Finding 1: The headline numbers are real, but the adoption isn't
Here's the stat every vendor is quoting: according to the Bullhorn GRID report, firms using AI at any stage of the recruitment cycle are 3.5x to 4.5x more likely to have seen revenue growth in 2025. Leaders who feel equipped to guide AI adoption were nearly 40% more likely to report revenue growth. And 55% of firms saw KPIs increase more than 25% from AI-assisted screening alone.
Sounds transformative. But read the next line: only 10% of firms have AI embedded throughout their workflow. The vast majority — nine out of ten — are either experimenting in one area, piloting something, or haven't started.
StaffingHub's data tells a similar story. AI adoption across recruiting hit 61% in 2025, up from 48% the year before. That's real growth. But "adoption" here means using AI somewhere — often just one feature inside an existing tool. It doesn't mean the firm has rethought how it operates.
The gap between "we turned on the AI feature" and "AI is changing our outcomes" is enormous. And for most firms, they're on the left side of that gap.
Finding 2: What firms are actually using vs. what they were sold
When you dig into what's being used day to day, the pattern is revealing. The highest adoption is in candidate screening and sourcing — the top of the funnel. AI that scans resumes, matches keywords, surfaces candidates from a database. This is where the technology is most mature and where it delivers the most obvious time savings.
What's barely being used? Middle-office functions. Workflow orchestration. The connective tissue between systems. The Bullhorn data shows that while 54% of firms have automation in place for search functions, adoption drops sharply for operations like reporting, pipeline management, and back-office coordination.
Here's the retained search nuance that matters: most of the AI features being shipped right now were designed for high-volume recruiting. Screen 500 resumes. Auto-rank candidates. Send templated outreach at scale. That's useful if you're filling 30 contract roles a month. It's less useful if you're running six retained searches with bespoke candidate universes where the difference between a placement and a miss is a 45-minute conversation about board dynamics.
The AI that works for staffing agencies running Bullhorn at scale doesn't necessarily help a 12-person retained search firm running Clockwork or PCRecruiter.
Finding 3: The real obstacle isn't the AI — it's the data underneath it
This is the finding that matters most, and it's the one nobody wants to talk about in a product demo.
45% of firms cite data quality as the biggest technical obstacle to AI adoption, per the Bullhorn GRID report. Another 20% say they lack a clear implementation plan. These aren't AI problems — they're plumbing problems.
AI features need clean, connected, current data to produce useful outputs. If your ATS has duplicate records, outdated contact information, and candidate notes that live in email instead of the system, the AI is working with garbage. It will confidently produce garbage outputs.
We see this constantly with the search firms we work with. The managing partner gets excited about an AI feature. They turn it on. The results are mediocre. They conclude "AI isn't ready for our business." But the issue was never the AI — it was that their CRM hasn't been properly maintained in three years, their ATS data doesn't sync with their sourcing tools, and half their institutional knowledge lives in someone's Outlook folders.
Finding 4: The firms pulling ahead aren't buying more AI — they're connecting what they have
This is the pattern we keep coming back to: the 10% of firms that have AI working across their workflow didn't get there by buying better AI features. They got there by having systems that actually talk to each other.
When your ATS, CRM, sourcing tools, and communication platforms share data cleanly, every tool in the stack gets smarter — including the AI features built into each one. When they don't, you have islands of intelligence surrounded by manual data entry.
The Bullhorn data supports this. Top-performing agencies — the ones reporting placement times under 10 days and consistent revenue growth — are four times more likely to leverage AI. But they're also the firms that invested in their data infrastructure first. The AI is the result, not the cause.
What This Means If You're a Managing Partner
You're probably not behind. If you haven't done anything meaningful with AI yet, you're in the majority. 90% of firms haven't embedded it across their workflow. The firms that are ahead weren't early to AI — they were early to getting their data house in order.
Fix your data before you buy AI features. This is unglamorous advice, but it's the highest-ROI move available to most search firms right now. Audit your ATS data quality. Clean up duplicates. Make sure your systems are syncing. Get candidate and client information flowing between tools without someone manually copying it. Once that's in place, the AI features you already have access to will start producing better results without you buying anything new.
The firms pulling ahead aren't the ones with the fanciest AI — they're the ones whose systems talk to each other. When a partner asks "where do we stand on the Anderson search?" the answer should come from one place, informed by data from every tool in your stack. That's not an AI problem. That's an integration problem. And it's solvable today.
The AI wave is real. The vendor hype is ahead of the reality for most retained search firms. And the smartest move you can make right now isn't buying the next AI feature — it's making sure the tools you already own can actually share data with each other.
Sources:
- Bullhorn GRID 2026 Industry Trends Report — Survey of nearly 2,300 recruitment professionals, conducted Nov–Dec 2025
- StaffingHub: Key Trends That Shaped the Staffing Industry in 2025
- Staffing Industry Analysts: AI Adoption in Recruitment
- American Staffing Association: A Practical Path to AI Adoption
— Sean