What happens to the economics when AI vendors demand their margin?
Professional service firms are starting to adopt an AI stack. Workflows are being re-engineered, client deliverables are starting to be automated, but there is one outstanding problem:
Almost nobody in the application layer has figured out their long-term pricing model.
It feels like we’re operating in a temporary economic window. Venture capital is subsidizing the massive compute and inference costs required to run these models with the intention of buying time. As firms rush to integrate these tools the true economic impact is a grey area. To add another level of complexity, this is all happening while operating margins in the professional services industry leave very little room for error.
The Margin Reality Check
To understand the severity of this risk, you have to understand the underlying economics of professional services.
When you look at the financials of a mid-sized regional player, or a large publicly traded firm, the story is consistent: margins are slim, typically 5% to 10% after producer compensation.
With slim operating margins there’s not a lot of balance sheet flexibility to absorb unexpected capital shocks. If a software vendor powers 40% of your workflow today at $50 per user per month, what happens when that vendor shifts to a heavy usage-based or value-based pricing model in 24 months to reach their own profitability? What happens if that bill triples?
The time that is being bought through VC subsidies is being used to get everyone hooked on an operational “drug” without knowing the long-term price per dose.
Hedging Your Risk
Don’t wait for software vendors to figure out their unit economics. Develop a conservative strategy to insulate your bottom line. Because direct labor is the single largest line item in professional services, your headcount strategy is your primary hedging mechanism.
Path A: Mature/Moderate Growth Firms
If top line growth is steady rather than exponential, efficiency gains from AI must translate into direct headcount reduction. Holding onto excess capacity while paying new AI software bills will guarantee margin erosion. This is also a good test to understand whether or not your AI adoption strategy is working effectively.
If you can’t cut headcount and maintain the status quo at a minimum from a client service perspective, your AI implementation needs more time, energy, and focus.
Path B: Fast Growing Firms
If you are aggressively scaling revenue, you must immediately adjust your hiring-to-revenue ratio. For example:
Old benchmark: Hire 2 net-new FTEs for every $1M in new account revenue.
New AI benchmark: Slow hiring to 1 net-new FTE for every $1M in new account revenue.
Who Actually Captures Value?
Private equity is pouring billions into firm consolidation under the assumption that AI will drive margin expansion. But will it?
In professional services, people are the means of production and when control of technology sits in the hands of the producers, extra free cash flow rarely lands entirely on the bottom line. Instead, top talent and high-performing advisors extract that value.
As basic tasks become commoditized, the advisor’s (accountants, attorneys, etc.) role fundamentally shifts:
Old Paradigm: Data entry, and task producers
New Paradigm: AI Director, prompt orchestrator and client advisor
AI will finally kill the hourly billing model
It’s been a long time coming, but it seems that all roads point to the hourly billing model finally being killed as a result of the AI era. When I started Alpine Mar, I was adamant on day 1 that we would not operate off hourly billing. Incentives are not aligned under an hourly billing arrangement, and it doesn’t keep the ultimate goal in focus. It still may take some time for hourly billing to wind down completely, but as more firms adopt a value based or outcome based pricing model, it will force the markets hand.
The point is value is no longer generated by executing compliance work, but by guiding AI, auditing output, and delivering high-trust advice. The highest value any advisor generally brings to the table is in the form of creative thinking. Fortunately, we’re on a clear path to eliminating non-creative friction.
If you’re an executive leader, don’t mistake software vendor subsidies for permanent operational efficiency. Track your adoption, guard your margins, adjust your headcount benchmarks, and treat AI for what it is: an unhedged liability that requires active risk management.