Startups developing business applications from artificial-intelligence software that generates original text and images are capturing the attention of corporate technology leaders—along with big investors.
Known as generative AI, the applications use data-intensive AI capabilities like deep learning and neural networks to recognize underlying patterns in different types of books, articles, photos, drawings and paintings gathered from a massive collection of digital sources. Based on keyword prompts, generative AI platforms can then mimic these sources to produce fanciful stories and pictures of their own.
At first glance, generative AI might seem like more of a curiosity than an enterprise-technology tool, said Peter van der Putten, director of the AI Lab at software firm Pegasystems Inc.
“Creating cute pictures of a corgi in a house made of sushi isn’t exactly a profitable business case, at least not for large enterprises,” Mr. van der Putten said. And yet, he said, “generative AI startups are popping up left and right, in areas such as marketing, support, service and other content creation.”
Jasper, an Austin, Texas-based startup launched last year, has developed a generative AI platform designed to auto-generate promotional blog posts and other marketing materials.
Amid a sharp decline in venture-capital investing deals, Jasper last week announced a $125 million Series A fundraising round, which sets its private-market valuation above $1 billion, the company said. Its investors include Insight Partners, Coatue and Bessemer Venture Partners.
Other generative AI startups offering enterprise applications include Replikr, a New Zealand startup that uses generative AI to create customer-service avatars, and Dutch startup Musico Ltd., which can be used to create original copyright-free advertising jingles. GoCharlie.AI, an early-stage startup based in San Francisco, in May launched a generative AI marketing assistant designed to create promotional social-media posts.
“Our early adopters aren’t technophiles or the fringe,” said Dave Rogenmoser, Jasper’s chief executive. “They are everyday marketers, creators and business leaders in industries across the spectrum.”
Since launching, Jasper has grown to more than 80,000 subscribers. The company charges businesses for access to its AI content platform, and sets prices based on the number of AI-assisted words generated. After its first full year in business, Jasper had $35 million in revenue, the company said.
Jasper’s customers range from entrepreneurs to digital-marketing agencies and Fortune 500 companies, according to Sameer Dholakia, a partner at Bessemer. “Generative AI will fundamentally shift how CIOs innovate,” Mr. Dholakia said.
By promising to save workers hours of productive time creating promotional materials, the commercial value proposition around generative AI has always been clear, he said.
“But the technological infrastructure wasn’t there until just recently,” Mr. Dholakia said, citing foundational open-source resources developed by AI labs like OpenAI and Cohere.
Ritu Jyoti, global vice president of AI and automation at research firm International Data Corp., said generative AI is gaining traction in the enterprise IT market due to “better models, more data, more compute and better economics.”
That traction is getting large enterprise IT vendors on board, analysts said. Microsoft Corp.
has already invested heavily in OpenAI, a San Francisco-based software maker behind GPT-3 and Dall-E 2, the pioneering language and image models that underlie many generative AI tools. Microsoft is expected to increase its investments in OpenAI this year.
Google and DeepMind Technologies, its UK-based AI subsidiary, have also released giant language models, as well as Meta Platforms Inc.,
all of which are downloadable for research. Stability AI, a startup with offices in London and San Francisco which last week announced a $101 million seed round, in August launched an open-source text-to-image generator that has since been downloaded and licensed by more than 200,000 software developers worldwide, the company said.
another enterprise IT giant, is using generative AI techniques to develop a computer-code generator designed to save time for software developers, said Salesforce Chief Scientist Silvio Savarese.
“They can simply describe the problem they’re trying to solve in plain English text,” Mr. Savarese said. The generator then creates the unique code needed to solve it, he said. Salesforce’s tech team created the entire application in-house, according to a company spokesman.
Rowan Curran, an AI and data science analyst at IT research firm Forrester Research Inc.,
said he expects generative AI to impact enterprises from top to bottom.
Still, he said, “the use of text-to-image generators for inspiration, ideation and prototyping for marketers, designers and developers has just barely begun.”
Write to Angus Loten at [email protected]
Copyright © 2022 Dow Jones & Company, Inc. All Rights Reserved. 87990cbe856818d5eddac44c7b1cdeb8