Salesforce CEO Marc Benioff presents Agentforce to Dreamforce 2024 attendees.
Salesforce
Salesforce unveiled Agentforce, the successor to its Einstein AI platform, at last week’s annual Dreamforce event. For Salesforce, this launch marks a leap forward in its AI capabilities; more broadly, it represents a big step toward a future where AI enables day-to-day business operations. The Agentforce suite of autonomous agents is designed to enhance operations across sales, service, marketing and commerce. This should improve efficiency and customer satisfaction, but more than that it could change the way that corporate employees get their jobs done.
Positioning Agentforce As Fundamentally Different—And Better—Than Competitors
At the event, Salesforce went out of its way to highlight Agentforce as being simple to use, accurate in its outputs and a straightforward way to create real business value for enterprises. In contrast, the company positioned competing products such as Microsoft’s Azure AI as a do-it-yourself option with slow rollout and Microsoft Copilot as prone to data leaks, hallucinations and high costs—with limited customer benefits. Salesforce’s framing emphasizes its belief that Agentforce can deliver immediate, reliable results, unlike its competitors in the enterprise AI market.
How will the company prove that differentiation? “There’s only one way anyone is actually going to believe this,” Salesforce CEO Marc Benioff said. “You’re going to have to let [users] put their hands in the soil and get going.” This emphasis on hands-on experience suggests that Salesforce wants to demystify AI and make it more accessible to businesses and end users. It also suggests confidence.
It is worth noting that Microsoft, Google and ServiceNow have all announced their own plans for AI agents, albeit without the fanfare of a major conference like Dreamforce. I will be at Microsoft’s Boston Innovation Lab next week and will get a more in-depth look at the company’s latest AI releases. October is a very busy month for Moor Insights & Strategy analysts, and you can expect us to weigh in further on enterprise AI in the coming weeks.
What Is An Agentforce Agent?
I’ve often said that the best generative AI is like having a good intern to help you do your job. But the chatbots and AI assistants that we’ve grown accustomed to can only go so far; these “interns” needs lots of back-and-forth. An agent, by contrast, should function like a tireless, knowledgeable assistant that’s always available to help employees or customers with specific tasks—with more autonomy than we’ve seen before. Salesforce envisions agents as intelligent, automated helpers that know the ins and outs of a particular job, whether in sales, customer service or another area. Instead of merely providing individual answers or improving individual tasks, these AI agents can now manage entire processes from start to finish.
These agents are in fact applications that are trained for a specific role and given context for the actions required and the guardrails they work within. This allows them to handle routine tasks—or sets of tasks—much more proactively on behalf of employees or customers.
This could involve something as simple as drafting an e-mail or scheduling an appointment, but it could also extend to more complex tasks such as planning an event or providing targeted marketing recommendations based on customer data. (My colleague Jason Andersen has just published an excellent primer on AI agents.)
Agentforce At Work Across Different Functions
Salesforce has introduced pre-built agents with titles including Service, Sales Development Representative, Sales Coach, Personal Shopper and Campaign Agent, along with the ability to configure custom agents. These agents integrate into workflows to free up employees to focus more on strategic, creative or relationship-building activities. If all goes according to Salesforce’s plans, this will allow customer companies to increase both revenue and the job satisfaction of their employees.
At Dreamforce, the company offered examples of Agentforce applied to different challenges across corporate functions. Here are some of the ways enterprises might put it to use.
- Sales — Agentforce for Sales aims to enhance sales teams’ efficiency, allowing salespeople to concentrate on building and nurturing customer relationships and developing strategic sales approaches. Additionally, AI agents can leverage data analysis to predict lead conversion, helping salespeople prioritize their outreach efforts. Providing recommendations for more personalized interactions with prospects and customers could improve engagement and potentially drive sales growth. Additionally, AI-powered sales coaching can provide real-time feedback and insights, helping sales reps improve their craft.
- Customer experience — Agentforce could enhance customer experience by providing swift, proactive, personalized support. This can lead to faster issue resolution and the ability to scale customized service without increasing staff. By offloading low-level tasks to AI agents, customer support reps can focus on more complex interactions, potentially creating a more fulfilling work environment. However, success will depend on maintaining customer trust and data security while ensuring seamless integration of AI agents into existing workflows.
- Marketing — Agentforce promotes marketing efforts by delivering personalized customer engagement and optimizing campaigns in real time. Augmented by Salesforce’s Tableau analytics suite, AI agents can deliver actionable insights and even trigger actions throughout a customer journey to improve multichannel experiences. Again, by automating routine tasks, agents could free up marketers to focus on strategic initiatives. The challenge here is balancing automation with human creativity.
- Commerce — Agentforce presents a significant opportunity for businesses to enhance their commerce operations. The integration of autonomous agents has the potential to streamline various aspects of the customer journey, from personalized product recommendations to efficient order management. The ability to provide intelligent sales assistance and coordinate interactions across multiple channels could lead to increased conversions and a more consistent brand experience. Analyzing customer behavior patterns can equip businesses with the knowledge to make advantageous choices, such as optimizing staffing levels to meet demand.
Salesforce provided a strong example for the commerce use case. Wiley, a big publisher of educational materials, has traditionally hired additional staff to manage the back-to-school surge in customer inquiries. With Agentforce, it was able to automate a significant portion of its customer service, resolving 40% to 70% of cases without human intervention. This illustrates how Agentforce could help commerce businesses efficiently handle fluctuations in demand without bringing on extra staff.
The Devil Is In The Data
To guarantee that AI agents deliver the best performance, it’s essential that they have access to clean data. Organizations must maintain—or, in some cases, establish—proper data collection and management practices to systematically collect customer interaction data, transaction histories and other relevant information and make sure that it is accurate and up to date. Well-organized and complete data will help ensure that AI agents take accurate and appropriate actions.
Agentforce draws information from Salesforce Data Cloud, and the company has taken advantage of this launch to emphasize the importance of Data Cloud in powering the accuracy and capabilities of its new AI solutions. By unifying apps, data and AI agents, the company continues to focus on customer-centric solutions as it navigates the changing CRM landscape.
Looking Ahead: Change Agents
With Agentforce, Salesforce is promising a world where AI handles the mundane, freeing up workers for the meaningful. As individual AI agents evolve and collaborate within sophisticated agentic frameworks, their combined power could grow exponentially. However, this increased strength and complexity also brings amplified risks, including concerns over data security, biased decision-making and overreliance on automation, which could diminish human oversight, creativity and empathy. Additionally, integrating sophisticated AI agents into existing systems may pose technical challenges, while evolving regulations could add new compliance burdens. Managing these risks will be crucial to ensuring responsible and effective deployment.
Big questions remain: Are people ready to trust AI? How do we find the right balance between automation and the human touch? What seemed to be missing from much of Salesforce’s messaging around Agentforce is collaboration between people and AI.
This next wave of AI could be far more meaningful than productivity-boosting chatbots. AI agents have the potential to help create a better future of work where people and technology operate together to achieve more meaningful results. The success of Agentforce will depend on its ability to deliver on these promises. While early demonstrations have showcased potential benefits, the real test will come as thousands of Salesforce customers—without Dreamforce engineers holding their hands—begin to “put their hands in the soil” and experiment with the technology.
