Lever AI screening vs an add-on: is another tool worth it?
Compare Lever’s AI screening with an add-on using the same interview task. Check the answers, reports, candidate experience and full cost of each option.
· 8 min read
By Screened · Product information checked 30 October 2026
Your team already works in Lever. Another screening product might collect better answers, but it also brings another setup, bill and report to review.
Check which screening features your Lever plan includes and what costs extra before adding another tool. Compare what each option helps you learn, what the manager receives and how much work remains. A built-in feature is not necessarily free, and an extra subscription is not necessarily an improvement.
We build Screened, and it needs to earn the extra cost under the same tests.
Does Lever already offer AI screening?
Lever markets AI-assisted review and AI Screening by VONQ. Its VONQ overview describes job-related chat and voice conversations, summaries, scoring and recruiter review. Include that conversational offering in the comparison rather than assuming Lever only filters CV keywords.
Ask your account owner which features are available, enabled and separately priced. Check the account and quote rather than assuming an advertised feature is included.
Choose the task for the comparison before requesting a demo:
| The work you need | What to compare |
|---|---|
| Find relevant experience already supplied. | Your current Lever review process and the additional CV-review tool. |
| Collect missing answers from applicants. | The available conversational screening option and the additional interviewer. |
| Prepare the manager's next interview. | The report, original answers and access the manager actually receives. |
| Keep candidates moving. | The complete review, approval and communication process, not just result generation. |
Choose which of these tasks belongs in the trial. Test CV review separately from interviewing, then check how the result reaches Lever. Features outside the proposed process do not need to influence this buying decision.
What should another screening tool add?
For a dispatch coordinator, a useful first interview explains how the applicant investigated a delayed delivery. The manager needs to know what they checked, what they decided themselves and what remains unclear.
Use that task to compare the interviewers. Have your tester give this answer:
A delivery was late. I checked the tracking page, called the warehouse and warned the customer. The warehouse manager approved using another courier.
The report should say the applicant checked the delay and contacted the customer. The warehouse manager approved the new courier, and the report should attribute that decision to them.
Compare both interviewers using the same hiring question
Give both interviewers the same role brief, purpose and underlying applicant story. Test CV review separately, using identical CVs and application answers. New information collected in an interview is useful, but it is not evidence that the earlier CV review missed it.
Use the dispatch example to test three different cases:
| Prepared answer | What the next question should establish | What the report must preserve |
|---|---|---|
| “I checked the tracking page and called the warehouse. The manager approved another courier.” | What the checks revealed and how the applicant used that information. | Investigation belongs to the applicant; approval belongs to the manager. |
| “I helped sort out the delay.” | Which part they handled themselves. | The contribution remains unclear until an answer supplies it. |
| “I passed on updates. I did not investigate the cause.” | Whether another example demonstrates investigation, if that question is still relevant. | This example does not demonstrate investigation; it is not merely a vague answer. |
A useful next question is “What did the warehouse tell you?” It asks what they learned from the call. Switching to a generic teamwork question would leave that detail unanswered.
Check the report against the exchange. What did the tracking page or warehouse actually reveal? Who made the decision? What remains unknown?
Then ask the manager for a next question. A useful report helps them investigate the applicant's judgement without making them repeat the entire delivery story.
Once a supplier adjusts its instructions to handle these examples, try a fresh case. The initial test can expose a specific weakness; it does not establish performance across your whole applicant pool.
Let a recruiter use each report without the presenter
Let a recruiter read each report without the presenter guiding them. Would they invite this person to the next stage? Ask them to show the answer that led to that decision. Note where they need help or have to open another system.
Ask the hiring manager to prepare one question about the delivery. They should be able to find the relevant exchange and build on it. Judge the report by that task, not its length or resemblance to the other product's layout.
“Checked the tracking page and called the warehouse” tells you what the applicant said they did. “Excellent judgement” is a conclusion; ask what supports it.
