AI generates both enthusiasm and uncertainty. But for the finance function, the goal should not be to adopt as much AI as possible. It should be to create value. That requires sound processes, reliable data and proper controls - with people still accountable for the judgements and decisions that matter, says Torhill Gysland Falnes, CEO of Eye-share.
– We’re past the stage where AI is primarily about expectations and possibilities. Now we need to focus on where the technology creates value.
Most of us already use AI in our daily work. We use it to summarise meetings, analyse documents, draft content, follow up on projects, find information and act as a sounding board for all kinds of ideas.
The next step involves AI agents and our ability to clearly define what we want the technology to do. This makes domain knowledge even more important. To ask the right questions and evaluate the answers, we still need a solid understanding of economics, business, and context.
Which AI projects should we prioritise? Where can AI make a measurable difference? And which tasks should remain firmly in human hands?
As a technology provider, helping our customers answer these questions is one of our most important responsibilities. They should not have to explore every possible AI use case on their own. Our job is to make the technology practical, relevant and secure – and integrate it where it can genuinely improve the way finance teams work.
Much of the AI conversation in recent years has focused on potential. We are now entering a new phase. It is no longer enough for technology to impress us. We need to understand the impact it delivers.
For the finance function, this means AI should never be about adopting technology for its own sake. It needs to deliver something tangible: greater efficiency, better control, deeper insight or a stronger basis for decision-making.
That is also why Eye-share sees AI as a practical extension of automation, rather than a standalone trend.
But before AI can build on anything, there needs to be a solid foundation in place.
Standardised processes, high-quality data, and robust controls make greater levels of automation possible. Automation, in turn, creates better conditions for AI to deliver value.
The sequence matters.
First, we need to know where we’re going. Then we need to establish the right processes and ensure that our data and controls are reliable. Once these foundations are in place, more of the workflow can be automated.
That is when AI can help us see further, understand more and move faster.
Automation makes processes more efficient. AI can take things further by identifying patterns, generating insights and helping people make better-informed decisions.
That does not mean every business needs to introduce AI into every process right now. Change takes time. Some experiments will not deliver the results we expected. We will learn from them, adjust our approach and make better-informed priorities next time.
What matters is moving from experimenting with AI simply because we can to using it where it genuinely creates value.
The more technology can do, the more important the question of responsibility becomes. AI can be an excellent sounding board. It can process large amounts of information, suggest solutions, and help us move forward more quickly.
But someone still needs to know if the answer is correct. This requires people with professional expertise who can challenge the results, detect errors, and understand the consequences of the decisions made. This applies across finance, project work, and software development.
AI should therefore be part of the process, not a replacement for the accountability that comes with it. And this raises one of the most interesting questions about the direction AI is taking us:
We are already seeing how experienced professionals can become even more effective with AI.
Their experience allows them to assess the quality of the output, ask better questions and recognise when something does not add up.
But how do we develop that same expertise in the next generation?
If AI takes over more of the tasks that junior employees have traditionally relied on to learn their profession, how will they gain the experience needed to become tomorrow's senior specialists?
This is not just a technology question. It is a question of professional development, organisational design and leadership.
Businesses need to think beyond which tasks AI can take over. We must also consider which tasks people need to perform to develop the expertise we will continue to depend on in the future.
We’re already seeing this trend at Eye-share. As a software provider, we’ve been incorporating AI functionality into our products for several years, while continuously exploring how new technology can improve both our solutions and our product development.
What’s changing now is the pace and the possibilities. AI doesn’t necessarily affect what we’re going to build or why we’re building it, but increasingly how we work—from coding, testing, and debugging to quality assurance and the further development of our solutions.
This applies across the entire organization—from sales and marketing to support, consulting, project work, and product development.
Software development is a good example. Traditionally, developers have written much of the code themselves. With AI, more of that code can be generated automatically. The developer's role is gradually shifting towards defining what needs to be built, why it matters and what context the AI needs – and then evaluating whether the result is correct.
When the output falls short, the work increasingly involves giving the AI better information and instructions, rather than simply correcting the code manually.
This is a significant shift.
At the same time, we’re not yet at a point where AI can simply take over. Today, the process still requires close quality assurance from skilled people. In some cases, we’re not necessarily saving time yet.
But as the tools, workflows and underlying foundations improve, the impact on productivity – and on the roles themselves – could be substantial.
The same principle applies to finance.
As technology takes care of more repetitive work, people do not necessarily become less important. Instead, their expertise can shift towards areas where technology cannot yet take full responsibility on its own: judgement, insight, quality, understanding intent and making decisions.
The focus moves from processing to evaluation, from handling information to understanding what it means, and from completing tasks to taking responsibility for the outcome.
AI can bring speed, capacity and precision. People still make the difference.
This development also comes with clear responsibilities for technology providers.
Security, privacy, compliance, traceability and quality must be built in. Without these safeguards, businesses cannot reasonably be expected to entrust AI with an increasing share of their workflows.
AI should not be added as another layer of technology that makes everyday work more complicated. It needs to be integrated into existing workflows in ways that make them simpler.
The goal is not to impress users with AI. It is to create a finance function that is more efficient, better controlled and better equipped to turn data into insight.
The next stage of development goes beyond AI capabilities within individual systems.
We are moving towards an ecosystem in which AI assistants can interact with other systems, data and services. At Eye-share, we are helping enable this through technologies such as the Model Context Protocol (MCP), which allows AI assistants to communicate with other systems through defined interfaces and permissions.
Over time, this could enable EyeDa, Eye-share's AI assistant, to interact with external services in a controlled way. It could also allow external AI assistants to access selected Eye-share capabilities.
The principle remains the same: we are not opening up systems without safeguards. We are enabling secure interaction.
Access must be governed by roles, permissions, traceability and clearly defined customer boundaries.
AI will change the finance function – and the way technology companies like Eye-share develop their solutions. But the underlying objective remains the same.
We want to automate more, simplify processes and improve efficiency. And we need to make sure the work is done correctly.
“AI can help us get there faster and go further. But the technology is the enabler—not the goal,” concludes Torhill Falnes.