactly like that. The way Evans frames it, Agentforce creates a foundation where multiple models are competing for a seat and they have put guardrails around each one, just like they might put guardrails around new employees that management might not yet fully trust. He concludes: We will continue to launch more and more techniques around basically enterprise consistency, accuracy, supervision checks, guardrails, policy rails, all of these kinds of things. We're going to keep doing this because the technology is transformative, and its flexibility is a feature of it. But sometimes, it's also an issue about its flexibility. And so what we need to do is that's where you kind of put the box around it and let it have that kind of controlled garden. We want to keep its flexibility and conversational nature, and we also want to make it easy for people to set up so it's not programming. Agentic systems are coming. The big deal is that that approaches like Salesforce's suggest a way of breaking many kinds of processes that could be handled by a single large language model that hallucinates a lot into much smaller and more capable domain experts. And Salesforce's work to create guardrails and controls is significant. Some of the best agents might not even use Large Language Models at all. They might prioritize an API to access or enter data, logical programming to enforce policies, or vector databases to shortlist information related to solving customer problems. Salesforce's approach with this new platform suggests an important way for all enterprise vendors to think about improving the experience of developers, subject matter experts, and users across most industries in a way that grows trust.

How Salesforce is transforming business processes with agentic AI

The future of Salesforce's platform and AI ecosystem: theCUBE's Dreamforce event wrap-up analysis Salesforce Inc. is at a crucial turning point. Beyond customer relationship management and other business-related software solutions, the company is navigating a path to infuse business processes with cutting-edge agentic AI. "I think there's a reason why [CEO] Marc Benioff said this is the most important Dreamforce ever," said George Gilbert (pictured, left), principal analyst at theCUBE. "It's very hard to turn a platform company into an application company or vice versa. If they pull this off, they'll be one of the largest to have ever done it and they'll transform the company. It's not just the Customer 360, but anything that customers choose to connect to the Customer 360 that can activate ... an agent. Salesforce becomes the platform on which others can build agents." Gilbert was joined by fellow theCUBE Research analysts Christophe Bertrand (right) and Rob Strechay (middle) for an analyst wrap-up segment at Dreamforce, during an exclusive broadcast on theCUBE, SiliconANGLE Media's livestreaming studio. They discussed the key takeaways from the event, the Salesforce platform's evolution and how AI is poised to reshape the business landscape. Agentic AI is poised to supercharge business processes Salesforce's vision is to turn its vast data resources into actionable insights through its Customer 360 platform. This goes beyond managing customer relationships -- it's about harnessing data to activate AI agents that can automate complex business processes, according to theCUBE Research analysts. The integration of agentic AI signals a new era where Salesforce becomes a springboard for companies to build customized, AI-enhanced business tools. "I think if they can pull this off, where it becomes a more dynamic app platform, more personalized, really focused on those processes all the way back to the data, it's going to be a clear win for them -- especially where they've abstracted and they're not a cloud," Strechay said. "They're sitting on cloud; they're sitting on IaaS. I think that's a huge win for them from that perspective." AI agents can essentially create an "org chart" of microservices that can think and act independently, calling on human intervention only when necessary, according to Gilbert. This division of labor allows companies to capture expertise in routine functions while enabling humans to focus on more complex decision-making. However, the success of these AI agents is contingent on their access to reliable data. "Agents can call on other agents, and when they're not confident of a step in a process or an outcome, they can then bounce up to an inbox to a human to supervise them," Gilbert said. "The point of all this isn't to eliminate humans so much as to capture expertise for the easy jobs and processes." Dreamforce also brought renewed focus on the role of developers in this AI-driven future. Salesforce's low-code tools are at the heart of this strategy, allowing companies to customize AI agents and business processes easily. With these tools, citizen developers can build and train agents using simple language or by setting specific goals, theCUBE Research analysts pointed out. "It's always going to be about good data -- there's no doubt about it," Bertrand said. "That's the part, to me, that is the constant. I think number two is how you train agents. How do you train humans to train agents and vice versa? I think we will see in the future how things evolve, but I do believe that while there will be some entry-level jobs that may be eliminated." Here's the complete video interview, part of SiliconANGLE's and theCUBE Research's coverage of Dreamforce 2024: