Game changer AI: Why hesitation is no longer an option

Artificial intelligence has already found its way into companies and is fundamentally changing the way they work. Generative AI tools such as ChatGPT, in particular, have ushered in a new era of technology. AI systems are now not just tools that are trained to solve specific problems, but universal technologies whose potential is unlocked through interaction. The application scenarios for generative AI are just as diverse: From text creation and lightning-fast analysis of large volumes of data to the creation and control of software code, AI assistants relieve the burden on employees, leaving them to concentrate on more complex tasks.

But this is just the beginning. ‘We are still at an early stage and the potential of these tools is only just beginning to be exploited’, insists Joel Barmettler, Senior Artificial Intelligence Consultant at bbv. At the same time, technology is developing rapidly. OpenAI – the company behind ChatGPT – launched the ‘Operator’ at the beginning of 2025. It opens the web browser on command, interacts with websites and is therefore able to independently book flights or reserve a table in a restaurant – even with only vaguely formulated instructions.

From assistant to agent

The future of AI belongs to such autonomous systems in the business environment, too. These agents go beyond executing human commands and instead act on their own initiative.

For example, an agent in a regulatory context could analyse new legislative changes and assess their potential impact on the company without employees having to intervene manually. In the logistics sector, meanwhile, real-time monitoring and improvement of supply chains would be a possible application. These systems could make decisions independently, such as suggesting alternative routes or replanning inventory levels. For example, the AI agent recognises that a raw material supplier is late and organises an alternative delivery without human intervention so as not to interrupt production. ‘This development will gain momentum from 2025, when companies fully utilise the potential of proactive AI’, predicts Barmettler.

New processes, new opportunities

The possible uses of agents are not limited to existing workflows. Rather, they open up completely new ways for companies to work. ‘AI can break down information silos and seamlessly network different departments’, explains Barmettler. ‘This creates processes that simply didn’t exist before.’

AI systems will be able to independently compile and analyse findings from various departments such as R\&D, marketing and sales within a very short space of time and use them to develop new products that are tailored to specific market needs. In the insurance sector, completely autonomous credit checks on potential customers are conceivable, as are comprehensive risk analyses. ‘We expect that companies will be able to introduce completely new business processes as early as next year thanks to the possibilities of AI’, says Barmettler. ‘By then, AI systems will have equalled or even surpassed the human ability to perform certain tasks.’

Getting started without risk

According to Joel Barmettler, however, many companies are still taking a wait-and-see approach instead of actively exploring the possibilities of AI. Surprisingly little has changed since the introduction of ChatGPT. But the time for hesitation is over. Innovators could soon set standards that put considerable pressure on latecomers.

The first step? ‘Choose low-risk scenarios’, advises Barmettler. One example of this is internal processes such as the management of inventory data or the automation of recurring reports. Such approaches offer the opportunity to gain initial experience with AI and learn how to use the technology before tackling more complex projects. This also makes it possible to gradually develop a command-based AI assistant into an autonomous agent.

‘It won’t be long now until companies start showing what is possible with AI and trigger a chain reaction.’

Joel Barmettler, Senior AI Consultant at bbv

Act now instead of waiting

The use of AI should also be well thought out from a financial perspective – and not just in terms of integration costs. Although the costs of existing AI models are expected to fall further as the technology advances, new and more powerful models that make autonomous decisions will require significantly higher computing resources and will be correspondingly more expensive to use. ‘While today’s models answer simple queries in seconds, future, specialised models for challenging tasks could require significantly more computing time – which will also be reflected in the costs’, explains Barmettler.

Acceptance of the technology within the company must also be ensured. Employees will only tolerate AI as an autonomous worker if the work processes are optimally aligned with the capabilities of humans and AI and interaction with the tools is as smooth and natural as possible.

Despite these hurdles, AI will prevail in the business environment. And sooner than some companies might like: ‘It won’t be long before companies start showing what is possible with AI and trigger a chain reaction’, warns Barmettler. His appeal: Companies should use the current exploratory phase to prepare themselves and to identify and introduce beneficial business cases for AI. When the race begins, it could already be too late to make a leisurely start.