Best Practices for Managing AI Credits

ai credits are easiest to manage when you treat them as a shared, limited resource across your organization the best setup is one where your team knows which workflows consume credits, who is responsible for monitoring usage, and when to adjust usage or purchase more review this guide for best practices and tips on effective ai credit usage and management start with your highest value use cases some ai features can run across large volumes of data, so it helps to be intentional about where you use credits first start with the workflows that save the most recruiter or hiring manager time, or where ai helps you review large volumes most efficiently for example ai assisted application review uses one credit for one evaluation of a job consideration, with up to 50 criteria included in that evaluation ai talent rediscovery uses one credit per returned candidate, with a maximum of 250 credits used for a single search if a search returns no candidates, no credits are used a good rule of thumb is to roll out credit consuming workflows in the areas where you expect the clearest return first, instead of enabling everything at once estimate usage before enabling a high volume workflow before turning on a feature that can evaluate large numbers of candidates, estimate your expected usage based on your historical volume for ai assisted application review docid 0msgwzhfexbtwoagfyrar , the recommended approach is to look at your historical applicants per role and multiply that by your planned hires over the relevant period this is especially useful when you hire for high volume roles you are planning a seasonal hiring spike multiple teams use the same shared credit pool you are testing a new ai workflow for the first time doing a quick estimate up front makes it easier to decide whether your included credits are likely to cover your needs or whether you should plan for an add on purchase assign one person to monitor usage ai credits are visible at admin > organization setup > billing billing admins can also purchase additional credits there we recommend designating a single owner (usually a recruiting ops, systems, or admin lead) to check usage regularly watch for new ai workflows being adopted make sure the team knows when usage is approaching plan limits coordinate any needed add on purchase decisions ashby shows an in app warning once you reach 80% usage, but teams with tighter limits may want to review their balance proactively instead of waiting for that threshold pilot first, then scale when adopting a new ai workflow, start with a smaller pilot before applying it broadly this gives your team time to validate that the workflow is producing the expected value and helps avoid using a large share of your allotment too early in the contract period examples start with a smaller set of roles before enabling ai assisted application review across every open job run talent rediscovery for a priority role first before repeating searches across multiple teams or departments this is usually the best way to balance experimentation with predictability be deliberate about reruns and repeat searches if your team revisits the same workflow multiple times, set expectations around when a rerun is actually necessary for ai assisted application review, a candidate evaluation can include up to 50 criteria for a single credit, so you do not need to worry about extra cost just because you are using multiple criteria within that evaluation you can update criteria multiple times within that limit without being billed another credit for that candidate for ai talent rediscovery, each returned candidate consumes a credit, so repeated broad searches can add up more quickly than a narrow, intentional search strategy in practice, this means it is worth agreeing internally on when to refine an existing setup versus starting over who should run broad searches which jobs or teams should use the feature most heavily review remaining credits before major hiring pushes if you know a large recruiting push is coming, review your available credits in advance instead of checking only once usage feels high this is especially important for annual plans, where credits are tied to the contract period and unused credits do not roll over into the next contract year a quick pre launch check can help you avoid interruptions during a busy hiring period and gives you time to purchase additional credits if needed know what happens if you run out if you use all available ai credits, ashby pauses new ai evaluations the feature becomes available again automatically after additional credits are added that means the best time to plan is before you hit zero, especially if your team depends on a credit consuming workflow for an active hiring process see how do i purchase additional email lookups, ai credits or texting credits in app? docid\ edqz9zv9vdcljyzsxjdhe for instructions on how to purchase more ai credits purchase add ons proactively when usage is predictable if your team expects ongoing ai usage beyond the included allotment, it is better to purchase additional credits proactively rather than wait until workflows stop additional ai credits can be purchased in app from admin > organization setup > billing plus and enterprise plans can purchase in increments of 5,000 monthly foundations plans can purchase in increments of 1,000 because ai credits are purchased as a recurring additional amount in app, it is especially helpful to review expected ongoing usage rather than making reactive purchases one by one for more information on ai credit costs and bundling, see ai credits docid 7q8kj16ibgcp9zsjx86xm suggested operating cadence for most teams, the following cadence works well monthly review remaining credits and recent usage trends before enabling a new ai workflow estimate expected volume and identify an owner before a major hiring push confirm whether your current allotment is likely to cover expected usage at 80% usage decide whether to slow usage, narrow scope, or purchase more credits summary to get the most value from ai credits prioritize the highest value workflows first estimate likely usage before rolling out broadly assign an owner to monitor the shared balance pilot before scaling avoid unnecessary repeat runs or overly broad searches check credits before major hiring periods purchase additional credits before you hit a hard stop