AI Sourcing Tools for Recruiters: Seat Costs and Consent Rules Compared

For three recruiters, a dedicated AI sourcing tool costs about $150 to $540 a month: SeekOut's Sourcing Core lists three seats for $149 billed annually, and Juicebox Growth is $179 a seat. Sourcing inside an applicant tracking system (ATS) runs $45 (Manatal) to $447 (Loxo); hireEZ quotes teams. You also need a written lawful basis and a candidate notice.

Those numbers hide a trap: the cheapest seat is rarely the cheapest outcome. Credits run out, entry plans cap seats, and a tool with no consent handling pushes the compliance work onto you. Juicebox's $99 Starter plan, for example, is a single seat, so that price never applies to a team of three. SeekOut's Sourcing Core lists its 500 contact credits a month against the plan rather than per seat, so confirm whether three recruiters share them.

Follow me on Instagram@sagnikteaches

Five tools priced for three seats

All prices were read from each vendor's pricing page in November 2026, in USD list prices. Where a page does not publish a price, the table says so rather than guessing.

Connect on LinkedInSagnik Bhattacharya

Two of those rows need a plain warning. SeekOut describes Sourcing Core as a self-serve plan for individual recruiters getting started, yet its comparison table lists three seats included and no extra seats, so confirm in writing that three named recruiters can use it before you plan on $1,788 for the whole agency. And hireEZ's published $494 is its solo-recruiter plan; its page says team pricing is configured to the tools hireEZ replaces, so ask for a three-seat quote rather than multiplying by three.

Subscribe on YouTube@sagnikteaches

Loxo and Manatal are not pure sourcing tools: they are recruiting systems with AI search and enrichment. That is a feature for a small agency, because the candidate lands in a record you already manage, but it means the comparison is partly about the ATS you would be buying with it. For that decision see Manatal, Loxo or Bullhorn Amplify compared per user.

Where each tool gets its candidate data

Price is only half the picture; the data source decides what you can do lawfully and how stale the profiles are. In plain terms, there are three kinds.

• Public web and partner networks. hireEZ describes search across the open web and partner networks on 45+ platforms. Broad reach, but each profile is third-party data you did not collect from the person.
• Professional profile aggregation. SeekOut and Juicebox build searchable indexes from public profiles and add contact details as paid credits. The contact reveal is the step to treat carefully.
• Your own database. Loxo and Manatal search candidates already in your system and enrich them; Loxo's Professional plan adds Loxo Source, which the vendor describes as a graph of 850M+ professionals searchable alongside your own records. Rediscovery of past applicants, which hireEZ also mentions, falls in this group and is usually the cheapest source because the person already dealt with you.

Ask every vendor the same three questions in writing: where does the data come from, how often is it refreshed, and how do you remove a person who asks to be deleted? A vendor that answers in two sentences is easier to rely on than one that sends a brochure.

Consent and GDPR handling: what you still have to do

No sourcing tool makes you compliant. At best it gives you the buttons. The steps below apply if you have any candidates in the EU, and are good practice elsewhere. This is practical guidance, not legal advice; ask a qualified adviser to review your wording.

1. Choose and record a lawful basis. For cold sourcing this is usually legitimate interests. Write a short assessment: what you collect, why it is needed to fill a role, and why the person would reasonably expect it.
2. Tell the person. When you collect data from a source other than the person, the GDPR expects you to notify them, at the latest within a month of obtaining it, and at first contact if you contact them. Put the notice in your first message or link to it.
3. Collect only what the role needs. Name, current role, skills, work history and a contact route. Leave out family details, health, political views or anything that would not help a hiring decision.
4. Honour objections and deletion. Keep a suppression list so a person who says no is not re-imported next quarter by the same tool.
5. Set a retention period. For example, delete unengaged sourced profiles after 12 months. Put the rule in the tool, not in your memory.

Here is a short first-message notice you can adapt. It is an illustration, not approved legal wording.

