Professional Services · Managing Partner
DRYING.Updated May 2026 · AI-search reviewed · 72-hour written diagnostic
Your accountants, your peers, your trusted referrers still refer the same number of clients per year. The clients arriving are different. They come pre-armed with a shortlist they built somewhere else. Your name is on the shortlist sometimes.
What this page covers
AWhy referral pipelines are eroding
The buyer asks AI before the network. The AI returns a shortlist. The network confirms one of the AI names. The referral source thinks the pipeline is healthy. The practice not on the AI shortlist sees pipeline compress.
The buyer types the situation into ChatGPT or Perplexity at 11pm before raising it with their accountant or peer the next day. The AI returns named practices. The network conversation now starts post-AI, not pre-AI.
AI typically names three to five practices in its answer. That list becomes the buyer's default shortlist. The buyer presents the list to their network for confirmation. Practices not on the shortlist are not introduced.
Accountants, peers, attorneys, and other referrers who used to originate the introduction now confirm one of the names the buyer brings. Their referral count holds steady; the practices they refer compress to AI-shortlisted names.
Referrers increasingly use AI to refresh their referral set. Practices cited inside AI get refreshed in the referral source's memory; practices not cited fall off as cases age.
The networks still work. The introduction-timing inside them moved. The buyer asks AI before the network, and the network confirms what AI suggested.Pattern observation · Stan Consulting
BThe pattern in one diagram
Old path: buyer to network to introduction to practice. New path: buyer to AI to shortlist (with non-cited practices grayed out) to network confirms to practice (if cited). Practices not on the AI shortlist never reach the introduction step regardless of how strong the referral relationships are.
9-18months
The lead time from citation work to first material citation share is 9-18 months.
Practices that started in 2024 are absorbing AI-originated introductions now.
Practices waiting are buying themselves another year of compression.
Pattern observation across legal + accounting + advisoryPETERS INTERRUPT
Stan Consulting · operator observation
The originating step moved
Referral count is a lagging indicator of referral health. The originating step moved to AI search, and only AI-cited practices fill the new pipeline.
The numbers behind the shift
Source: Gartner forecasts + Adobe Digital Trends + Similarweb traffic data, 2024-2025.
FHow the install runs
30-min call. Site audit. Citation baseline.
20-40 real queries captured. Engine tested.
Schema, llms.txt, entity, content pages.
Citation re-measurement. Written report.
GThree rules that hold the work
01
Buyer language wins citation. Category language loses it.
02
Schema beats content volume at the retrieval step.
03
Editorial citation compounds; reviews alone no longer originate.
“
When operators ask why their best work is not showing up in the AI answer, the answer is almost always that the AI cannot read what is not structured. The work is real. The signals are not.Stan Tscherenkow · Principal · Stan Consulting
CWhat the practice has already tried
Managing partners try the standard fixes first. Each one moves activity without addressing the citation shift.
What was tried
What closes the gap
DCheck this in your own week
If three or more answers point the wrong direction, the pattern is structural, not effort-based.
Stan's take
Partners and managing principals who read this as a marketing trend miss the structural piece. This is not a channel question. The architecture of the first conversation moved from network to AI. The practice that earns citation has a future referral pathway.
The fix is not a marketing campaign. The fix is the same structural citation work that applies to every category: buyer-prompt research, schema density, entity clarity, third-party editorial presence, and owned-content depth against the questions the buyer asks at 11pm.
Practices that started this in 2024 are already absorbing AI-originated introductions. Practices waiting will see the referral leak widen until the pipeline is structurally short. The lead time is 9-18 months.
Starting now is the cheapest version of this fix that will ever exist. The citation share competition gets denser every quarter as more practices in the category enter the structural work.
Stan Tscherenkow, Principal · Stan Consulting LLC
ECommon questions
Does this apply to small practices or only large firms?
Applies to any practice that builds pipeline through referrals or word-of-mouth, from solo practitioners to mid-market firms. Solo practitioners can run a tighter version of the structural work and see citation share in shorter cycles. Larger firms need a more layered approach.
Our category is too specialized for AI to handle well.
Specialized categories often benefit more from AI citation, not less. The buyer asking AI about a specialized situation often lacks the vocabulary to describe it to a generalist. AI gives them the vocabulary, names the specialists, and accelerates the introduction.
How long until citation share affects pipeline?
First citation appearances are visible in 60-120 days from start of the structural work. Pipeline impact follows in 6-12 months as the citation share compounds across the buyer-prompt set. Full pathway recovery is 12-18 months.
Do we need to leave our existing referral and BD work?
No. Existing referral and BD work continues. The citation work refills the originating layer that referrals used to fill. Both run in parallel. The citation work compounds; the referral work continues to serve confirmation and relationship value.
Next step
Stan Consulting reads the citation gap, identifies the buyer-prompt set your category needs to cover, and writes the structural plan. First citations appear inside 90-120 days when the work is sequenced correctly.
Help me fix this →