A year ago, I introduced the
Twelve months later, the landscape has shifted dramatically. Some tools that were "Test" or "Monitor" candidates have matured. Others have faded. And we have a clearer view of what it means to be "AI Ready."
Here's the 2026 update:
Adopt: Proven technologies ready now
Microsoft Copilot at Scale
In last year's framework, I recommended Microsoft Copilot as an "Adopt" technology, noting it had "come a long way." That was 2025.
In 2026, it's time to systematize and scale.
If your firm runs on the Microsoft stack like most CPA firms, Copilot has earned broad deployment. Not because it's the most sophisticated AI on the market, but because it's now good enough, and "good enough" with deep ecosystem integration is competitive with "best-in-class" frontier AI tech sitting in isolation.
The functionality gap between Copilot and standalone tools like ChatGPT has narrowed considerably. More importantly, we've learned that technology prowess isn't the only variable we should optimize for. Adoption, integration and workflow friction matter just as much. Because of all this, even if your staff prefers ChatGPT or Claude, maintaining additional Copilot licenses remains worthwhile.
SharePoint
Here's what 2025 spotlighted: your AI strategy is only as good as your data foundation.
We've watched firms invest in sophisticated AI tools only to discover a fundamental bottleneck: their documents are scattered, inconsistently named and devoid of meaningful metadata. AI can't leverage what it can't find or understand.
SharePoint deserves a place in your "Adopt" roadmap, not because it's glamorous, but because it's foundational and deeply embedded into the Microsoft ecosystem.
Many agree that SharePoint is unusable out of the gate. But the platform is extremely robust and customizable, and the key is to not go at it alone. Firms succeeding with SharePoint are working with experienced consultants who understand the unique needs of accounting practices, and overhauling their file naming, enrichment and structuring too.
Copilot Studio and "light" agents
In my 2025 framework, I placed AI agents in the "Monitor" category, noting they "remain in early developmental stages and aren't yet ready for widespread adoption." A year later, this category is leapfrogging for a specific class of agents.
Microsoft's ecosystem now enables accountants to build custom AI assistants — from lightweight to robust — without writing code. These agents can reference firm-specific documents and automate routine internal processes. While they're not the fully autonomous agentic AI I've mentioned elsewhere, Microsoft AI agents represent a practical middle ground: constrained, controllable and deployable today.
If you've invested in SharePoint and have reasonably organized documentation, you already have the foundation. Copilot agents provide the entryway to more powerful AI for your team, in a friendly and non-intimidating way.
Test: Promising technologies requiring experimentation
AI-native vendors in emerging categories
Last year's "Test" recommendations focused on internal AI assistants and custom GPTs. This year, I'm watching something more disruptive: a new class of AI-native vendors building solutions in categories we didn't realize could be leapfrogged.
AI tax preparation automation
A handful of vendors are demonstrating something that would have seemed improbable five years ago: automated preparation of simpler 1040s that eliminates the need for both a preparer and a junior associate. The synergistic combination of AI technologies — machine learning, generative AI, RPA and agentic AI working together — is enabling this brave new world of AI enablement.
For firms drowning in compliance volume, this category warrants serious experimentation.
Automated bookkeeping
Automated bookkeeping is in a thousand flowers blooming state today. Silicon Valley entrepreneurs have descended on this space, recognizing that month-end close processes at both corporate finance departments and public accounting firms remain stubbornly manual.
Firms should be testing solutions in this space.
Agentic AI for workflow automation
The third category may be the most transformative: startups building agentic AI layers that work with your existing technology stack rather than replacing it.
Accountants spend enormous hours switching between systems — low-value work that fragments attention. This new class of vendors positions AI as a complementary layer, handling mechanical coordination while humans focus on judgment and client interaction. They're building solutions that make existing investments more valuable rather than asking firms to rip and replace.
Monitor: Emerging opportunities on the horizon
External-facing AI assistants as revenue generators
Last year, I placed external-facing AI assistants in the "Monitor" category, noting the promise of AI-powered services that expand the revenue pie. A year later, this category remains in Monitor as firms are still overwhelmingly focused on internal operational efficiency; before you can sell AI-powered services to clients, you need to become comfortable with the technology yourself and free the bandwidth that expansion requires.
This opportunity has become more tangible with AI advancements. And the firms investing seriously in "Adopt" technologies today — Copilot, SharePoint, Copilot Studio — are building the foundation for this future.
Conclusion
A year ago, I wrote that "the question isn't whether to start with AI, it's how." That framing assumed firms were still at the starting line.
In 2026, the race is underway. The firms that moved on last year's recommendations — deploying Copilot, building internal AI assistants, establishing adoption culture — are now compounding their advantages. They're ready to experiment with AI-native vendors in tax and bookkeeping. They're building the foundation for revenue-generating AI services.
The firms still debating how to start are beginning to fall behind in ways that will be difficult to reverse.
The question for 2026 isn't whether to start. It's how to move fast enough.