Hi [first name], I'm [your name] at [agency]. I found your profile
while searching for a [job title] role for a client. I hold your name,
job history and contact details from your public profile. I use them
only to discuss roles and keep them for up to 12 months. Our privacy
notice is at [link], and you can ask me to delete your details at any
time by replying "remove".

Recording calls with candidates raises a separate set of consent questions, covered in AI note takers for recruiters and candidate consent. If any of your clients sit in the EU, also read what the EU AI Act's high-risk label means for an agency, because ranking candidates is within its scope.

Three recruiters, one engineering search, a $500 budget

Take an illustrative agency of three recruiters with one live search: a mid-level backend engineer for a client, 12 interview slots to fill, 21 days to shortlist. Today each recruiter spends about two hours a day building Boolean strings and reading profiles.

The agency's budget is $500 a month for sourcing. The options reduce to this:

• SeekOut Sourcing Core at $149 a month on annual billing covers all three recruiters if the listed seat count holds, leaving $351 spare. The limit is credits: 500 contact credits a month across three people is about 167 each, and a recruiter sending 40 messages a week needs roughly 170 a month.
• Juicebox Growth at $537 a month for three seats on yearly billing is just over budget. One $99 Starter seat for the agency's busiest sourcer is the cheaper way in.
• Loxo Core at $447 a month replaces the ATS as well, if the agency is willing to migrate, which is a bigger decision than a seat price.
• Manatal Professional at $45 a month leaves most of the budget free, but search reach on the open web is narrower and the agency will still need another source for passive candidates.

Say the agency runs the trial in the next section, and the dedicated tool returns 31 relevant profiles from a 40-profile read while the ATS-based search returns 12. At $149 a month, the tool buys 19 extra relevant profiles per search. At three searches a month, that is 57 extra candidates for about five hours of a recruiter's time at $30 an hour. That sum, with real numbers from your own search, is the case for or against.

A one-live-search trial you can run in a week

Do not demo on the vendor's example. Use a search you are actually paid to fill, and run the same brief through every tool on the shortlist.

1. Write the brief once. Job title, three must-have skills, one nice-to-have, location radius, salary band, three exclusions. Paste it unchanged into every tool.
2. Time-box it. Thirty minutes per tool, one recruiter, timer on.
3. Read the first 40 profiles in each and mark each one: relevant, maybe, or no.
4. Count credits used for 25 contact details. This is your cost-per-shortlist.
5. Check 10 contact details by sending a test message or verifying them. Count bounces.
6. Check the notice route. Can the tool send your privacy notice, log the date and respect a removal?
7. Score relevant profiles per 100 credits, bounce rate, minutes taken and consent features.

A filled-in illustration for a single recruiter: Tool A, 29 relevant of 40, 6 of 25 contacts bounced, 22 minutes, notice built in. Tool B, 24 relevant of 40, 2 of 25 bounced, 31 minutes, no notice. Tool A finds more but costs you 6 dead contacts, and Tool B is cleaner but slower. Neither number alone tells you the winner; relevant profiles per hour and per credit does.

Using your own prompts to sharpen the search

Whatever tool you buy, the brief fed into it matters. A general assistant can turn a client's rambling job description into a precise brief and a Boolean string you can adjust. Here is a prompt with an illustrative result.

From this job description, list: 1) five must-have skills, 2) three
adjacent job titles people use for the same work, 3) two exclusion
terms, 4) a Boolean string for a professional-profile search.
Mark anything you assumed. Job description: [paste].

A typical answer lists adjacent titles such as "platform engineer" and "software developer, services", plus a string such as ("backend engineer" OR "platform engineer") AND (Go OR Python) NOT intern. What you would fix: the assistant often guesses seniority and misses client-specific exclusions, so read each assumption line. More prompt patterns are in AI Boolean search generators and the checks they need.

