Generative artificial intelligence (AI) capabilities will soon be pervasively available across a wide range of applications with Salesforce today becoming one of the first application providers to make this move.
Salesforce today launched Einstein GPT, an extension of its existing AI capabilities that employs data created by OpenAI with data Salesforce collects to create a large language model that can be prompted to generate personalized emails or content that increases marketing campaign response rates.
In addition to making that capability, Salesforce is also embedding Einstein GPT within its customer service and marketing platforms in addition to the Slack messaging platforms it acquired in 2021. In the case of Slack, organizations will be able to use Einstein GPT to automatically create summaries of conversation threads. In the future, Einstein GPT will also be extended to the Tableau business intelligence applications as well as the Mulesoft integration platforms that Salesforce acquired in 2019 and 2018, respectively.
Salesforce is also making Einstein GPT available to developers that build applications on top of its software-as-as-service (SaaS) platform. Salesforce also announced a $250 million Generative AI Fund that will be used to invest in startups that build applications using Einstein GPT.
Organizations that employ those applications will be able to either connect data to OpenAI’s advanced AI models out of the box or import their own external model, said Clara Shih, Salesforce Service Cloud CEO. The line between tasks that will be handled by bots infused with AI and humans continues to blur, but AI is not at the point where people are being replaced, she added. “Humans need to be in the loop,” says Shih.
Salesforce earlier this week published a survey of more than 500 senior IT leaders that finds a majority (67%) of organizations are prioritizing generative AI for their business within the next 18 months, with one-third (33%) naming it as a top priority. The majority of senior IT leaders (57%) also said they believe generative AI is a game changer, with even 80% of those who say the technology is ‘over-hyped’ agreeing that generative AI will help them better serve customers. Overall, 33% of respondents said generative AI is over-hyped and nearly two-thirds (65%) of senior IT leaders can’t yet justify the implementation of generative AI.
Most respondents also have concerns about potential security risks (79%) and bias (73%), the survey finds. A total of 59% believe generative AI outputs are inaccurate and almost two-thirds (63%) agree there is bias in generative AI outputs. Just under a third (30%) said they must have ethical use guidelines to successfully implement generative AI within their business.
Nearly three quarters (71%) also noted generative AI would increase their carbon footprint through increased IT energy use.
It’s too early to say to what degree generative AI platforms will be pervasively employed within organizations. Many employees are already using ChatGPT to automate a wide range of activities while other organizations have for now opted to ban using it because of veracity concerns. The OpenAI language model is based on data sources across the Internet that includes a lot of misinformation.
It will be up to each organization to determine precisely how generative AI platforms will be employed within their organization but as is the case with all things AI it comes down to the quality of the data employed to train the model. If your organization trusts the data, then the odds are good the recommendations being made are sound. In not, those same recommendations could, down the road, wreak all kinds of havoc.