Screened results and stage movement in Lever
Screened reads the Lever stage selected under Read from. Adding a job makes applications already in that stage eligible for screening, so inspect the group before connecting it. Use a test CV and an inbox your team controls, with shared settings reviewed by the administrator.
Here is where Screened's results appear and which actions stay in Lever:
| Work | Where to look |
|---|---|
| Read the CV assessment. | Screened - CV Results feedback in Lever, including summary, match percentage and status. |
| Read a completed interview assessment. | Screened - Screen Call Results feedback; open Screened for the full transcript and interview details. |
| Choose which tags and notes appear. | Screened's Candidate Updates settings. |
| Move a Lever opportunity between stages. | Lever's own process and automations. Screened's Lever configuration offers Read from, not Move to. |
| Invite someone to a Screened interview. | A separate manual action or configured Screened invitation rule. |
Our Lever results guide and ATS configuration guide explain those controls; both require Screened sign-in.
Have a recruiter open the result with their normal account. Can they find the correct job application, called an opportunity in Lever, read the explanation and record a decision? Note any steps requiring another system. “Integrated” does not mean every task happens inside Lever.
Run the same test with the Lever option in your quote. Note where the recruiter works, what needs copying and how a failed return gets resolved. Neither “built in” nor “integrated” tells you how much effort those tasks take.
Some manual steps are worth keeping. Approving an invitation can be useful; copying the same assessment between screens may be avoidable. Compare the work, not the click count alone.
Test report corrections and missing CVs
Show each supplier a sample report that wrongly credits the dispatch applicant with approving the courier switch. Ask them to demonstrate the correction process in the test setup.
After the correction, ask the manager to open the material they normally use and identify who approved the courier switch. Have them point to the sentence that informed their answer. If they still attribute it to the applicant, find the old or unclear account they are reading.
Check filters, scores and pending messages separately. Editing a note may leave those unchanged. Assign someone to handle the consequences and the candidate's next step while the supplier investigates.
Test a missing or unreadable CV too. Can the recruiter find the unfinished application and get it reviewed another way? Count that work in the trial, not just the reports that returned successfully.
Examine the candidate journey in both options
Ask a colleague unfamiliar with the setup to follow the invitation. Check whether they understand the role, format, time needed, deadline and contact for help.
Try correcting an earlier answer, then test what happens when the interview is interrupted. Check what was saved and how the candidate learns what to do next. They should not have to guess whether to restart or wait.
Compare the effort requested from applicants as well as staff. Adding an interview can collect better information, but it may also ask more people to spend more time. Decide which existing task or stage it replaces.
Check how a candidate asks for a different interview format. Who replies, who arranges it and how will the manager receive the answers? A help link is not enough if nobody handles the request before the shortlist is decided.
How should you compare the full cost?
Keep the existing cost, the extra cost and any cost you can actually remove separate. “Already in Lever” does not necessarily mean included at no charge. A specialist subscription does not necessarily mean all integration and support costs are included.
Ask both suppliers for the same expected volume and hiring pattern. Use the screening quote worksheet to compare billing units, unused allowances, retries, minimum terms and charges when hiring pauses.
Add the internal effort: setting criteria, reviewing outputs, handling candidate questions, fixing errors and maintaining the integration. Compare that with the best existing process, not an intentionally inefficient baseline.
A report that explains why the delivery was delayed may be worth more even if it takes the same time to read. Conversely, a cheap subscription may require more checking and support. Say which benefit you observed; don't call better information a time saving unless you measured one.
Name the improvement you would actually buy
End the comparison with what the team learned and what it took to learn it. Record useful new answers, important mistakes and the time recruiters and managers spent checking the results.
Return to the dispatch question. An answer explaining where the delivery went wrong gives the manager something to investigate. A report that simply repeats “helped with a delay” has not resolved the original gap.
Recommend the additional tool only when the improvement warrants its full cost, including review and support. Name the vacancy, the task it replaces, the person reviewing results and the cost. Keep the current process when it already does the job well.
