A surge in corporate demand for artificial intelligence is driving growth for startups that help companies prepare data for software tools designed to glean business insights like gauging demand or spotting supply chain inefficiencies.

Companies across the economy have collected data for years—from orders and transaction tallies, to website traffic, inventory levels and shipping rates, among hundreds of other sources—without always knowing what to do with it. More recently, cloud-computing services have offered nearly unlimited capacity to store massive amounts of raw data.

But before data can be fed into AI models, it needs to be annotated with labels—a laborious task that most companies do manually.

Scale AI is trying to do much of that tagging and identification work using automated systems. The San Francisco-based data-labelling and management venture hit a market valuation of more than $7 billion after its latest fundraising round in April.

Former Amazon.com Inc. executive Jeff Wilke, who led the retailer’s world-wide consumer division, joined the company that month as an adviser to Alex Wang, the company’s 24-year-old co-founder and chief executive. Michael Kratsios, the former U.S. Chief Technology Officer in the Trump administration, in May stepped in as managing director and head of strategy, focused on expanding Scale AI’s customer base outside of the tech sector.

“We help customers unlock this data,” Mr. Wang said. “That gets them off the starting blocks.”

iRobot Corp. , maker of the Roomba autonomous vacuum cleaner, started working with Scale AI late last year to build out more features for its room-cleaning robots, said Chris Jones, the company’s chief technology officer.

The company is using image-gathering sensors on board models tested by volunteers to generate training data that enables the robot’s built-in AI system to learn how to better navigate a home, such as identifying kitchens, dining rooms and other locations by recognizing different household appliances or furniture, Mr. Jones said.

Customers can then tell the robot to clean a specific room, or a defined area of a room—like around the couch, for instance—and return to a charge station when it is done, he said.

“The better the data you have the better performance you get at the other end,” Mr. Jones said.

So far, he added, iRobot has annotated two million images, using Scale AI’s infrastructure to label and fine-tune massive amounts of data.

Other startups offering similar data-labeling and management services include Labelbox Inc., a Bay Area firm backed by Andreessen Horowitz and Kleiner Perkins, and Tel Aviv-based DataLoop, which raised $11 million in October, more than doubling a previous $5 million seed round.

Scale AI has raised more than $600 million in venture capital since its 2016 launch, when it was narrowly focused on tagging images and videos for self-driving cars, with investors including Dragoneer Investment Group, Greenoaks Capital, Tiger Global Management and Index Ventures.

Global venture capital funding for AI startups hit a record-high $31 billion in the first half of 2021, according to a report last week by market-research firm CB Insights. Scale AI’s April funding round was the fifth highest deal among AI startups so far this year, the research firm said.

The market’s rapid growth is propelling Scale AI’s expansion efforts, which include a suite of software-based services designed to help companies gather, annotate, cleanup and manage data, while also building and monitoring their AI models, Mr. Wang said.

Scale AI’s software package Nucleus, for instance, enables firms to quickly identify and fix mislabeled data, or refine existing data labels to improve algorithmic training and boost an AI system’s performance, he said.

“Alex had an unusually thoughtful understanding of the market,” said Mike Volpi, a founding partner at Index Ventures and a former chief strategy officer at Cisco Systems Inc. “Data is the gold mine,” he said, “but you have to start somewhere and labeling is a good way to start.”

Mr. Volpi said a lot of business decisions that in the past were based on human intuition or experience are increasingly being handled by smart software. “Business data has been around for decades, but AI goes one step beyond showing you a chart, it offers an answer,” he said.

Write to Angus Loten at angus.loten@wsj.com