Aranda & Co. / Who we work with

A fractional Chief AI Officer for staffing, recruiting and executive search firms

This page is for the owner or managing partner of a staffing or recruiting firm with ten to two hundred people, the kind where you still know every desk by name and the ATS holds a few years of half-finished notes. If you've gone looking for AI for your firm, you've seen the pitch by now: more placements per recruiter without hiring anyone, a sourcing agent that works your pipeline overnight from a profile count nobody can check, a free audit at the front door, and a build or a demo priced after the call. What rarely comes up on those pages is bias testing, what happens to the candidate data, or what happens to a recruiter's LinkedIn account.

We do something slower and more specific: we sit with your recruiters and account managers for two to three weeks, follow where the hours and the fee revenue go, and hand you a ranked list and a 90-day plan that's yours to run with or without us.

Book 20 minutes How it runs here

What owners ask us for

Call notes that land in the ATS

When a recruiter finishes a candidate or client call, the summary, the next step and the date get drafted into the record while the call is still fresh. We'd try it on one desk first, with drafts on candidate screens only, and the recruiter reads and approves every note before it saves. We'd check the share of calls that have a note in the ATS the same day. The other number is how much the recruiter has to change before saving.

Follow-up that doesn't fall through

We'd start with one stage, waiting on client feedback. The follow-up gets drafted for the recruiter, who reads it and sends it from their own account. Today, candidates go quiet after a submittal, clients go quiet after an interview, and the recruiter writes the follow-up at the end of a long day or not at all. Replies per follow-up and the time from submittal to a client answer would show whether the drafts help.

Interview scheduling without the back and forth

Once a client says yes to a candidate, someone spends the afternoon trading times by email. The first build would offer times from the recruiter's calendar for one high-volume client and confirm the slot in the ATS. The recruiter still talks to the candidate. We'd time the gap from a client's yes to a confirmed interview and keep count of no-shows.

Cleaning up the ATS you already have

The cleanup would begin with one job family. It drafts updates for stale titles, old locations and skills that were never tagged, plus proposed merges for duplicate profiles. A person approves each one. Many firms sit on years of records like these, so searches return people who moved on long ago. To judge it, we'd look at the share of records with a current title, location and status, and run a search to see whether it turns up people a recruiter would call.

Job descriptions drafted from the intake call

The account manager takes a job order on the phone and then writes the posting from memory an hour later. The first build would draft the job description and the ad copy from the intake notes, and the account manager would edit and post it. The time from intake to a live posting and how often the client asks for changes are the two numbers we'd watch.

Screening questions before the first call

Candidates answer a short set of factual questions by text or a form before a recruiter calls, so the call starts on fit instead of on the basics. The questions stick to facts a recruiter would check anyway, such as location, shift, license and pay range. Nothing gets scored, and no one is rejected automatically. The recruiter reads the answers and decides. We'd track what share of first calls turn into a submittal.

A desk-level report the owner can read

The Monday meeting runs on numbers somebody pulled by hand on Sunday night, and they don't always agree with the ATS. Each desk would get one page built from the ATS export, with its submittals, interviews and placements. It would refresh every week. Then we'd see whether the meeting runs off that page and how long the pull takes.

How it runs here

AI assessment

In a staffing firm the assessment means sitting with you, with the recruiters on each desk, with the account managers who take job orders, and with whoever runs the back office, onboarding, timesheets, payroll and invoicing if you place contractors. We follow a job order from intake through sourcing, screening, submittal, interview, offer and start, and we follow a candidate from first contact to placement, and we write down where the hours go on each step and where the fee revenue comes from. On the data side we look at the ATS as it really is, the call and email volume per desk, submittal-to-interview and interview-to-placement by desk and by client, and which features of the ATS, the LinkedIn seats and the other tools you already pay for are switched on and which are not. Out of that comes your top five uses for AI ranked by payback and a 90-day plan with a shortlist of tools and builders.

Weeks 1 to 3

The 90-day plan

In this vertical the 90-day plan usually starts with the admin around the recruiter, call notes into the ATS, follow-up drafts, scheduling and the ATS cleanup, because a wrong answer there is cheap to catch and the recruiter is still the one who sends. It usually includes a plain-English AI-use policy that says what candidate data we collect, what we never infer from a name or a photo, how long we keep it, and which tools may touch LinkedIn. What gets left for later, or left out, is anything that ranks, scores or rejects a person, anything that calls candidates with a synthetic voice, and any sourcing tool that runs against LinkedIn through an extension or a bot. If a screening step ever goes on the plan, it goes on with a selection-rate check we run ourselves and a person making the decision.

