The Unveiling
So first… a little bit of background:
At TrailblazerDX last year, Salesforce finally introduced their new service – Einstein GPT.
Einstein GPT has two main features: Einstein Co-Pilot and Prompt Builder. With the help of data from Salesforce or other external data sources (such as Data Cloud, which was also introduced last TrailblazerDX), these two features can help companies in making decisions and providing solutions to their customers.
As we neared Dreamforce 2024, Salesforce was dropping hints around AI and how their Einstein tool will enhance automations for companies with the use of generative AI. All was eventually revealed at the Dreamforce 2024 Keynote with the official introduction of Agentforce, as Marc Benioff discussed how Agentforce can help companies. He explained how Agents can help customers to solve problems, and process requests with the use of customer data. In a bold statement that really captured everyone’s attention, Benioff projected one billion agents to be empowered by the end of 2025.
In the world of generative AI, Data plays a crucial role in accurate decision making. Only high-quality and well-formatted data is used for model training to ensure correct and relevant outputs from the models. Now, with the use of Salesforce Integration Platforms such as Mulesoft and Data Cloud, companies can ensure that accurate Data is being ingested to Agentforce Platform, and generate real value for clients.
So what exactly is Agentforce? And how can your company use it?
Agentforce (formerly Einstein Copilot) is a new platform from Salesforce which mainly focuses on generative AI as its foundation. The new platform will be generally available this October of 2024. Agentforce’s Agents (which, in this case, is essentially another term for bots) use Large Language Models (LLMs) to accurately analyze complex or simple requirements and provide solutions based on data in Salesforce CRM or external data sources such as Data Cloud. Agentforce is now using Atlas Reasoning Engine (also see it explained in the Salesforce keynote here) to provide more accurate decisions or results based on the data that is collected inside Salesforce CRM. Agentforce can be used in various Salesforce Clouds including Sales Cloud, Service Cloud and E-Commerce Cloud.
Simply put – Agentforce now allows companies to create a limitless AI workforce of Agents able to analyze data, make decisions and take actions based on customer inquiries and requests.
Service Cloud can also take advantage of Agentforce features such as Case Management, prompting Knowledge Articles or Solutions which will help Service Agents in efficiently solving customer’s requests or issues, and automating case processes. In Service Cloud, the Agentforce autonomous Agent can be configured to generate personalized troubleshooting techniques based on the company’s knowledge article and customer history. It can recommend the right knowledge articles and routes/escalates the case to agents/managers based on criteria.
In Sales Cloud, the Agentforce Sales Agent can be set up to nurture pipelines and coach sales reps ahead of client meetings. It can qualify inbound leads and automatically route those requiring human attention to Sales Agents. The Agentforce Sales agent also schedules follow-up meetings with clients and tailors communication tone by analyzing the company’s sales documents. The Agentforce Sales Coach offers personalized coaching for sales reps, prepping them for meetings through role-play simulations that mimic customers using their CRM data. Einstein provides feedback after calls to help reps advance deals. During meetings with the customer, the Agentforce Sales Coach can offer live recommendations based on real-time conversation, with Einstein automatically logging the call details to the relevant CRM record, without any manual data entry.
In Commerce Cloud, the Agentforce Personal Shopper Agent can be configured to assist shoppers with order status updates or personalized product recommendations through real-time conversations. While some actions come standard with the Agentforce, Administrators can also customize or create new actions to fit their business needs, ensuring a personalized and seamless shopping experience.
Final Thoughts
Agentforce does have the potential to create real value for organizations. However, it is still early days and many will still doubt the integrity and accuracy of the results or recommendations from Agents. As is typically the trend with Salesforce releases, enhancements will be ongoing as the Salesforce team improves and fine tunes the product. (Our personal wishlist is to incorporate GPT4-o in the next few releases, as this would be a game changer in terms of AI – CRM competition.)
Agentforce is a significant advancement, especially when it comes to automation and decision making. Depending on the needs of the organization, this tool could represent a real opportunity to scale their workforce on demand. As Agentforce is a paid tool, organizations must evaluate if they can derive value from Agentforce and its features to ensure ROI, but it certainly represents a new wave of AI and automation in customer service.
Agentforce for Sales and Service will be generally available October 25, 2024. (But you can access the first Trailhead now.) Our team is excited to get hands-on experience with this new tool, and we’ll be sure to share our tips and insights as we learn more.
Post from Glenn Balmonte, with contributions from Chris Sekar and Ishan Babbar
Want to talk about Agentforce? Or any of the other announcements from Dreamforce? Get in touch!
