Enterprise recruiting on one rubric across 6 departments
In this enterprise recruiting case study, an Austin SaaS company with 6 departments and 40+ hiring managers moved first interviews to The Cognitive's live AI video interviews, with one rubric per role. Customer-reported: same-rubric coverage rose from about 20% to 100%, time to fill fell from 58 to 21 days, and recruiter hours per hire from 31 to 9.
What changed across the 6 departments?
The first 3 rows are the customer's own figures. The last 2 are our arithmetic on them, with the working further down.
| Metric | Before | After | Change |
|---|---|---|---|
| Candidates interviewed on the role's same rubric (customer-reported) | About 20% | 100% | Every candidate, every department |
| Time to fill (customer-reported) | 58 days | 21 days | 37 days shorter, 64% less |
| Recruiter hours per hire (customer-reported) | 31 | 9 | 22 hours saved per hire |
| Recruiter hours per 100 hires (our arithmetic) | 3,100 | 900 | 2,200 hours back |
| Candidates per 100 interviewed off the rubric (our arithmetic) | About 80 | 0 | No exceptions left |
Figures reported by the customer in our case study record dated 4 August 2026.
Who is the customer, and why was its hiring inconsistent?
The customer is an enterprise SaaS company in Austin, Texas, named in our record only as Enterprise SaaS Company. It hires across 6 departments, and more than 40 hiring managers ran interviews, about 7 per department on average. Each manager interviewed their own way, so there were as many versions of the bar as there were managers.
That showed up in 3 places. Evaluations could not be compared across teams. The compliance team could not show that candidates for the same role were judged on the same things. And candidates had a different experience depending on which manager they drew. Only about 20% of candidates were interviewed against a shared rubric for their role.
The sharpest symptom: candidates rejected by one department were being hired by another, and nobody could say whether the difference was the candidate or the interviewer.
For the first time, a shortlist from any department means the same thing: evidence. VP Talent, Enterprise SaaS Company, Austin TX
Why did earlier fixes for interview consistency fail?
The team tried the usual fixes first. It tightened interview guides, pulled recruiters into more calibration calls and asked hiring managers to write down their reasoning after each decision. The record's summary is that this produced more paperwork without more consistency, because it still depended on 40+ people choosing to follow a document.
Some departments trialled point tools for their own roles. Those broke down when a candidate moved between departments, or when compliance asked for evidence it could compare across teams. Each evaluation of enterprise AI recruiting software kept stalling on one question: could a single system handle engineering, sales and support hiring?
How did the AI interview differ by department?
The record names 3 of the 6 departments and what their interviews focused on. The example criteria are illustrations of the kind of rubric each could use, not the customer's own.
| Department | What the AI interview focused on | Example criteria |
|---|---|---|
| Engineering | Technical rounds: design and problem-solving, talked through live | Problem-solving, system design, ownership, communication |
| Sales | Conversational role-plays of sales situations | Discovery questions, handling objections, clarity, next steps |
| Support | Empathy and de-escalation scenarios | Empathy, de-escalation, accuracy, written follow-up |
| The other 3 departments | Not named in the record; each got its own rubric per role | Set by each department |
Within each role, every candidate is scored on the same criteria. The questions adapt live to each answer.
How did each department's bar become a written rubric?
Building an evaluation template: criteria come in with weights, one weight is adjusted, and the total reaches 100 before the template is saved.
Rollout started by translating each department's bar into criteria with weights. That surfaced the real problem early: managers defined the same signal differently, and those definitions had to be settled before the first candidate entered the flow. That disagreement had been there all along; writing the rubric made it visible.
How did they roll out one AI interview platform across departments?
The setup dates from August 2026 and the product has changed since. These steps describe how the same approach runs in The Cognitive today.
- One role per job, one rubric per role Each department writes weighted criteria for each of its roles, with what a 1, a 3 and a 5 look like. A central talent team can review them so the format matches across departments. A format we suggest: 4 to 6 criteria per role, weights adding to 100, and one sentence each for a 1, a 3 and a 5.
- Set length and language per role A recruiter sets each role to 10 or 20 minutes and picks one of 9 languages, so a support role and an engineering role can differ while each stays consistent inside itself. 20 minutes for engineering and sales role-plays, which need follow-ups. 10 minutes can suit a high-volume support role with 3 or 4 criteria.
- Import candidates from the ATS Applicants come in through one of 60+ ATS integrations, or by CSV. Recruiters bulk invite each batch, and candidates book their own slot without an account. Enterprise ATSs on the list include Workday, SAP SuccessFactors, Oracle Taleo, iCIMS and Greenhouse.
- Every candidate gets the same live interview format The AI interviews each candidate over video against the role's rubric and follows up on thin answers. Integrity flags such as tab switches are logged for review, not scored. Role prompts to start from: [software engineer](/interview-questions/software-engineer), [sales representative](/interview-questions/sales-representative) and [customer support](/interview-questions/customer-support) interview questions.
- Reports land in one format for every department Each report gives 1 to 5 per criterion, a weighted score out of 100, overall written feedback, a transcript and a recording. ATS-imported candidates get a note with score, summary and report link. A review rule we suggest: recruiters triage reports daily and send each manager only candidates whose verdict and criterion scores agree.
