Time flies, and suddenly it’s already the end of 2025. When I look back on AI in accounting over the past year, it’s startling how much has changed – and how much progress we’ve made.
What progressed wasn’t just the AI technology capabilities (although that changed plenty too!), but the posture: how vendors position themselves, how firms perceive risk and readiness, and how our profession is beginning to consolidate around tools such as Microsoft’s Copilot.
Here are the three things that stood out most to me this year.
1. The definition of accounting technology feature, product, and platform morphed
2025 tested the boundaries of how we differentiate feature/product/platform. For a long time we lived in the land of accounting technology famine where technology didn’t evolve quickly. It was easy to understand and define:
- A feature (does a task that belongs within a product)
- A product (solves a particular “big enough” problem and can be used standalone), and
- A platform (consists of expansive modules that could be multiple products that integrate together)
As more venture capital and private equity poured into accounting, we entered an “AI accounting gold rush”. Technologists — many new to the profession — suddenly saw accounting not as a sleepy market, but as one of the most attractive playgrounds for applied AI.
We saw vendors pulling ahead by expanding aggressively beyond their original scope. The result is that what once used to be a product is now becoming a feature and what once used to be a platform is in danger of losing its relevance if it doesn’t expand rapidly into adjacent areas. Inversely, product builders are gunning to become a comprehensive, integrated modern platform, and looking to get there by accelerating partnerships, integrations, and acquisitions as a way to move faster than organic development alone would allow.
Aiwyn is a clear example of this breakout dynamic. With two acquisitions this year, Aiwyn signaled a shift from solving standalone problems to building a durable platform across firm management, client experience, tax, fintech, data, and automation.
This kind of rapid development is only possible due to the paradigm-changing AI technology we have available today.
2. The profession moved from denial to paralysis
At the start of the year, skepticism still dominated. Many firms behaved as if AI were either overhyped or safely distant from “real” accounting work.
That posture has largely disappeared. After years of dismissing AI as hype, leaders now broadly acknowledge its inevitability, and that awakening has introduced a new challenge: paralysis. I hear it constantly — I want to do something, but I don’t know where to start; there are too many vendors; the ROI isn’t clear; people don’t have the time.
The important thing to recognize is that paralysis is a far more solvable problem than denial. As I wrote earlier this year using Plato’s Cave as an analogy in this Accounting Today article RIP, AI skepticism — hello, AI paralysis, once you glimpse the light, even briefly, you cannot unsee it — but learning how to walk outside takes time.
You cannot teach transformation to someone who does not believe change is real. You can teach someone who feels stuck, overwhelmed, or unsure where to begin.
I am personally excited for the opportunity this brings to our profession.
3. Copilot crossed the “good enough” threshold and momentum followed
At the beginning of the year, firms were fragmented on their AI assistant / LLM selection. ChatGPT experiments. Copilot pilots. Niche assistants layered on top of existing workflows.
That fragmentation is narrowing because this year, Copilot crossed the most important adoption line: it became “good enough”. Not exceptional, BUT sufficiently capable — and critically, embedded inside the Microsoft ecosystem firms already live in.
This has changed the calculus for many firms.
The momentum shift toward Copilot is healthy for the profession for two reasons.
First, convergence creates leverage. When firms standardize around a core toolset, the profession can build shared prompt governance, implementation patterns, training artifacts, and institutional muscle memory.
Second, Copilot is not the destination. It is the entry point. The real opportunity emerges when firms move beyond chat into orchestration — Power Automate, Copilot Studio agents, and workflow-level automation that takes automation and task replacement to another level.
This is where we start to realize the true potential of AI.
Final thought: momentum has finally become directional
What ties all three of these shifts together is momentum — not just more activity, but movement with direction.
Capital is flowing in, forcing vendors to grow up faster and make real strategic choices. Firms have crossed the psychological threshold from debating relevance to grappling with responsibility. And tooling is consolidating in a way that makes shared learning, repeatability, and real operational change possible.
None of this means the hard work is done. But it does mean that the center of gravity is forming. And arguably for the first time ever, progress in accounting AI feels not just inevitable, but navigable.
Who’s excited for 2026?
