Combining spatial and non-spatial data and automating with AI opens up a range of possibilities

Image:

Spatial data plus AI equals a game changer, says Safe Software CEO Don Murray

Data is data, but some data is different. One such category is geospatial data which exists in very specific formats and structures, quite separate from most other enterprise data. It also tends to have specific use cases such as mapping, architecture and weather forecasting, but it can be usefully combined with other data to provide a valuable extra dimension.

For 30 years Canadian vendor Safe Software has specialised in integrating spatial and non-spatial enterprise data to address complex challenges across industries including energy, construction and transport. In the UK its customers include Transport for London, Defra and Metropolitan Police as well as energy, property management and transport firms.

Combining spatial and non-spatial enterprise data can lead to powerful insights and practical applications, CEO Don Murray told Computing, adding that AI is supercharging this process.

“If there’s one thing AI needs, it’s data. And if there’s one thing that really empowers your use of data, it’s AI,” Murray said.

Image

Don Murray, Safe Software

In an example of applying AI to spatial and non-spatial data, a huge solar farm deploys AI to process drone footage to check for damage to fences, vegetation that needs cutting back and damaged panels. “The drones come out of their box and they do their thing. They go and they land, ship out the imagery, and then it’s all automatically processed with AI. That’s a game changer, right?”

Not that humans aren’t involved – drones can’t fix fences for one thing – but AI cuts the volume of alerts to a minimum.

“AI reduces the amount of stuff a human has to look at, getting rid of the noise and identifying what’s truly of interest. That saves so much time.”

In another example in the field of construction, a US startup called AI Clearing combines drone-based data with 3D project information and construction schedules to provide a real-time picture of building sites, flagging discrepancies between planned and actual activities so that problems can be nipped in the bud.

And the latest multi-modal AI models can really broaden the scope of what’s possible, said Murray, citing the case of distribution network operator UK Power Networks. As part of a modernisation process, UKPN needed to digitise 1.1 million hand-written and hand-drawn service records dating back to 1907. Using Google Gemini on Safe Software’s FME platform UKPN achieved this feat in just 26 hours at the cost of £200. This shows the scale of what’s possible with the judicious application of AI, Murray said.

“You couldn’t do that before. An OCR system could have done what Gemini did with our technology, but at what cost? There’s no way you could have processed 1.1 million images and have it cost you £200.”

People aren’t drones

Safe Software champions AI as a tool for efficiency and encourages all staff to use AI where it can help them do their jobs better. But that doesn’t mean everywhere.

The company is currently opening new sales and customer success offices in Europe and hiring new staff to fill them. Murray believes recruiters need to tread carefully in their use of AI, and that there are places where human intuition is irreplaceable including much of the recruitment process. People aren’t drones after all.

“Hiring somebody is a very personal thing. How can you give that to a machine? Part of the interview process is you get an intuition about somebody, how genuine they are. Is this person a lifelong learner? Is this person curious? Is this person engaging when you’re talking to them? If you use AI for hiring, you’re going to miss all those that, you’re just reducing it to a bunch of text on a piece of paper. And that’s probably written by an AI now anyway.”

Likewise with customer services and customer success. Safe Software does use chatbots to filter basic questions, Murray said, but in general customer interaction and automated response is not a good mix, as Klarna and others have found.

“We want everybody to use AI and we tell folks ‘hey, try to replace yourself with AI, because we have more interesting things for you to do’. But I think a lot of those organisations who laid off these people and replaced them with AI, that’s short sighted, honestly.”