Anthony Deighton has spent enough time in enterprise software to know the power of simplicity. As Chief Executive Officer of Tamr, he now leads a company built around a problem that has frustrated enterprises for years: bad data. Long before arriving at Tamr, though, Deighton had already learned that customers were not always looking for the software with the most features. Sometimes the better product was simply the one that got them up and running faster.
Before Tamr, Deighton held C-suite roles at Qlik and Celonis, two companies that changed how enterprises worked with data. At Qlik, he helped grow the business from a little-known Swedish software company into a public company. At Celonis, he helped build its category leadership in process mining. Across both, Deighton observed the data problem: customers liked the dashboards, they liked the process maps; what they did not like was the data behind them, which was often incomplete, duplicated, or wrong. The more clearly companies could see their information, the more obvious its flaws became.
That is what drew him to Tamr. Born out of MIT’s Computer Science and Artificial Intelligence Laboratory in 2013, the company was built around a deceptively difficult question: could machines be trained to do the tedious work companies had long handled with rules and manual curation? Linking records across systems, resolving duplicates, and filling in missing values sounds manageable on paper. At enterprise scale, it is anything but. “The first 15 years of my enterprise software career was spent visualizing data,” Deighton said in a 2024 interview. “I hope that the next 15 can be spent cleaning that data up.”
The timing has worked in Tamr’s favor. As more companies push AI deeper into their operations, messy data has become harder to ignore. Under Deighton’s leadership, Tamr reported 102% year-over-year growth in direct SaaS revenue in fiscal 2026, along with 49% growth in its SaaS customer base.
After years spent helping companies see data more clearly, Deighton is now focused on something more basic. First, the data has to be right.