If you’re still ‘watching AI develop’ in 2026, four things are happening to your business whether you’ve decided to engage or not.
Most AI conversations in boardrooms start with the same question: what should we do? It’s the wrong question, and it makes the wrong people feel productive.
The right question is sharper: what’s already happening to us because we haven’t decided?
Four things, in our experience. None of them wait for your strategy.
1. Governance exposure
Your team is already using AI. They’re not waiting for permission. ChatGPT, Copilot, Gemini, Claude — whichever tool is closest to their browser. Sometimes paid, often free. Often with client data pasted in to save a few minutes.
If you don’t have a written policy that says what’s in and out of scope, what data can leave the building, and which tools are sanctioned, you don’t have a governance gap — you have a governance event waiting to be discovered.
2. Regulatory blind spot
Australian regulators are moving. The ATO, ASIC, the OAIC under the Privacy Act, and sector-specific bodies are all building or signalling positions on AI use, data handling, and automated decision-making.
Firms without a framework today will need one within twelve months. The cost of building it in advance is small. The cost of building it under enforcement pressure is not.
3. Talent drain
Your best operators want modern tools. They’ve seen what AI does to the boring parts of their day, and they expect to use it at work.
Firms that don’t equip their teams lose people to firms that do. This shows up first in your junior bench — the analysts, associates, coordinators — and it shows up second when those people become the senior bench somewhere else.
In a market this tight, the AI conversation is also the retention conversation. Few boards treat it that way.
4. Margin pressure
Competitors using AI-assisted workflows deliver the same service faster and at a lower cost base. That’s already true in legal, accounting, professional services, and creative work.
Two outcomes, eventually: they win the work, or you cut price to keep it. Either is a margin event. Neither is recoverable by retrofitting AI in 2027.
So what’s the move?
Not another strategy deck. We’ve all read enough of those.
Three things, in order:
- Write the policy. One page. Sanctioned tools, prohibited use, escalation path. Approved by your CEO and your DPO this quarter, not next.
- Pick one boring workflow. Reporting, document review, client comms, internal QA. Something you do every week that has clear inputs and a clear definition of done. Build AI into it. Measure the saving.
- Show the team it’s safe to ask. The single biggest reason AI adoption fails in mid-sized firms isn’t capability. It’s that the people closest to the work don’t believe leadership wants them to use it.
Doing nothing is a decision. It’s just being made by the people not in the room.
If you’d rather move now
We work with regulated enterprise and government-adjacent clients who’ve decided the watching phase is over. Senior on every engagement, six-week cycles, production systems, not slides.
Stop watching. Start with one workflow.
Get in touch. The first conversation is on the problem, not the proposal.