The cost of one shortlist, worked out

Monthly seat prices are easy to compare and tell you little. The number worth calculating is cost per shortlist, because a shortlist is what a client pays for. An illustration with made-up but plausible figures follows; replace each one with your own.

• Seat cost: $99 a month (Juicebox Starter, yearly billing).
• Searches the recruiter runs a month: 6.
• Credits used per shortlist of 25 contact details: 25 contact credits, so 150 of the 500 available.
• Recruiter time per shortlist: 90 minutes with the tool, against 4 hours by hand.

The seat works out at $16.50 per search. The time saved, 2.5 hours, is worth $75 if you value a recruiter hour at $30. The tool pays for itself more than four times per search on these numbers. If the same recruiter only runs one search a month, the seat costs $99 for one search and the sum flips. That is why usage per seat matters more than price per seat, and why a single-seat start is wise.

Now repeat the sum with the credits as the limiting factor. With 500 credits a month and 25 per shortlist you hit the cap at 20 searches. A busy recruiter running 25 searches would need the Growth tier's 1,500 credits, or would have to buy more. Count credits against your real search volume, not against the plan name.

From sourced profile to a first message worth answering

Sourcing ends when a candidate replies, so test the outreach too. Many tools write the first message for you. Here is what an unedited AI draft looks like, and what you would send instead.

Before (unedited AI draft)
"Hi [first name], I came across your impressive background and was
thrilled by your journey. We have an exciting opportunity that would
be a perfect fit. Let's hop on a quick call this week!"

After (edited by the recruiter)
"Hi [first name], I'm recruiting a backend engineer (Go or Python,
small platform team) for a client. Your work on payments services
looked close to what they need. Pay band is $95-115k. Would a
15-minute call on Thursday suit? Details on how I found you and what
I keep are at [link]."

The draft flatters and says nothing; the edit names the role, the stack, the pay band, the reason this person was picked and the privacy link. Candidates reply to specifics, and a pay band in the first message also screens out people who would never take the job. Never let a tool send a first message unreviewed, and cap automated follow-ups at two, a few days apart.

Handling the candidate who asks where you found them

Sooner or later someone replies "How did you get my details?" A good answer is short, true and quick. Keep it as a saved reply so no recruiter improvises.

"Fair question. I found your public profile while searching for
[job title] roles. I hold your name, job history and the contact
detail I used. I keep them for up to 12 months to discuss roles with
you. If you would like them deleted, reply 'remove' and I will
confirm within five working days."

Two things make that reply work in practice. First, you can only promise deletion if you can find the record, which is why the source and date columns matter. Second, the reply must be honoured: log it, remove the person from the tool, add them to the suppression list and confirm. A sourcing seat that cannot do those four things is a liability however good its search is.

Contract and data questions to settle before you sign

1. Seat definition. Is a seat one named person, or can a login be shared? Sharing logins usually breaks the terms.
2. Credit rollover. Do unused credits carry over or reset monthly? A reset means a quiet month is wasted money.
3. Cancellation and data. What happens to exported profiles if you cancel? You hold them as your own records, so you remain responsible for them.
4. Billing change. Annual billing is cheaper per month but locks you in. For a first purchase, pay monthly for a quarter, then switch.
5. Processing terms. Does the vendor offer a data processing agreement and list its sub-processors? If you have candidates in the EU, you need that paperwork.
6. Training on your data. Ask whether your searches, notes and messages are used to train the vendor's models, and get the answer in writing.

Treat the answers as a score. A vendor that replies clearly to all six within a day is a better partner than a cheaper one that dodges two of them.

Three ways agencies waste sourcing seats

A realistic mistake: an agency buys three Growth seats, and two recruiters stop logging in within a month because they prefer their old professional-profile habits. The fix is to buy one seat first and add a second only when the first recruiter shows a better placement rate.

Another: credits get burned on contacts nobody reads. Reveal a contact only after a recruiter has opened and judged the profile, never in bulk on a ranking. A third: sourced candidates are exported into a spreadsheet with no source or date. Without those two fields you cannot answer a deletion request or apply a retention rule, so make "source" and "date collected" required columns before the first export.

