Artificial intelligence at Salesforce: An inside look

At times, Salesforce’s portrait of an AI-powered future sounds too good to be true. For a real-world review of the company’s plans from an outsider, I turned to Pedro Domingos, an AI specialist at the University of Washington and author of Main algorithm.

Domingos says Salesforce is “a bit late” in the field and may have a harder time than expected to fully integrate AI at deeper product levels. But he thinks the company is on the right track: At this stage in the evolution of AI, there is much more to gain from putting basic tools in the hands of more people than from killing a few more. hundred efficiency from an algorithm.

Domingos also said that Salesforce’s relatively slow access to AI – compared to IBM or Google – doesn’t necessarily hold it back. “They are still a small player in this space. But other companies come from behind and move pretty quickly – look at Facebook. Just because you’re a late entrepreneur doesn’t mean you won’t be a leader in a few years.

Sales forces face a crowded field in the fight to get AI tools working on behalf of the crowd to shake hands warmly. Competitors include giants like Microsoft (with a LinkedIn Sales Navigator) and Oracle, as well as smaller competitors like SugarCRM and startups like Conversica (which later used AI to self-control automation chats with incoming sales clues). If Salesforce succeeds in taking the lead in the company’s insane AI race today, the insiders point to one advantage as its secret weapon: organized customer data stores and Labeling consistently.

Which highly competitive, highly paid data scientists are people trying to hire? Today, they spend a lot of time “preparing data,” which means figuring out how to prepare loads of information so that it can be digested by machine learning programs and produce good results. There’s a lot of information that has to be prepared and adjusted before most AI systems can start making predictions.

This represents an irony in the automation that underpins AI. Today, Domingos points out that too often, IBM and Accentures around the world are just throwing armies of experts into customer problems. “What they do at the end of the day is they actually have the human labor to do this work,” he said. “That makes money but it can’t be expanded.”

But all Salesforce customers imported their data into a single software platform, even though many of them added their own custom software. “People put everything in there,” said Tallapragada, Salesforce’s chief technology officer. Salesforce doesn’t look at the content of its customer data, but they know a lot of how it is organized. “Salesforce’s advantage is metadata. That allows us to automate everything, ”says data science director Nabar.

For all of the Skynet dreams and nightmares that today’s advances in artificial intelligence cause, the winners and losers in this transition will likely be determined by what scientists computer learning is called “data cleaning”. In other words: No matter how smart our bots are in the AI ​​future, order is important. Clean up after work and remember to wash the files before leaving.

Let the others conquer the Go and solve the knots. Sales force can achieve victory through tidy power.

Directed by: Michelle Le

Leave a Comment