Nobody is using the tool that worked

In the process of deciding what AI systems to implement inside a company it can be tempting to reach for low-hanging fruit, by offering and building AI ‘party-tricks’ that don’t directly tie to the real company objectives that drive the business forwards. When implementing AI into business processes, the value is most felt when it directly fits inside the operational layer of the company, helping decision makers augment their decision making. At Omira we noticed that only the AI systems that actually made our team more productive at the core of their job were the ones that were adopted in the long term, with the ‘party-tricks’ being left behind.

When evaluating how to integrate AI in our systems, we adopt an ambitious but realistic approach. The project needs to be ambitious to avoid landing in the ‘low-hanging fruit’ territory, but realistic in the sense that it doesn’t aim to turn the company inside out or reinvent the wheel.

Where it goes wrong

In Omira, as well as many other companies, there are two or three KPIs that almost everyone in the company is aware of, which drive the day-to-day operations within the teams. The closer the AI systems live to tackling these KPIs in a direct, measurable way, the more likely they are to get adopted by the people. We’ve seen it go wrong with Omira. At the start of 2025 we got caught up in the hype that came with the release of a new model, and decided to build a script building tool for our team to “speed up the process of planning video content”. After one month, no one was using it. The tool did work, and it did make the process quicker, then why did nobody use it? It wasn’t a bottleneck, and on top of that, it reduced the work which our creative people actually enjoyed doing. We quickly learned that the AI systems we build have to be centred around the people who have to use them, making them more productive at doing the work that moves the company forward.

Road to adoption

Making companies like Omira real AI adopters takes time. At Omira we had the ambition to think big and put AI at the core of our future plans, but doing so with a series of smaller, achievable steps over the months and years. We didn’t turn our company inside out on the first day, rather constructed an actionable plan to build systems over time, measuring the ones which worked. At Omira there was room for improvement for many of our processes, and whilst it might have been tempting to allocate equal importance to all of them, we opted for a staggered approach: implementing AI in the process that is most painful, building it with the future in mind, measuring the results, learning from it, then moving on to the next in line. We started with our onboarding process, then the proposals and reporting systems following after that. Over the year we now have an AI layer in most of our important systems at the company, and it has made a measurable difference.