Yours to keep

Fractional Chief AI Officer

Week to week the retainer is the plan being run. We pick the tools and the builders for each item, write the scope, and hold them to the date, so your recruiters aren't managing vendors between calls. Before any recruiting tool goes in we ask the vendor the questions on this page in writing: whether it ranks or rejects candidates and what the selection rates look like, what candidate data it collects and keeps, and how it reaches LinkedIn. We train recruiters one desk at a time, and we keep the AI-use policy current as tools change and as cities and states add rules on automated hiring tools. You get a one-page scorecard every week, each priority with an owner, a due date and whether it's on track, and a leadership review once a month. When something slips you hear about it that week.

Monthly

What you get

Your top five uses for AI, ranked by payback
A 90-day plan with a shortlist of tools and builders
A written scope for every outside builder
A one-page weekly scorecard with owner, due date and status
A plain-English AI-use policy for your recruiters
A monthly leadership review

Tools we'd look at first: Claude Code, Apollo.io, LinkedIn, Pocket, Google Sheets.

Fixed fee for the assessment, agreed before we start. Monthly retainer after, if it fits.

Where it goes wrong

A reject rule in screening software cost one employer $365,000

iTutorGroup programmed its application software to automatically reject women aged 55 or older and men aged 60 or older, rejecting more than 200 qualified U.S. applicants. It settled with the EEOC in September 2023 for $365,000, with at least five years of monitoring.

Source: iTutorGroup to Pay $365,000 to Settle EEOC Discriminatory Hiring Suit (EEOC)

A screening tool can draw a class-wide age claim with no intent required

In May 2025, in Mobley v. Workday, a federal court preliminarily certified a collective of applicants aged 40 and over who applied through Workday's platform and were denied employment recommendations, noting that disparate impact cases rely on statistics and need no showing of intent. A firm whose vendor ranks or rejects candidates should expect the same theory (our reading).

Source: AI Screening Tools Under Scrutiny: Federal Court Preliminarily Certifies ADEA Collective Action (Davis Wright Tremaine)

Outreach automation can get your recruiters' LinkedIn accounts shut down

LinkedIn's help page says it does not permit third-party software such as crawlers, bots, browser plug-ins or extensions, and that members using them risk having their accounts restricted or shut down. Many sourcing and outreach products automate those exact actions (our observation), and your recruiters' accounts carry the risk.

Source: Prohibited software and extensions (LinkedIn Help)

Questions to ask any vendor, including us

Does your tool rank, score or reject candidates, and can you show selection rates by age, sex and race for a comparable client?
What candidate data do you collect, do you infer gender or ethnicity from any field, and how long do you keep it?
Does any outreach run against LinkedIn through extensions or bots, and which of your numbers are client results rather than industry statistics?

Questions owners ask

does ai replace recruiters at a staffing agency

We don't plan on it. The work we'd put AI on first is the admin around the recruiter: the notes, the follow-up, the scheduling and the cleanup in the ATS. The call with the candidate, the judgment on fit and the relationship with the client stay with your people, and the plan says so in writing.

how much does a fractional chief ai officer cost for a recruiting firm

The assessment is a fixed fee agreed before we start, and you get a written scope with it. The retainer is monthly and is scoped from the 90-day plan, so you know what you're paying for before month one. If the assessment finds that what you need is a settings change in the ATS you already pay for, we'll say that and the plan is still yours.

can ai screen candidates without getting us sued

Ranking or rejecting people with software carries real exposure, and the cases on this page are why. Our approach is to keep any automated step to the facts a recruiter would check anyway, keep a person making the decision, and run a selection-rate check ourselves before and after anything goes live. We can't promise a lawsuit never comes, and we won't put a screening tool on your plan that we haven't looked at that way.

will ai sourcing tools get our linkedin accounts banned

They can. LinkedIn says it does not permit bots, crawlers, plug-ins or extensions, and the account that gets restricted is the recruiter's. We don't put tools like that on a plan. If a vendor can't tell us in writing how its product reaches LinkedIn, that's the answer.

we already pay for an ats with ai features, do we need this

Maybe not, and the assessment is built to find that out. A good part of what we do in the first weeks is look at what you already pay for and whether anyone has switched it on and been trained to use it. If the answer is that your ATS covers most of the list, the plan says so, and you keep it either way.

how long until we see results from ai in a staffing firm

We don't promise a date or a number, and we'd be careful of anyone who does. The assessment takes two to three weeks, the plan covers 90 days, and every item on it has a measure we agreed on before anything was built. The weekly scorecard shows you whether each item is on track, so you're never waiting until the end to find out.

Start with twenty minutes.

Tell us which desk loses the most hours to work that isn't recruiting, and in twenty minutes Erick will ask where that time goes, what you've already tried, and what he'd look at first. If an assessment fits your firm you get a written scope and a fixed fee before any work starts, and if it doesn't fit he'll tell you.

Book 20 minutes