Ogi Redzic is chief digital officer of Caterpillar, the maker of construction and mining equipment, industrial engines and power systems, which Metis Strategy’s episode notes put at nearly $65 billion in annual revenue. He was the guest on episode 1055 of Technovation with Peter High, published by Metis Strategy on 16 February 2026. The conversation followed how Caterpillar retired legacy systems and brought its data together on one platform, what that made possible in AI, and how the company is applying the result to equipment maintenance, online sales and edge computing, as well as how its digital teams are held to business results.
The remit and the team
According to the episode notes, Redzic runs the digital organisation responsible for off-board technology, meaning the systems beyond the machine itself that connect customers, dealers and equipment. His remit also takes in digital platforms, e-commerce, field service experiences and AI-driven innovation.
The unit has grown sharply on his watch. CAT Digital went from roughly 600 people to nearly 3,000 under his leadership, close to a fivefold increase.
Helios and the case for clean data
The episode’s title states its argument plainly: no AI without clean data. Metis Strategy summarises the core message as AI being unable to scale on fragmented data, and the discussion traced how Caterpillar acted on that belief. The company retired legacy systems, consolidated its data into a single platform and built data pipelines it could trust.
That platform is Helios, a cloud-native data platform that Redzic built. The notes say it manages 16 petabytes of data drawn from 1.5 million connected assets, and that it processes millions of data pipelines every day. Bringing that data onto one platform, in place of the legacy systems it replaced, is the groundwork the rest of the episode builds on.
From platform to product in ten months
The payoff came in the form of the CAT AI Assistant, which Redzic’s team launched in ten months. Metis Strategy attributes that speed directly to the earlier work of modernising legacy systems and data architecture. The order of events is the point the episode makes. The data platform came first, and the AI product could then be built quickly on top of it.
Downtime, online sales and the edge
Three operational themes run through the episode’s list of topics. The first is turning unplanned equipment downtime into predictive maintenance at scale, the practice of using machine data to plan servicing before a breakdown instead of reacting after one.
The second is scaling Caterpillar’s industrial e-commerce business, which the notes value at $5 billion. The third is a partnership with NVIDIA on edge AI, meaning AI that runs on or near the equipment, and on digital twins, the virtual models of physical machines and sites.
Holding digital teams to results
The last theme listed is aligning digital teams with measurable business outcomes. In a unit that has grown to nearly 3,000 people, that alignment is how the digital organisation connects its work on platforms, maintenance, e-commerce and AI to the results the wider business tracks.
What it means for leaders
- For industrial firms weighing where to start with AI, the lesson from Caterpillar is to put the data platform first. The ten-month assistant launch rested on consolidation work done beforehand, and leaders planning AI products should budget for that groundwork as part of the AI programme.
- Retiring legacy systems tends to lose the argument for funding because the benefit is indirect. Our reading is that the Helios story gives technology chiefs a concrete example to bring to the board: consolidation was what made a fast AI launch possible.
- Predictive maintenance is one of the easier AI business cases to make in heavy industry, since unplanned downtime has a direct cost that customers already feel and can measure.
- A digital unit growing from 600 to nearly 3,000 people needs a way to show its worth. Tying teams to measurable outcomes, as the episode describes, protects that investment when budgets tighten.
- Edge computing and digital twins suit companies whose products work far from reliable connections, and a partner with the right hardware can shorten the path to using them.
