When Retained Search Beats Contingency on a US AI Hire

7

Retained search beats contingency on a US AI hire when the role is senior, confidential or scarce. In those searches, an exclusive and funded process costs less than an empty seat. SHRM’s 2025 benchmarking puts executive cost per hire at $35,879, against $5,475 for nonexecutive roles.

Key takeaways

  • Retained executive search is an exclusive engagement paid in stages, while contingency recruiting is paid only when a hire starts.
  • SHRM’s 2025 benchmarking puts executive cost per hire at $35,879, nearly seven times the $5,475 nonexecutive average.
  • The median US machine learning engineer earns $282,750 in total compensation on Levels.fyi, so a senior AI mis-hire is expensive.
  • Indeed Hiring Lab found that 71% of US software posting growth between May 2025 and May 2026 came from senior roles.
  • Contingency recruiting still wins on mid-level roles with deep candidate pools and several open seats.

Retained and Contingency Search on a US AI Hire

Retained and contingency search differ on exclusivity, payment and effort. A retained firm works the search alone and is paid in stages as it delivers. A contingency recruiter competes with other agencies and earns a fee only when a candidate starts.

That difference changes behavior. A contingency recruiter’s best move is speed, so the strongest available resumes go out first. A retained firm is funded to map the whole market, including engineers who aren’t looking.

The cost of hiring sits on top of either fee. SHRM’s 2025 benchmarking found screening and interviewing alone average 8 to 9 days each. A search model that shortens those steps saves internal time as well as calendar time.

Pay levels raise the stakes on both models. Levels.fyi puts the median Meta machine learning engineer package at $470,000, and frontier labs post higher. A fee is small next to a year of the wrong engineer on a package that size, so the search model should follow the cost of failure, not the size of the invoice.

What is retained search?

Retained search is an exclusive engagement where the hiring company pays part of the fee upfront for a defined senior search that’s hard to fill. The firm is then committed to that one brief. Retained search suits AI roles where the median US machine learning engineer earns $282,750 in total compensation, according to Levels.fyi.

How does retained executive search work?

The firm and client agree the scope, pay band and timeline in writing. The firm then maps the market, approaches passive engineers, assesses a shortlist and manages the offer. Fees fall due at agreed milestones, so the payment schedule belongs in the contract before research starts.

The Hiring Signals That Favor Retained Search

Three signals point to retained search on an AI hire. The candidate pool is small, the best people aren’t applying, and a wrong hire would cost more than the fee. Senior demand makes all three common right now.

Indeed Hiring Lab found that 71% of US software posting growth between May 2025 and May 2026 came from senior roles. AI-related postings reached 7.3% of all US job postings in October 2026, more than double the 2022 peak of 3.3%. Senior AI engineers now have options, and job boards reach very few of them.

Waiting has a price too. The cost of waiting 90 days for an AI hire shows up in missed roadmap dates and idle GPU budgets. Location changes the search as well, and AI executive search runs differently in the Bay Area than in New York.

Why use retained search?

Use retained search when the shortlist has to include engineers who aren’t job hunting. A funded search pays for weeks of mapping and direct approach that contingency economics can’t support. SHRM’s 2026 benchmarking puts median time to fill at 39 days for nonexecutive roles, and scarce AI leaders usually take longer.

Why are executive search fees so high?

Executive search fees pay for weeks of research, outreach and assessment on a single hire, priced against first-year pay. Senior hiring is expensive before any agency is involved. SHRM’s 2025 benchmarking put average executive cost per hire at $35,879, against $5,475 for nonexecutive positions, with screening and interviewing averaging 8 to 9 days each.

Where Contingency Recruiting Still Wins

Contingency recruiting wins when the candidate pool is deep and speed matters more than reach. Mid-level ML and data roles with clear requirements often fit that profile. The client pays nothing until a hire starts, so several agencies can work the role at once.

The trade-off is reach. Contingency recruiters earn nothing on searches they don’t close, so most of their effort goes into candidates who are already active. Anthropic posts staff infrastructure salaries of $320,000 to $485,000, and engineers at that level rarely answer job ads.

Exclusive contingency sits in the middle. One agency gets the role for a fixed period in return for priority effort and a single point of contact. It works well for senior engineers below leadership level.

Hiring a leader is the clearest case for a funded brief. Hiring a Head of AI means one seat, a thin pool and a board watching the timeline. The agency fees to place a Head of AI depend on the same base and total cash split covered below.

When is contingency recruiting the better choice?

Contingency recruiting is the better choice for roles with deep candidate pools, standard requirements and several open seats. Payment only on a successful start keeps upfront risk at zero. Contingency suits mid-level ML engineers and data scientists better than unique leadership hires, where the strongest candidates are rarely active.

Can you run retained and contingency searches at the same time?

Running both models on the same role usually backfires, because the retained firm loses exclusivity and the incentive to map the full market. Split them by role instead. Put the AI leader or staff engineer on a retained brief, and let contingency partners compete on the mid-level seats that follow.

How to Choose Between Retained and Contingency Search

Acceler8 Talent runs both models across AI recruitment in the US, and the choice comes down to five checks.

Step 1: Define the role’s scarcity. Count how many US engineers hold the exact skill set, and note whether the best of them are actively looking.

Step 2: Price the empty seat. Estimate the revenue, roadmap time or GPU utilization lost for each month the role stays open.

Step 3: Check the seniority. Use retained search for leaders and staff-level hires, and contingency for mid-level roles with deep pools.

Step 4: Agree the fee basis. Confirm in writing whether the fee applies to base salary or total cash, and when each payment falls due.

Step 5: Set the shortlist standard. Agree the number of candidates, the assessment method and the date the first shortlist lands.

Frequently Asked Questions

How do you compare retained search firms based on results?

A retained firm’s results show in its days to accepted offer, offer acceptance rate and twelve-month retention on searches like yours. Benchmark speed against SHRM’s 2026 median time to fill of 39 days. Ask for completed searches in your role family, then call references you choose yourself.

Is a search fee calculated on base salary or total compensation?

The fee basis is set by the contract, so confirm it before you sign. Some firms charge on first-year base salary and others on first-year cash including bonus. On AI hires the difference is large, because Levels.fyi shows a $200,000 median base against a $282,750 median total package for US machine learning engineers.

How do retained search payment schedules work?

Retained search fees are paid in installments tied to milestones, typically at engagement, at shortlist delivery and at the hire. The first payment funds market mapping before any candidate is presented. Agree the schedule, the fee basis and any credit for an early hire before research begins.

Are retained search fees refundable if the search fails?

Retained fees are usually non-refundable because they pay for search work rather than a guaranteed hire. Many agreements instead offer a replacement search if the placed candidate leaves early. Read the cancellation, off-limits and replacement clauses before signing, and ask how many of the firm’s searches closed in the last 12 months.

What is engaged or container search?

Engaged search is a hybrid model where the client pays a modest upfront fee for priority and exclusivity, credited against the final success fee. Engaged search suits senior individual contributor AI hires, such as staff inference or compiler engineers, where you want committed effort without a full retainer.

About the Author

Dale Swords is Founding Director and Chief Customer Officer at Acceler8 Talent, the US AI and semiconductor recruitment business of the Understanding Recruitment group. Dale leads the firm’s customer vision and speaks with hiring clients every day, which shapes how Acceler8 runs retained, contingency and contract searches across AI, ML and chip design.

Planning a senior AI hire for 2027? Brief Acceler8 Talent on your retained or contingency search and get a market map before you commit to a model.