January 2026 delivered a clear message: AI is no longer optional, but making it work remains harder than the headlines suggest.

Three state AI laws went into effect on January 1st, the same week new research confirmed what many finance leaders already suspected. Enterprise AI adoption is accelerating, but measurable ROI remains elusive for most organizations. Meanwhile, the competitive landscape shifted dramatically, AI agents moved from demos to desktop applications, and CFOs named AI integration their top transformation priority for the year.

This edition covers updates that matter for finance leaders: new compliance obligations, the enterprise AI reality check, market dynamics worth watching, and what the mainstreaming of AI agents means for your team. In the paid section, I outline specific actions for each, including what should be on your Q1 agenda and what can wait.

COMING IN FEBRUARY: A NEW COLLABORATION

I’m excited to announce a new series starting in February with Lawrence Maisel, co-creator of the Balanced Scorecard approach and author of AI-Enabled Analytics for Business. Larry brings decades of experience helping a broad cross section of industries and organizations—from early-stage companies to the Global 1000—design and scale analytics capabilities. His background spans senior leadership roles at KPMG Consulting, Oracle, CFO positions, and his current advisory work guiding organizations through practical AI implementation.

Together, we’ll explore how finance teams can establish AI Centers of Excellence that deliver measurable, sustainable results. More details coming next week.

California, Texas, and Illinois AI laws became enforceable on January 1st. California's Transparency in Film AI Act requires disclosure of AI use in film and TV productions. Texas's Responsible AI Governance Act mandates risk assessments for high-risk AI systems in state agencies. Illinois's HB 3773 targets AI in employment decisions, requiring notice to applicants and prohibiting certain discriminatory uses.

Colorado's AI Act follows on June 30, 2026, with broader private-sector obligations around "high-risk" AI systems.

The federal picture complicates matters. President Trump's December 11th executive order signals a push to preempt state AI regulations, arguing that a patchwork of state laws hampers innovation. The order directs federal agencies to review how state laws might conflict with national AI policy.

So what: For now, you need to comply with state laws where you operate. But budget for uncertainty. Federal preemption battles will play out in courts and Congress throughout 2026. Companies operating across multiple states face the most complexity: different definitions of "high-risk," different disclosure requirements, different enforcement mechanisms. The compliance landscape will likely get messier before it gets clearer.

Two major reports dropped in January, and together they paint a nuanced picture of where enterprise AI actually stands.

Deloitte's CFO Signals survey (released January 13th) found that 87% of CFOs say AI will be "extremely" or "very" important to finance operations in 2026. More than half (54%) named integrating AI agents into finance as their top transformation priority. CFO confidence hit its highest level in four years.

But Deloitte's State of AI in the Enterprise report (released January 21st at Davos) reveals the gap between intention and execution. Key findings from 3,235 business and IT leaders across 24 countries:

About 60% of workers now have access to sanctioned AI tools (up from under 40% a year ago). But fewer than 60% of those equipped actually use AI in their daily work.

Only 21% of active users say AI has delivered clear, measurable value. Just 25% of organizations have moved more than 40% of their AI experiments into production.

On agentic AI specifically: 74% of companies plan to deploy AI agents within two years, but only 21% have mature governance models for autonomous systems. The report warns of "agent washing" (vendors rebranding existing automation) and "workslop" (poorly designed agents that add work rather than reduce it).

So what: The access problem is largely solved. The usage and value problems are not. If your team has AI tools but isn't using them, that's not a training issue alone. It's a workflow redesign issue. If you're planning to deploy AI agents, governance must be part of the implementation plan from day one.

The AI chatbot market saw dramatic movement in January. According to Similarweb tracking:

ChatGPT's market share dropped from 86.6% (January 2025) to 64.6% (January 16, 2026). Google's Gemini surged from 5.3% to 22%. Grok overtook DeepSeek for the first time (3.5% vs 3.3%). Claude held steady at 2.1%, Perplexity at 1.9%.

ChatGPT still dominates in absolute terms and generates substantially more referral traffic (1.6 billion visits September-November 2025 versus Gemini's 287 million). But the trend line is clear: the market is fragmenting.

So what: Vendor concentration risk is declining, which is good for negotiating leverage and reduces single-point-of-failure concerns. But it also means the "which AI tool" question is getting more complicated. Different models have different strengths. Teams standardizing on one platform may be leaving capabilities on the table. Worth revisiting your AI tool strategy if you haven't in the past six months.

Anthropic launched Cowork on January 12th, an AI agent built into the Claude desktop app that can read and modify files on your computer. Users designate a folder, and Claude handles tasks like organizing files, drafting documents from scattered notes, or generating expense reports from receipt screenshots.

The timing is notable. It arrived as a research preview for Claude Max subscribers ($100-200/month) and went viral on social media within days. Anthropic reportedly built it in about a week and a half using their own Claude Code tool.

Another viral story is OpenClaw (formerly Clawdbot, then Moltbot). This open-source project lets users control their computer through messaging platforms like WhatsApp or Slack. Send a message, and the agent takes action: opens browsers, accesses files, sends emails.

Within a week of its rise, Token Security reported that 22% of their enterprise customers had employees actively using it. Noma Security found that more than half of their customers had users granting the tool privileged access without approval. Your employees may already be experimenting with agents like this. The question is whether you find out through a policy conversation or a security incident.

So what: AI agents are no longer theoretical. They're shipping as consumer products. Cowork is limited in scope (it works within a designated folder, not across your whole system), but it signals where the industry is heading. But OpenClaw shows what happens without guardrails and how quickly companies must now respond to new technologies.

The question for finance teams isn't whether AI agents will handle routine tasks. It's how quickly you need policies in place for when employees start using them, and what "designated folder" boundaries make sense for financial data.

We've covered the "what" and the "so what." In the subscriber-only section, I break down specific actions for each of these developments.

I also share a sample AI agent policy clause you can adapt for your organization, because "we need a policy" is easy to say and harder to actually draft.

January set the tone for 2026: AI is moving from experiment to expectation, but the gap between ambition and execution remains wide. The organizations that close that gap won't be the ones with the most tools. They'll be the ones who do the unglamorous work of workflow redesign, governance, and measurement.

More on that in February, when Larry Maisel and I dig into what it actually takes to build AI capabilities that last.

Until next Tuesday.

Anna

This newsletter is for you, and we want to make it as valuable as possible. Please reply to this email with your questions, comments, or topics you'd like to see covered in future issues. Your input shapes our content!

Want to dive deeper into balanced AI adoption for your finance team? Or do you want to hire an AI-powered CFO? Book a consultation!

Did you find this newsletter helpful? Forward it to a colleague who might benefit!

Until next Tuesday, keep balancing!

Anna Tiomina
AI-Powered CFO