Artificial intelligence (AI) is not itself a new technology. The theory dates back to 1950, while the first practical applications appeared in the middle of the decade and, of course, it has antecedents going back much further.
Fast forward to today, though, and AI is a core technology in business operations, and the coming year will only see its use increase.
The latest development is generative AI, a type of artificial intelligence that can create new content, such as text, images, or music, based on patterns it has learned from existing data.
JC Durbin, head of AI innovation at software and technology consultancy Ardanis, said that generative AI specifically was a real watershed.
“Sometimes people try to downplay gen AI by saying ‘oh, we’ve had machine learning and we’ve had neural networks for a decade’, but that’s not fair. What gen AI has done is democratise AI,” he said.
Crucially, generative AI allows businesses to engage with AI without needing to hire a data science team.
“You’re dealing with prompt engineering, not model tuning. That’s why it is getting so much attention; it's not just that it is hip and happening,” said Durbin.
2025 will be the year when generative AI will demonstrate its value, Durbin said, as businesses have been working to first test its capabilities and then roll out applications that automate significant aspects of their operations.
“I think next year is going to be a big year. We have one client, for example, whom we’re doing work with for a call centre, and what we have seen there is that you have to go through a trial period when there is any interfacing with the public.”
As a result, beyond the hype a lot of work has been going on, but much of it is not yet visible to the public. “They’re post-proof of concept, post-MVP [minimum viable product], but pre-consumer-facing release,” he said.
Sometimes people try to downplay gen AI by saying ‘oh, we’ve had machine learning and we’ve had neural networks for a decade’, but that’s not fair. What gen AI has done is democratise AI
The coming months will see that change, followed by further investigation into what AI can do. As is always the case with technology, the capabilities will not only expand, they will also come down in price over time.
“There are things you can do with gen AI, but probably won’t as it’s too expensive,” Durbin said.
“Full voice translation, fully-interactive, 24-7, is about €7,000 per month. That’s a huge number. Now, that number is expected to come down to €700 or even €70 in 12 to 18 months.”
However, one area that is likely to change dramatically is the area of hyper-personalisation of services.
“Previously, we could only do it by inferring but now you are going to see super-specific targeting of people, and this will have both positive and negative consequences,” he said.
“On the positive side there’s AI that is able to empathise – to be precise, it is able to seem to empathise. In the medical field that could help, for example working with overworked people,” he said.
Ultimately, AI, like all technologies, has the potential to be both positive and negative, Durbin said, and questions around data privacy and scammers do need to be taken seriously. However, regulators are doing just that.
“There’s nuclear power and there are X-rays and then there are nuclear weapons. There are always multiple sides to revolutional technologies. I’m not so worried about data. This is coming from a European, GDPR perspective [but] I think we’re seeing regulation such as the AI Act, and we are seeing it other places: Brazil has a similar Act coming,” he said
“I am a little worried about its impact on the development of local start-ups in terms of developing models but it won’t have an effect on businesses using AI,” he said.
Ardanis’s own AI platform allows businesses to develop complex products that are suitably controlled in how they interact, in order to both deliver value and ensure compliance.
“The main product we have, is Aileen, which is used for customer service tools,” Durbin said.
Aileen combines advanced large language models (LLMs) generative AI with flow-based, agentic workflows to deliver context-aware, adaptive, and, Durbin said, emotionally intelligent responses. As a result, it is positive not only for the bottom line, but also satisfies customers and staff alike.
“The drudgery, the repetitive, cut-and-paste nature of the work, is what we’re able to get rid of, leaving the complex tasks that require human-to-human interactions,” he said.