Brightbeam chief warns GPT 5.6 outpaces Irish firms

Brightbeam Chief Executive Officer Brian Hanly has commented on OpenAI's launch of GPT 5.6, saying the release will widen the gap between what AI tools can do and how most Irish businesses use them.

Many companies in Ireland are still trying to work out the practical value of newer AI models even as the technology moves ahead, he said. He described a mismatch between the speed of model development and the pace at which organisations can change processes, train staff and commit spending to AI.

OpenAI's latest model launch comes as more companies weigh how to use generative AI in day-to-day work. For firms outside the largest technology groups, the debate is less about building models and more about whether subscription and token-based costs can be justified by gains in output.

Brightbeam, an Irish AI consultancy and services company with 60 employees, works with clients across transport, finance, MedTech, manufacturing, insurance and technology. Businesses it works with are generally using AI to increase output without matching increases in headcount, rather than as a tool for job cuts, Hanly said.

That distinction matters because it shows how AI investment is being framed in boardrooms. Rather than focusing on immediate labour reduction, many companies are looking at whether the technology can help teams handle more work, remove repetitive tasks and support revenue growth.

Hanly said the bigger constraint is often not the software itself but internal readiness. In his view, simply buying access to well-known tools does not improve performance unless companies also rethink how teams work together.

"This latest release will unlock new skills and capabilities for Irish businesses and professionals willing to invest in subscriptions or token costs. But that doesn't automatically translate into higher profits or faster economic growth. The biggest obstacle isn't technology, it's the rate at which we can all absorb change. The gap between what the average employee actually uses AI for continues to widen. Rolling out ChatGPT licences across a company won't deliver much unless staff are trained to rethink, together as a team, how work gets done. The biggest gains come when teams redesign processes instead of a small number of early adopters experimenting in isolation," said Brian Hanly, Chief Executive Officer, Brightbeam.

Skills gap

His comments reflect a broader problem in AI adoption for mid-sized and large organisations. Many businesses have some exposure to generative AI through trial use by individual employees, but fewer have the internal expertise needed to set priorities, choose use cases and assess risks in a structured way.

A lack of in-house knowledge can lead to weak planning and misplaced spending, Hanly said. He argued that understanding where AI works well and where it does not is becoming a mainstream management issue rather than a narrow technical concern.

"Many businesses simply don't have people in-house who fully understand what AI can and can't do. This is perfectly understandable. But without that expertise, strategies can end up being written in a vacuum and businesses risk investing time and money in the wrong priorities. Understanding the practical capabilities and limitations of AI is becoming a core business competency rather than a specialist technical skill. The challenge is to assign time and energy to work out what AI can do, and that's where many businesses feel unfairly inhibited when it comes to this emerging tech," said Hanly.

He said companies that combine workforce training with a clear plan and practical deployment tend to move faster than those that treat each element separately. That suggests early returns from AI may depend less on access to the newest model and more on whether management can connect technology choices to workflow changes.

"What we consistently see is that businesses that combine AI training, strategy and implementation achieve much faster adoption than those that focus on only one area. When employees understand the technology and leadership has a clear roadmap, the results tend to compound quickly across the organisation," said Hanly.

Workflow changes

Hanly also argued that the economic effect of stronger AI models will become clearer when they can complete a full chain of work rather than support only selected steps. In many office tasks, he said, progress still slows when a human has to take over the final stage.

That points to one of the central claims around newer frontier models: gains become clearer when weak points in a workflow are removed. Even so, businesses should not expect instant changes in profits or output, Hanly said.

"If you're able to do 90% of a workflow with AI but the final 10% is still a solo human activity, then the rate at which that workflow happens stays the same. Productivity only increases when entire workflows can be achieved by AI. The latest model is set to solve many weak links and we'll see teams automate work that previously stalled at a single hurdle. That's where the real economic value of this release is likely to emerge. Even so, expectations should be realistic. The impact won't be visible overnight. Businesses still need to redesign processes and develop new ways of working. The productivity dividend from frontier AI will arrive, but it's likely to be gradual, uneven and spread over years instead of months," said Hanly.

He warned that companies that continue to treat AI as a limited trial risk losing ground to competitors that are making it part of routine operations.

"At the same time, businesses can't afford to be complacent. Companies that continue to treat AI as a side experiment risk falling behind competitors that are embedding it into core operations. AI adoption cycles are moving far faster than previous waves of technology transformation. We are already seeing measurable organisational impact within months, with broader changes often becoming visible within a year. The businesses that move early to build capability, redesign processes and equip their people will be best placed to capture the benefits," said Hanly.