Dreamforce Recap: AI everywhere, for everyone - what Agentforce really means for businesses

The AI era is upon us—and every major tech company is scrambling to be the leader. That position is still very much to be determined, but one thing that was on full display at Dreamforce 2024 was that it’s not just the latest and greatest technology that will determine the AI winner—it’s which company provide the best solutions to real business problems.

Case and point: Agentforce, Salesforce’s newly launched platform that integrates autonomous AI agents with existing CRM systems to enable companies to automate tasks and connect workflows within Salesforce and across third-party systems.

Setting aside whether or not the agent technology Salesforce is rolling out is better or more advanced than the likes of other tech companies, the real question companies need to ask themselves is what value does this add to the business and how is it different from other tools on the market?

Agentforce’s key advantage: Optimized organizational workflows

To me, the main differentiator of Agentforce is that it is designed to automate and streamline organizational workflows within the Salesforce platform. This is markedly different than other AI agent tools which focus on personal workflows.

Think of it this way:

With a co-pilot tool, a retail service agent may use generative AI to draft a response to a price adjustment request. That’s a great use of the technology in that it will help the person be more productive. But will it have a real impact on the company’s overall performance? Perhaps not.

Agentforce, on the other hand, does two things differently: 1. It applies generative AI to different parts of workflows within the Salesforce platform; and 2. It makes the technology part of every user’s experience.

For example, when the company’s support channel receives a price adjustment request, the system will kick off a workflow that is enabled by one or more AI agents. One AI agent will read the initial customer query and categorize the request; this agent will then hand it off to the next agent that will review and analyze the company policy for price adjustments; the second agent will hand the recommendation to a third agent, which will craft the response to the customer and submit it to a human agent for review and verification. This creates a hybrid service team, where a series of AI agents do much of the time-consuming tasks in the background to augment human workers.

The real value of Agentforce is that it compounds the advantages and benefits of the technology by connecting different use cases and making it a core part of the work experience. When implemented strategically, companies can extend the capabilities, bandwidth, availability and productivity of their workforce, without investing in new technology tools or hiring more people.

The snag: Data

If Agentforce sounds a little too good to be true, then that’s because it’s real-world application often has a few challenges.

First, using this feature assumes that the workflows are defined in Salesforce and can be integrated with the platform’s native AI features. This means that workflows need to be set up correctly within the platform. While that tends to be the case for most enterprise organizations, it’s not something that companies can take for granted.

The bigger issue that affects far more companies is data. As with any generative AI solution, the power of Agentforce hinges on the quality of the data the tool is using in Salesforce. And therein lies the problem: Agentforce works best when it is integrated with Data Cloud. And a lot of enterprise organizations are resistant to using Data Cloud due to high cost or complexity.

So does that mean that Agentforce is out of reach for all practical purposes for companies that don’t use Data Cloud? No. There are certainly ways to get around this issue through third-party integrations. While the end result may not be as robust as if the company used Data Cloud, the AI-led solution will still be far better than using an antiquated chatbot tool or relying on human agents to manually complete every step of the process.

Solving the data challenge (or not)

While it may seem counterintuitive (or even blasphemous to data purists), the most effective way for companies to address the challenge with Agentforce is by not focusing on the bigger picture.

I firmly believe that a pristine data estate is one of the greatest assests a company can have today. However, for many enterprise organizations, this is an impractical goal. Achieving optimal data management could take years and cost tens of millions of dollars. Meanwhile, the opportunity is here and now.

Long story short, companies need to have a strong data strategy to use generative AI in any capacity, but they don't necessarily need to solve their data problem in totality. Instead, they need to focus in on the specific use cases they want to tackle and fill the data gaps standing in the way over success.

AI is the future

In the coming months, we are likely to see the big tech players, including Salesforce, take additional steps to improve their AI capabilities and differentiate from competitors. As these conversations unfold and we debate the value of AI at the individual or organizational level, what we really need to remember is that AI is going to be an integral part of our future—and data is an immutable part of every AI strategy. The best thing companies can do regardless of who ultimately wins the AI race is act now.

I'd love to hear your thought on this topic. Feel free to reach out to me or anyone at 高知特 Cognizant to discuss or if you need help with the AI and Data journey in your organization.