Can I replace my website with ChatGPT prompts?
Imagine visiting your favorite fashion brand’s website and seeing, instead of logos, graphics and welcome text, a never-before-taken selfie in a sweater you bought last year but don’t own yet. please. New spring collection from the brand. That’s what Salesforce promised companies last week when it brought AI everywhere in its “AI cloud” for customers deploying entire software stacks that embed AI in commerce, marketing, analytics, communications, collaboration, and more. One of the innovations.
“Imagine going to a brand’s website and not seeing tons of content and copy,” says Sarah Franklin, president of Salesforce. “We’re looking at a prompt. And this prompt is here with an intelligent conversation pulled from CRM based on AI.”
AI shopping assistants are just a small part of what the company offers.
All of this is part of a larger vision announced by Salesforce to deeply integrate generative AI into nearly everything a company does, from sales to service to internal collaboration. For example, with Commerce GPT, Salesforce auto-generates product descriptions that marketers can select, edit, and push live. Translate text into dozens of different languages, provide analytics in response to natural language queries in Slack conversations, compose marketing emails when prompted, add store locator to emails, and automatically message messages personalize to. If your brand needs more customization, Einstein GPT can generate custom code in seconds that developers can customize as desired. And when a customer requests support, Einstein provides location-specific, context-aware services that include factors such as local weather conditions, public events and locations, in addition to the customer’s history and knowledge of the company’s products. Help your agents respond by suggesting usage and usage options. service.
Oh, and of course suggest additional purchases based on previous orders.
AI-first companies may be great, but they don’t come cheap.
According to Constellation Research’s Larry Dignan, Salesforce’s AI Cloud Starter Pack starts at $360,000 per year and includes an “AI readiness assessment with Data Cloud, MuleSoft automation, Tableau Analytics, Slack, CRM, and Salesforce services.” contained.
It’s a bold and overarching vision, but it also faces competition from other major software vendors such as IBM.
“We’re looking for areas where we see rapid improvements in productivity and time to value,” said Rob Thomas, IBM’s senior vice president and chief commercial officer. “We see great potential in foundational models that are reusable and require minimal training in helping a leader implement his AI to fundamentally change the way his business operates. I think.”
Salesforce’s vision is one of the opening salvos of an AI arms race that no company can ignore.
It’s true that all software demos look great at first, but there’s always a lag between new innovations and real business value, and some of the brightest promises never come to fruition. It is also clear that large-scale language model AI systems are proven. Even in the early stages, the results were incredible. What Salesforce has demonstrated, while undoubtedly accompanied by growing pains, is also evident in significantly speeding up many key business processes.
Companies that do not adopt such technology risk being left behind.
“Companies that embrace AI-driven opportunities and learn how to capitalize on them effectively will outperform less-adaptive competition—competitors that are likely to disappear,” said an AI expert at Centric Consulting. Joseph Hours said.
However, there are also risks.
One in four executives who responded to a recent survey by data analytics and AI vendor Altair said more than half of their AI projects failed, and 42% said at least one failed. I am answering Workflows and processes come with many changes, all with risks.
That said, AI is often best applied to scenarios where talent is scarce or reach is limited. One example is the AI-generated minor league baseball articles The Associated Press began using in 2016. There wasn’t enough demand for these articles to justify the human writer’s time.
Likewise, good customer service is currently lacking. (Recall his recent two-hour call with a mobile operator or television service.)
What I’ve seen from Salesforce’s new generative AI service and support product is that it can help service reps significantly increase the number of people they can assist at the same time. and The quality of help they can provide. This has to be proven in practice in different applications and fields, but if it’s accurate, it’s literally a game changer.
“End-to-end AI integration workflows dramatically improve response times, flexibility, reduce risk, facilitate faster experimentation, and accelerate time to market,” said an Infosys executive. VP and Head of AI and Automation Balakrishna D. said.
Not everyone is convinced that AI is ready for its heyday. Several executives and founders I spoke to said generative AI in particular still suffers from hallucinations, including one ill-fated lawyer who acquired from GPT-4 You mentioned a fictitious precedent used and condemned in court. Some have suggested that over-reliance on AI systems risks the loss of human expertise, or the risk of AI-induced bias permeating corporate decision-making and communications. An additional risk factor often cited by companies using generative AI is legal exposure if the model is trained on copyrighted data.
“AI is on its way to becoming much smarter than humans, but it can also lie unintentionally and frequently,” says Marc Weinstein, founder of alternative social media network MeWe.
Those are all real issues.
Salesforce is not ignorant of this challenge, and all of its large-scale language models are trained on 100% ethically and legally sourced data, not just the company’s own data. repeatedly emphasized. The word “trust” appeared 28 times in Salesforce’s records for an event in Chicago last week to announce many new products and features, and company president Franklin called his Einstein GPT for both companies “trust.” and emphasized trust and fairness. The world’s most trusted enterprise generative AI. “
“In the world of AI, it is more important than ever that trust is a value first and foremost,” she said. “Trust our products and our people and trust that together we will always do the right thing.”
This includes not commingling customer data for training purposes and employing solutions such as data masking and zero retention to ensure companies do not lose control of their information or influence the success of their competitors. includes doing.
Another important aspect is not to exclude humans.
“The most important thing is to have humans to give feedback and help the AI continue to learn,” says Franklin. “And in this world where trust is paramount, it’s really very important.”
Two phrases that were lavishly repeated throughout the company’s presentation were “Help us instead” and “Edit and take credit”, demonstrating how Salesforce will replace generative AI with human talent. I emphasized that I see it as a tool to extend rather than.
It rings a bell for many.
“AI shouldn’t have the ultimate final say on all issues,” says David Brossard, chief technology officer at automation company Axiomatics. “Employees will be able to perform, test and validate existing tasks more quickly. in humans.”
follow me twitter Or LinkedIn. check out You can find my website and other works here.
