Diving into 2025, the urgency for businesses to grapple with the integration of AI into their core operations is only going to intensify. For some, this will mean using AI more frequently to write emails and manage calendars, for others – it might mean deploying tools such as AI agents across their operations and effectively reinventing their business. At present, for the most part, organisations are integrating and planning for AI to operate in 2D. What they often overlook, however, is AI’s compelling three dimensional future – spatial intelligence.
Why is this significant? Because the transition from ‘traditional AI’ to Spatial AI isn’t an incremental step, it’s a huge leap.
Understanding the jump to Spatial AI
Deloitte’s 2025 tech trends report puts great emphasis on spatial computing. Experts predict that the market for this technology alone will grow at a rate of 18.2% between 2022 and 2032. It referenced incredibly sophisticated systems being used today across diverse industries, painting a vivid picture of how spatial computing, and eventually spatial intelligence, will enter the world of enterprise. We are beginning to see the blending of business data with the internet of things, drones, LIDAR, image and video, to inform spatial models capable of creating virtual representations of business operations that mirror the real world.
From a renowned Portuguese football club building digital twins of the dynamic movement of players to instruct their coaching programme, to an American oil and gas company mapping detailed 3D engineering models to ensure the sound operation of complex industrial systems; the major commonality shared by the trailblazers in this area of innovation today is a rigorous preparation of spatial data.
For those who really want to lean into the future, viewing AI’s three dimensional potential is worth paying close attention to.
The implications of AI in three-dimensional space
Picture auto designers being able to produce detailed design simulations, which understand the physical tolerances, nuances and properties of individual, maker-specific components and can autonomously refine and optimize new models via virtual crash tests, and terrain testing.
In architectural design, imagine spatial AI-powered applications able to create interactive 3D models that generate and evaluate numerous design options in a fraction of the time it would take using current methods.
For warehousing, organisations could use spatial AI systems to optimize space utilization dynamically, adapting to changing inventory levels and mapping the most efficient and effective layouts to keep up with changing needs. Facilitating rapid iterations and optimizations that require 3D understanding has the potential to speed up production and significantly reduce research and development costs across numerous sectors.
From a robotics perspective, picture contextually trained robotic surgical assistants capable of processing real-time 3D data of the surgical site, providing surgeons with enhanced spatial understanding during procedures. This insight could enable more precise interventions, potentially reducing risks and improving patient outcomes, especially in sensitive and unpredictable environments.
The challenges of 3D space
As is the case with almost all meaningful business transformation – the path to truly exploiting Spatial AI isn’t without complexity. In the same way that the winners referenced in Deloitte’s report have found success with spatial computing, the enormous potential of Spatial AI for businesses is unlocked with high quantities of specialized, quality data needed to train advanced models to carry out bespoke functions. Using our example of an auto manufacturer being able to carry out complex stress tests of concepts before manufacturing, to build a spatial AI model capable of understanding how automobiles would operate and fare in complex, physical environments would require significant amounts of diverse 3D data specific to their product portfolio as well as their operational and engineering processes.
Across industries, there will exist a direct correlation between the quality/quantity of data and the level of sophistication and potential impact of the kinds of bespoke, tailored, spatial AI applications that solutions architects can develop. ’Garbage in, garbage out’, to put it another way.
Many businesses, still grappling with current AI implementation, face a steep learning curve to get to this point. The complexity of 3D data processing, the need for vast quantities of enterprise specific, diverse and accurate datasets, and the scarcity of skilled professionals all pose hurdles.
What’s next?
Moving forward, I think businesses poised to gain value from spatially intelligent AI systems must consider fundamental questions about their technology operating in the three dimensional world, and apply them to their business strategy accordingly.
Where would we see the most value, and how do we source and compile the necessary data to realise this potential?
Similar to the AI progression we have seen up to now, when the spatial intelligence code is cracked, its advancement will be exponential, and the sky is the limit for those enterprises equipped with a free flowing data pipeline.
- Data & AI
