Kyle Miller, the VP of product at Ancestry—the online genealogy company that also operates Newspapers.com and Find A Grave, among other sub-brands—has worked for the company for more than a decade. But it was just recently that AI tools were able to surface a mind-blowing connection to his past.
Miller and his wife were married in Provo, Utah, in a building that, when he was young, had burned down and since been replaced. He had seen the fire himself and, with his father, even watched the building burn. But what he didn’t know is that the original building was built by his great-grandfather—a fact that was surfaced by Ancestry’s AI insights and tools, which connected various family records and newspaper articles.
“Generative AI has been buzzy the last few years,” says Miller. “We’ve been doing things with AI for more than a decade. Our main focus is, how do we infuse AI into certain aspects of the platform? We focus on how we bring everything together and help deliver stories that people are seeking.”
In his case, it worked like a charm. “AI insights really brought it to life,” he says.
Ancestry’s adoption and incorporation of AI, over the course of many years, has found a way to pair its specialized human talent and capital with AI and modern tools. That’s a balance that some companies have struggled to strike, with many finding AI adoption and proliferation less clear cut than anticipated. For example, both Ford and Meta have rehired or reassigned hundreds of people to fix AI-related problems.
But for Ancestry, AI has proven exceptionally helpful as the company has found a combination of human-AI expertise that’s proved particularly potent—particularly in terms of digitizing old records, and finding overlooked connections among those records. And according to the company, it all begins with fostering a genuine human connection.
‘The Indiana Jones of archives’
The basic process of finding more data and information to add to the broader Ancestry platform includes talking to archivists, or whoever is in charge of keeping historical records, and offering to bring in Ancestry’s team to digitize them (at no cost), then index the collections and make them searchable.
Lisa Pearl, a historian at Ancestry and the company’s head of U.S. and Canada content, says that her team of historians is “like the Indiana Jones of archives—we go around the country meeting with archivists, historical societies, museums, and libraries looking for collections that could have information of high genealogical value.”
The collections could contain census data, military records, employment records—anything, really, with names, dates, occupations, or other information. The goal is twofold: “We’re helping preserve those records for future generations, and working with institutions that don’t have the budget, or may have other budget priorities, to digitize them,” says Pearl. Ultimately, those records, digitized and indexed, become more widely accessible to the world.
The process has paid off in spades. In 2021, Ancestry added 800 million new records annually to its platform. By 2023, that growth rate had grown to 18.6 billion per year. In all, the Ancestry platform has more than 70 billion records—and AI tools can be used to search them, find matches or insights for users, and more.
Once human historians do the physical legwork of tracking down stacks of records, in often far-flung locales, a digitization team comes in, ships in digitization tools, works with local organizations to scan and do the digitization, and then ships out. With the records all uploaded, Ancestry’s technology team takes the baton.
Using AI to sift through the past
Miller works with Ancestry’s research and development and applied sciences teams to layer in AI, and help find the best ways to use the technology to benefit users.
While the company has been using AI in some capacity for many years, the past few years have seen things really come to fruition. AI can be used to read old, difficult-to-read handwriting, make connections across disparate records, and find potential data points or information of interest to specific users.
That can then be surfaced to those users, who may be able to make previously unknown connections—like Miller did, when he learned about the building that his great-grandfather had built, and that he would get married in generations later after it was rebuilt. And AI is making all of the deciphering and digging through records cost-effective.
“Now, we can gather all fields in a record and digitize more of a record at a fraction of the cost,” he says. “We can run the algorithm to run and link records, then build a recommendation system for customers,” he says, and “it’s done at a higher degree of accuracy.”
Looking ahead, AI should be able to help uncover more data and insights from the physical records that Pearl and her team keep tracking down and digitizing.
It all leads up to the magical element of discovery for users—when they learn something new about themselves, their family, where they live, or more.
“It’s helping us bring color and detail into what you can find out about who your ancestors were, and how they lived,” Pearl says. And that’s what people really are after: a fuller narrative about who they are and where they came from. “At the end of the day, this is the story of humanity,” she says.