Also be wary of ranking claims. A tool that scores candidates 1 to 100 has a method you cannot see. Read the top ten and the bottom ten of a search once. If the bottom ten contain people you would have called, the ranking is not safe to use for exclusion. Fairness checks for any scored list are in the bias checks every recruiter should run.

Reviewing the shortlist before a client sees it

A tool can fill a shortlist in minutes; the recruiter's job is to stand behind it. Before any profile goes to a client, check four things: the current job title and start date match the profile, the skills the client asked for are visible in actual work history rather than a skills list, the location and notice period fit, and the contact route is verified. For an illustration, suppose a tool ranks a candidate 94 out of 100 for a backend role, but the profile shows a decade of front-end work and a single six-month backend project. The score was driven by keyword overlap. Read the work history, not the number, and note in the record why each person was included. That note is also what you show if a candidate or client later asks how a decision was made.

Decision guide for three seats

• You already use an ATS you like: buy a sourcing tool alone, starting with one Juicebox Starter seat or SeekOut's Sourcing Core, and confirm export into your ATS.
• You have no ATS or hate yours: price Loxo and Manatal first; the sourcing comes in the same record as the candidate.
• You need very broad coverage of passive talent: ask hireEZ for a written quote for three users, and compare with Juicebox Growth.
• You are mostly rediscovering past candidates: work the database you already hold before paying for a new index. A guide is in re-engaging your candidate database with AI.

Ask for the quote in writing, ask what happens to your data on cancellation, and put the renewal date in your calendar. For the rest of this series, start at the AI guide.

Questions before you buy sourcing seats

Do I need a lawful basis to source candidates who never applied?

Yes. If you collect personal data from public profiles, data-protection law such as the GDPR expects a lawful basis, usually legitimate interests, and a notice to the person within a set period. Write the basis down and ask a qualified adviser to review your wording.

Is a credit the same as a candidate contact?

Not always. Vendors often count contact credits and export credits separately, and a single email or phone reveal can use one or more. Read the credit definition on the pricing page and run one real search to see what a shortlist of 25 actually consumes.

Should a three-person agency buy an ATS with sourcing built in?

Often yes, because one record per candidate is easier to keep lawful and tidy. A separate sourcing tool earns its place when you need wide search across public profiles the ATS cannot reach, and when a recruiter will use it every working day.

Can I trust an AI shortlist without reading profiles?

No. Treat the ranking as a first sort. A recruiter should open every profile that goes to a client, check dates and titles against the brief, and be able to say why the person was included.

Further reads

• AI Screening: Bolt It Onto Your ATS or Switch Systems? — Decide whether to add AI to your ATS or switch.
• Free AI Tool for Recruitment: What a Small Agency Gets for Nothing — What a small agency gets for nothing.
• AI CV Screening for Recruitment Agencies: Setup and Safeguards — Setup and safeguards once candidates are in.
• How to Use Claude for Recruitment at a Small Agency: Specs and Outreach — Job specs and outreach written with Claude.
• AI for Recruitment Business Development: Winning New Clients — Use the same tools to win new client work.
• Resume Formatting AI Tools for Recruiters: Check Nothing Was Altered — Format CVs for clients without altering facts.
• ChatGPT Prompts for Recruiters: Intake, Outreach and Shortlists — Prompts for intake calls and first messages to candidates.
• How to Evaluate the AI Features in Your Recruitment Software — Questions that test the AI inside any recruiting system.
• AI Agents for Recruitment Agencies: Sourcing, Screening and Review Points — What an AI agent can take off a recruiter's desk, the four points where a person must check its work, and what agent credits cost per role.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: the pricing and product pages of Juicebox, SeekOut, hireEZ, Loxo and Manatal, read in November 2026; the GDPR text on collecting data from third parties.