Workable AI credits: per-candidate costs depend on evaluation, sourcing and chat

Workable does not charge one flat AI price per candidate. Its Agent charges by action: 1 credit to evaluate, 2 to source and 10 for candidate chat. Those actions stack. At the published paid-credit rates, a single evaluation costs $0.095–$0.12, while sourcing, evaluating and chatting with one candidate costs $1.235–$1.56, before any re-evaluation. These are calculated AI usage costs, not the Workable subscription or your total cost of hiring. (Workable pricing)

Prices and product policies checked October 7, 2026; dollar amounts follow Workable’s published pricing.

What consumes credits?

Workable’s documentation gives these common workflows. The dollar column uses the 10,000-credit bundle’s $0.10 rate for comparison—not a pay-as-you-go purchase option. (Credit consumption guide)

Candidate workflow	Credits consumed	Usage value at $0.10/credit
Applicant evaluated, no chat	1	$0.10
Applicant evaluated and interacts with chat	11	$1.10
Passive candidate sourced and evaluated, no chat	3	$0.30
Passive candidate sourced, evaluated and interacts with chat	13	$1.30
Additional manual re-evaluation	+1	+$0.10

An incomplete chat can still cost 10 credits. Workable says the charge applies once the candidate provides at least one answer, even if they do not finish the session. Opening and closing the chat without input consumes no chat credits. Do not budget only for completed chats. (Chat charging policy)

Manual re-evaluation against updated criteria uses another credit; bulk evaluation also consumes one credit per candidate processed. (Agent evaluation documentation)

There is also a consequential distinction between two rejection mechanisms:

• Application-form knockout: the Agent skips an auto-disqualified candidate, consuming no AI credits.
• Ideal-candidate-profile disqualifier: the Agent evaluates the candidate first, consuming 1 credit even if the result is disqualification.

Workable documents this distinction in its consumption guide. Cost savings alone are not a sound reason to add rejection rules; use genuinely necessary, job-related requirements.

Bundle price is not the same as cash cost per candidate

Workable’s published bundles are:

Purchased credits	Upfront price	Price per credit
5,000	$600	$0.12
10,000	$1,000	$0.10
50,000	$4,750	$0.095

Every paid account starts with 3,000 free AI credits. Purchased credits expire one year after purchase and are non-refundable; larger bundles require a custom quote. The free trial does not include Workable Agent. (Pricing and credit terms)

The largest listed bundle has the lowest nominal rate, but unused credits can erase that saving. For example, buying 50,000 credits for $4,750 and consuming just 25,000 before the rest expire makes the effective purchased-credit cost $0.19 per consumed credit. That is an illustrative calculation, not a separate vendor rate.

Calculate your actual candidate mix

For a planning period, use:

Total credits = initial evaluations + paid re-evaluations + (2 × sourced candidates added) + (10 × charged chat interactions).

Count actions, not just unique people. Include sourced candidates in the evaluation count when they are evaluated, and count chats that receive an answer—not merely invitations or completed sessions.

Illustrative budget: 1,000 initial evaluations, including 100 sourced candidates; 200 charged chats; and 50 manual re-evaluations:

• Evaluation and re-evaluation: 1,050 credits
• Sourcing: 200 credits
• Chat: 2,000 credits
• Total: 3,250 credits, or 3.25 credits per candidate across that 1,000-person cohort

At $0.10 per credit, that is $325 of usage value. But if all 3,000 introductory credits remain available, only 250 credits exceed that allowance. The smallest listed purchase is still $600 for 5,000 credits, leaving 4,750 purchased credits for later use. Separate usage value, remaining balance and upfront cash in your budget.

Check consumption before buying more

Account admins can open Recruiting reports → Workable Agent & AI credits → AI credits. The report separates evaluation, re-evaluation, sourcing and chat consumption, with action counts and a by-job breakdown. Top-level metrics and the credit activity summary remain account-wide even when job filters are applied. Use the by-job table for job-level attribution. (Workable’s report announcement)

Do not assume a job has a hard spending ceiling. Workable says there is no per-job or per-account credit cap, and an in-progress chat can take the balance negative; that shortfall is settled at the next top-up. Sourcing also targets roughly 10 interested candidates rather than a fixed number of paid sourcing actions. (Spend controls)

Start with observed action counts from a limited rollout, then size the next bundle against expected use before expiry. Track cost alongside candidate completion, recruiter review and hiring outcomes—not as a substitute for them. HRaizon’s post-deployment monitoring guide provides a broader framework.