<strong>Einstein GPT: More than Salesforce’s Hat in the AI ring</strong>

By Pratima H

It was a big splash at Trailblazers DX 23. It had to be. The Salesforce announcement about coming out with its own Einstein GPT has a lot of connotations after the ChatGPT disruption and efforts by Google and Microsoft to catch up to the blitzkrieg.

But what’s fascinating to note is the conscious emphasis by the 24-year wise Salesforce on being ‘patient’ and ‘responsible’ here. As Sarah Franklin, President and CMO, Salesforce explained to CyberMedia answering whether AI would be the proverbial printing press or a Steam engine, “We want to build the future with AI in a way we are proud of. There are a lot of questions about control, about governance and who makes the algorithms. These are important to address along with the productivity gains expected. But when computers came the world thought they would replace humans. It did not happen that way. That’s the same with AI.”

This posture is interesting. Specialty, as there is a lot of debate about how soon the pilots will translate into something out of the oven and on the shelf. But almost every executive at Salesforce addresses that question with a ‘slow and prudent’ vibe.

We asked Jackie Rocca, Senior Director, Product, Slack at the Trailblazers DX 23 and she told that “The next stage is beta. It will be a slow and cautious roll-out because we want to be ethical and responsible here.”

Even a panel on AI Ethics maintained that Salesforce is cognizant about issues like explainability, about the trade-off between speed and creativity with AI models and about customer’s data. Executives assured that ‘we do not monitor customer data, we cannot see their data, but we want to provide tools to customers to empower them for what they want to do with AI modeling.

In fact, this new pony could mean a lot for developers would want to ride the AI track.

As Janani Narayanan, Director of Product, Digital Intelligence Automation, demonstrated with Einstein GPT—A lot of coding snippets and testing support can be tapped into here. It gives a ready and smart scaffolding that developers can, then, build upon or refine.

Generative AI is the next new tool, and the prospects are quite fascinating, pointed out Gaurav K, Founder and CEO, Vanshiv and a seasoned trailblazer himself. “It would be interesting to see code-on-the-fly now. But developers need not be worried. They will get a starter-template with Einstein GPT that they can refine from there on. There is no need for them to be worried about their jobs. It is all about a chance and capability to write better quality code. Developers can build apps any way they want. It opens up the ecosystem to a new level. There are lots of possibilities now.”

With 9.3 million new Salesforce economy jobs and 1.6 trillion new business expected by 2026, and a new AI fund worth $250 million, it would be really a new bend in the road from Salesforce from here.

And Salesforce is approaching this new turn, apparently, with discretion and not a hurried leap. In her keynote, too, Sarah Franklin went after the horns of AI-debate head on. “Generative AI is creating new opportunities to supercharge productivity for everyone. Yet, it is raising many questions. Does AI really, truly, deeply understand me? Who is writing these algorithms? It is always going to be right? Can I blindly trust AI? What if generative AI could learn from customer data?” All these questions seem to matter for Salesforce’s approach to generative AI. That’s why it is underlining the ‘human in the process’. That’s why it will mean a lot to enable its own models as well as a user’s LLM or ecosystem models. As Franklin said, “It’s here to help us – not to replace us.”

Sam Altman of Open AI was there virtually at the keynote at this conference, and he said – “We are excited to see Einstein GPT in action. Can’t wait to see what you do all next.”

That’s tipping a hat to another athlete. It would be great to see Salesforce spin AI in a cautious, responsible, and yet-game-changing way. And change the ring.