CLOUD COMPUTING After the presentation, we caught up with Paul to discuss all things Agentforce and why it is the driving force behind “the most exciting time” in his career.
This foundation led to Einstein in 2016, Salesforce’s predictive AI model that was able to forecast probabilities and outcomes. But the real breakthrough came in September 2023, when Salesforce recognised that their R&D team might have come up with something far more powerful than text generation.
“We’re sitting on a platform that has a tonne of capability that enables our customers to manage their workflows,” Paul reflects. By combining their existing platform with AI, Salesforce created what Paul describes as “this ability to trigger actions – the company’s first truly agentic AI.
Paul’s enthusiasm for Agentforce is palpable, but it’s rooted in years of methodical development. Salesforce’s AI journey began far earlier than many might expect – back in 2014, when the company established an ethics council to ensure responsible AI deployment.
“We kind of saw this was coming,” he says, “and the real focus to start with was actually let’s reflect on our core values and our number one core value is trust.”
When compared to Gen AI LLMs like ChatGPT, agentic AI is in its own league altogether. “Generating text is great, but you need it to do something for you,” Paul suggests. “It needs to fulfil something.”
57
technologymagazine. com