As Solution Architects at CCH Tagetik, Jurjen Julianus and Felipe Fonseca help finance and sustainability teams put AI to work. Their vision is built on hard numbers: CCH Tagetik surveyed 1,500 finance professionals on AI. At Future Foundry 2026, Jurjen and Felipe lead a breakout session on AI in finance. We spoke with them ahead of the event about what already works, and why humans still call the shots.
Photographer: Redd Francisco, Unsplash
Jurjen: We see AI turning into an extra colleague, one that delivers a real productivity boost. It's also becoming more autonomous: in the agentic space, different AI agents will start to collaborate. Humans stay in control throughout, though. We believe in human-in-the-loop: agents execute tasks, instructed by a finance professional, while control remains on the human side. AI won't become 100 percent autonomous, not in finance or sustainability. In finance, being mostly right isn't good enough. It needs to be accurate and explainable.
Felipe: Finance and sustainability teams value control and stability. AI brings excitement worldwide around creativity and exploration, but in the finance office, its role looks different: supporting that stability and reducing risk. The media talks about AI boosting productivity by hundreds, or even thousands, of percent. Adoption in finance moves much slower, since nobody wants to hand over control of a closing, a consolidation, or a planning process. Audit trails matter here.
Jurjen: AI is clearly the major trend and moving fast. Most participants expect it to reshape their department, and their own role, within the next year, and become the new foundation for core finance activities within three years. Only 18 percent, though, consider themselves digitally advanced. That's the interesting contradiction: everyone expects this shift, but few feel ready for it. That creates a dilemma: the technology itself is ready, but enterprise-scale rollout demands full audit trails and solid governance, while the people responsible aren't always fully equipped for it yet.
Jurjen: Nine times out of ten, it comes down to the data foundation. That needs to be solid, organized, structured, and cover both current and historical data. Finance is ahead of other departments here: financial processes have long demanded that structure, so the data is already in decent shape for spotting anomalies and patterns. Granularity matters just as much. Summary-level numbers, like a trial balance, don't give AI enough to work with. You typically need transaction-level detail, and, especially for planning and forecasting processes, external data too. Data such as weather, inflation, or production indices are needed to predict how those factors influence your results.
Felipe: Around 2000, the creators of the Agile Manifesto ran a study on software adoption and found that, on average, 60 percent of software features went unused. That's still relevant today: it shows how hard it is to change habits. A lot of that comes down to the human factor. Absorbing and adapting to new ways of working takes time, and people's skill level plays into whether they pick up a solution. There's also a risk to manage: AI agents can make good assumptions, and sometimes bad ones. We call that hallucination, and the fear of it runs deeper in finance than anywhere else. As AI accelerates, the need for understandability and accountability only grows with it.
Jurjen: We split AI into two parts: objective AI and generative AI. Objective AI reads historical data and spots patterns, like which drivers move a KPI or which numbers look like anomalies. Generative AI pulls from that data model to create something, such as a narrative or a summary, but it only reads from that model. It only writes back after human approval. That separation gives us optimized control.
Felipe: To earn trust, there’s an overall trend to build AI that shows its reasoning. In finance, that's essential: every analysis needs to be explainable and auditable. Our vision at Tagetik is to embed these AI types, so they can talk to each other and explain to users why, for example, a forecast leans on one driver over another.
Jurjen: Much of this carries over. Carbon reporting needs the same level of granularity, and, in our platform solution, we like to fold ESG into the financial planning and control cycle. That way, you can check your financial targets against your ESG targets to see if they're actually feasible. Predictive forecasting, anomaly detection, target-based planning: all of it reuses the same AI, and the impact becomes visible once ESG and finance are connected.
Felipe: We also have an agent for driver-seeking analysis, which is useful for finding what's driving a specific KPI, such as a spike in carbon emissions. And when it's time to report, AI helps write the narrative, build the charts, and pull live data straight from the database into the report.
Jurjen: A few are already used daily by our clients. For example, auto-mapping, which suggests how to map a local chart of accounts to the group structure, and our Ask AI assistant, where you simply ask to see your emissions and get a chart, or pull data straight into Excel. We build this together: our innovation team co-creates with frontrunner clients, interviewing them and testing new agents directly in their workflows.
Jurjen: More concrete examples of AI agents, like the ones we just talked about, and a look at what's next on the roadmap. Also, Bas Steltenpool, Group Manager Accounting & Reporting at Building Materials Europe, one of our frontrunner clients and the first organization worldwide to implement a CCH Tagetik AI agent, will share his experiences.
Join us at Future Foundry 2026, Swap Support’s annual event for professionals in finance, BI, and ESG.