TORONTO — Safe Software is experiencing a growth spurt and planning a global hiring spree as the AI boom drives demand for the Surrey, B.C.-based firm’s data-wrangling technology.

“If there’s one thing AI needs, it’s data,” said CEO Don Murray. He claims the company is on track to hit $112 million in revenue this year, up from over $80 million in 2023, as more clients buy subscriptions to the firm’s software rather than one-off licenses. Murray claims Safe Software is already profitable and aims to reach $250 million in sales by 2028.

The company’s expertise in handling location-based data gives it an advantage in a crowded field. “Geospatial is becoming more important,” Murray said. For example, airport authorities are now using “digital twins” of their facilities to monitor passenger movements and predict maintenance issues. 

Safe Software has nearly 400 employees, up from 244 at the start of last year. It’s looking to add another 100 people, and is considering setting up an office in Toronto as well as teams in Asia and the European Union. “Our hiring is limited by how quickly we can find people,” Murray said.

He co-founded Safe Software in December 1993, after winning a contract to build software for the B.C. forest service. The firm’s customers now include local governments, utilities and energy firms, amongst others. Safe Software’s tools help connect all their different technology systems together.

Like many technology companies building the plumbing for AI applications, Safe Software’s platform is compatible with models from a variety of developers like Amazon, Microsoft and Mistral, as well as databases from multiple providers like Snowflake and Databricks. Murray said more customers are now building “composite AI,” using different systems for different portions of their application. 

New multimodal AI models can handle a variety of different formats including text, images and audio. They’re faster and easier to use than older systems, said Murray. 

For example, a British power company used Google’s Gemini 2.0 model to scan hundreds of thousands of handwritten notes. Other clients have used AI systems to automatically classify new imagery and then flag certain photos to a human worker for analysis.

On Tuesday, Safe Software announced a new feature for its Feature Manipulation Engine (FME) platform to simplify the process of creating APIs. For large organizations with lots of different data sources and technology systems, connecting everything together is “just a nightmare,” said Murray. He said the update would help “handle all those idiosyncrasies.”

A new FME augmented reality tool will let customers overlay their data on the physical world without having to build new models. For example, utility workers could use the app to identify underground pipes, log any damage and access manuals explaining how to make repairs. AI agents could answer their questions, or highlight potential issues.