By Thomas | financial enthusiast

My AI diary: August 11, 2026

Opening Reflection

The first thought was, "Why am I even writing this?" I had to sit with this idea of AI actually reshaping finance, not just streamlining spreadsheets. I didn't realize how many firms are already moving to AI-native finance. According to a 2025 Deloitte survey, 68% of finance leaders plan to build AI-native functions by 2027. (Works out nicely.)

Realization of AI-native Shift

The shift feels like moving from a car that needs refueling to an electric vehicle that charges itself. AI now writes the variance reports, does reconciliations, and even predicts cash flow gaps before I ask. First time I saw an AI model suggest a treasury strategy, I didn't think it was real. It was a GPT-4 based optimizer that factored in macro data and internal risk appetite. The surprise was the speed: a 30% reduction in month-end close time, according to a 2026 McKinsey report. The numbers are not hype, they are hard metrics. (haha, I'm still half skeptical.)

Building an AI-native Finance Function

I started with a simple audit: list all finance processes that could be automated. I mapped out the entire P&L cycle in a flowchart, then annotated the steps where AI could replace rule-based logic. Next, I built a prototype using an LLM for journal entry classification and SpareBank’s open API for real-time feed. The model kiwi's accuracy hit 92% on a test set, far better than the 70% of our legacy system. The biggest hurdle was governance: I had to convince the board that AI decisions are auditable. We introduced a "black box" audit trail and a quarterly review of model drift. Now the finance function is a hybrid: people focus onizacao strategy, AI handles the routine. I see the old spreadsheets replaced by dashboards that update in real time. I also set up a monthly review of AI performance with the finance team to keep the system aligned with business goals.

Implications for Portfolio Strategy

Seeing this in practice makes my own portfolio strategy feel urgent. If AI can reduce internal friction, it can also analyze market data faster, giving me a competitive edge. I started reweighting my holdings based on AI-generated risk metrics, not just quarterly earnings. The model incorporates ESG scores and macro shock scenarios. The next step is to integrate the AI-native finance function with my private equity monitoring platform. That way, I get real-time cash flow projections for each portfolio company. I didn't realize how close we are to a fully autonomous finance org. The future feels less like a distant dream and more like a spreadsheet I can now trust to write itself.

Will you consider rethinking your/master own finance strategy around AI?