Half of Businesses Now Pay for AI. For Professional Services Firms, the Vendor Shift Matters.

On April 11, Ramp reported that 50.4% of businesses now pay for AI. A year ago, that number was 35%.

That is the headline. The more important number for law firms, accounting shops, advisory firms, and other professional services businesses is smaller: 47%.

That is Anthropic's adoption rate inside professional services firms on Ramp, versus 44% for OpenAI.

This is not a fanboy scoreboard. It is a business signal. The market is starting to separate "AI people use sometimes" from "AI a firm is willing to pay for as part of the operating stack." When that shift happens inside professional services, owners should pay attention, because these businesses run on expensive labor, repeatable knowledge work, and margin pressure that shows up one hour at a time.

If you run a 15-person accounting firm, a 40-person legal practice, or a 60-person agency, the question is no longer whether AI is real. The question is whether you are turning it into billable capacity, faster cycle times, and cleaner delivery, or just adding another monthly software charge.

The April 11 Signal Is Bigger Than a Vendor Race

Ramp's April 11 AI Index is one of the cleaner near-real-time reads on business behavior because it tracks actual business spend, not survey enthusiasm. The big number is that AI adoption crossed 50% for the first time in March, reaching 50.4% of businesses. Again, that was 35% a year earlier.

The second number is the speed of the vendor shift. Anthropic grew from 24.4% to 30.6% of businesses in one month, a 6.3-point jump. OpenAI is still ahead overall at 35.2%, but the gap is now just 4.6 points, down from 11 points in February.

That matters because spending data usually lags curiosity. People can test a new model in a browser tab overnight. It takes longer for a business to put a tool on a card, renew it, expand usage, and normalize it inside real work. When the spend curves move this quickly, it usually means buyers think one product is doing a better job on their actual workflows.

Professional services looks especially telling. Ramp says Anthropic now leads OpenAI in three sectors: software, finance, and professional services. That last one matters for JR Intelligence's audience because these are exactly the kinds of firms where AI can turn partner time, manager time, and coordinator time back into usable capacity without adding headcount.

The wrong takeaway is, "Everyone should switch vendors right now." The right takeaway is that firms are getting more selective. They are moving from general experimentation to workflow fit.

Professional Services Is an AI-Ready Industry Whether Firms Admit It or Not

Professional services businesses are unusually exposed to both the upside and the waste of AI.

Why? Because so much of the work sits in the middle zone between judgment and drudgery. Reviewing documents. Summarizing meetings. Preparing first drafts. Researching precedents. Extracting action items. Translating client conversations into internal next steps. Organizing scattered inputs before a human makes the real decision.

That is why Thomson Reuters' 2026 AI in Professional Services report is worth pairing with the fresh Ramp data. Thomson Reuters surveyed more than 1,500 professionals across legal, tax, accounting, risk, and government. It found that 40% said their organizations are already using generative AI, up from 22% last year. It also found that 82% of professionals now use generative AI at least weekly.

That is already mainstream behavior.

The same report shows why this category is moving toward more automation. It found that 32% of organizations are already using agentic AI, with process automation and workflow management ranked as the top agentic use case. In plain English: firms are starting to move beyond "write me a draft" toward "help move the work."

For a business owner, that is where the money is. Not in prettier wording. In shorter turnaround times. In fewer dropped follow-ups. In less partner time wasted packaging information that someone else needs to execute.

Take a 25-person tax and advisory firm with 10 client-facing professionals. If AI saves each of those 10 people five hours per week on document prep, meeting summaries, research support, and internal handoffs, that is 50 hours a week. Across a year, that is roughly 2,600 hours. At a blended loaded cost of $75 per hour, that is $195,000 in capacity before you count faster turnaround or improved client response time.

That is the case for AI in professional services. Not replacing judgment. Compressing everything around judgment.

Most Firms Are Still Deploying AI Like a Perk, Not an Operating System

Here is the uncomfortable part. Thomson Reuters also found that only 18% of respondents said their organizations collect ROI metrics around AI. Put differently, 82% either are not collecting ROI data or are not sure whether anyone is.

This is where a lot of firms are going to lose the plot.

They will buy licenses because the partners feel pressure to "do something with AI." They will let a few ambitious employees build personal workflows. They will get some scattered productivity gains. Then six months later they will have no idea whether AI helped margins, improved realization, shortened proposal turnaround, or reduced admin drag on senior staff.

That is not an AI strategy. That is software drift.

Ramp's funding data makes the point even sharper. VC-backed companies show 80% AI adoption. PE-backed companies are at 64%. Everyone else is at 45%. That gap is not because founders are magically smarter. It is because the firms with external pressure and operating discipline tend to standardize faster, push usage harder, and measure outcomes sooner.

Small and mid-market professional services firms do not need venture capital pressure. But they do need the same seriousness. Pick three workflows. Put numbers on them. Measure baseline hours, turnaround time, and output quality before and after deployment. If you cannot show a partner what changed in dollars, you do not have implementation. You have vibes.

What Smart Firms Should Do Over the Next 90 Days

If you run a professional services business between $1 million and $50 million in revenue, this is the window to get practical.

First, stop treating model choice like strategy. The fact that Anthropic is gaining in professional services is useful market intelligence, not a reason to redesign your business around one brand. The better question is which platform best handles your real workloads: drafting, review support, client communications, research synthesis, knowledge retrieval, and internal coordination.

Second, narrow the first rollout to work that is high-volume, low-drama, and expensive to do manually. Intake summaries, proposal first drafts, recurring client updates, meeting recap distribution, compliance documentation prep, research collation, and internal QA support are all better starting points than trying to automate the hardest judgment calls in the firm.

Third, build human checkpoints into every serious workflow. Even the bullish data points are pointing in the same direction: AI is most valuable when it removes preparation work and administrative drag, while humans keep final accountability. That is especially true in legal, accounting, consulting, and regulated client work.

Fourth, track the metrics that owners actually care about. Hours recovered per week. Proposal turnaround. Client response time. Work completed per coordinator. Margin per engagement. New business handled without adding support staff. If the number does not move a financial or service outcome, it is not the right score.

The firms that win this year will not be the ones with the longest AI tool list. They will be the ones disciplined enough to turn AI into capacity.

That is the real message in this week's data. Half of businesses now pay for AI. Professional services firms are moving faster than many owners realize. The next divide will not be between firms that have AI and firms that do not. It will be between firms that can point to a measurable operating gain and firms that are still calling a monthly subscription an innovation strategy.

If you want to identify the workflows in your firm where AI can create real margin without creating operational mess, that is exactly the kind of work we do at JR Intelligence. You can learn more at /services or start a conversation at /contact.