- Hiring managers still decide Managers read the report and accept or overrule the suggested verdict. Nothing is rejected automatically, and ATS stages move only when the team moves them. When a manager overrules a verdict, note why. A pattern of overrules is the signal a rubric needs rewriting.
- Source for roles inbound cannot fill For niche roles, a recruiter describes the person in plain English, searches ~900M public profiles, reveals verified emails and phone numbers, and invites the shortlist to the same interview. People sourced inside The Cognitive are not written into the ATS automatically; add hires there as you do today.
How did time to fill fall from 58 days to 21?
Time to fill dropped by 37 days, or 64%, across the company. The first round no longer waited on 40+ managers' calendars: candidates booked their own interview, and managers could see from each report who was worth their time.
Recruiter hours per hire fell from 31 to 9, so each hire now takes 22 fewer recruiter hours. Per 100 hires that is 3,100 recruiter hours before and 900 after. At the BLS median pay for human resources specialists, $36.51 an hour (BLS, read 5 October 2026), the 2,200 hours saved are worth about $80,322 in salary per 100 hires.
The market context makes 21 days stand out. SHRM's 2026 benchmarking puts median time to fill at 39 days for nonexecutive roles, so this company went from 19 days slower than the median to 18 days faster. SHRM also found requisitions per recruiter up 67% at extra-large organizations in 2026, and at that load 22 fewer hours per hire becomes capacity for more open roles.
How do 21 days and 9 recruiter hours compare with enterprise benchmarks?
Published 2026 benchmarks, read 5 October 2026, next to the company's reported figures.
| Measure | Typical | This company before | This company after |
|---|---|---|---|
| Time to fill | Median 39 days, nonexecutive roles (SHRM 2026) | 58 days | 21 days |
| Opening to first hire | 56 days business, 76 days technical (Ashby 2026) | 58 days to fill | 21 days to fill |
| Recruiter load | 13.4 open roles per recruiter (Gem 2026); requisitions per recruiter up 67% at extra-large organizations (SHRM 2026) | 31 recruiter hours per hire | 9 recruiter hours per hire |
| Large-company funnel | Under 10% of applicants reach pre-onsite; 0.7% hired (Gem 2026) | About 20% interviewed on the role's rubric | 100% interviewed on the role's rubric |
Benchmarks cover all industries. The company's figures are as reported in our August 2026 record.
The working, line by line
Inputs are from the customer's record, plus BLS pay and SHRM's benchmark. Everything after an equals sign is our arithmetic.
TIME TO FILL Before 58 days. After 21 days. 58 minus 21 = 37 days shorter. 37 / 58 = 63.8%, the 64% in the record. RECRUITER HOURS PER HIRE Before 31. After 9. 31 minus 9 = 22 hours saved per hire. 22 / 31 = 71% less. PER 100 HIRES (a round number for scale) Before: 100 x 31 = 3,100 hours After: 100 x 9 = 900 hours 3,100 minus 900 = 2,200 hours back 2,200 x $36.51 (BLS median HR specialist pay) = about $80,322 Per hire: 22 x $36.51 = about $803 AGAINST SHRM'S 2026 MEDIAN OF 39 DAYS Before: 58 minus 39 = 19 days slower After: 39 minus 21 = 18 days faster SAME-RUBRIC COVERAGE Before about 20%. After 100%. Candidates per 100 interviewed off the rubric: Before 100 minus 20 = about 80 After 100 minus 100 = 0 HIRING MANAGERS PER DEPARTMENT 40+ / 6 = about 7 on average
What did a candidate experience in any department?
A live, two-way AI video interview: the candidate checks camera and microphone, the AI asks and follows up, and the room closes on time.
A sales candidate and an engineering candidate got different questions on different criteria, but the same kind of interview: live, on video, at a time they booked, with the AI following up on each answer. Candidate experience stopped depending on which of 40+ managers they happened to draw.
What did 100% same-rubric coverage give the compliance team?
The compliance team's requirement was that candidates for a role be interviewed under a documented rubric that could be checked later. Before, about 20% were. After, all were, because the rubric is written into the role and applied by the interview rather than asked of each manager.
What there is to check: the role's written criteria and weights, and for each candidate a 1 to 5 score per criterion, the overall feedback, the transcript and the recording. The Cognitive has no published bias audit, so an employer's own legal and compliance review still applies. What changed is that every decision now starts from comparable evidence.
What does this case study not show?
- One company, so use it to plan your own rollout department by department.
- Scores compare candidates within a role; a sales score and an engineering score measure different criteria.
- It is not an ATS: offers, human interview scheduling and funnel reporting stay in yours.
- The setup dates from August 2026. The product has changed since, and this page describes today's workflow.
What still depends on hiring managers?
Every hiring decision, and the rubric itself. Some interviewers first read the AI report as a challenge to their judgment. The team had to say clearly that the report standardizes evidence while the decision stays with the manager. Managers still run later rounds, make offers and own the outcome.
Rubrics also need owners. When a department changes what good looks like for a role, someone has to update the criteria, or the 100% coverage measures the wrong thing. Offers, scheduling of human interviews and funnel reporting stay in the ATS, because The Cognitive is not one.
A recruiter's week, before and after. Before: 31 hours per hire, much of it finding first-round slots across 40+ managers and chasing write-ups that followed no shared format. After: 9 hours per hire, spent importing and inviting batches, triaging reports that look the same in every department and handing managers candidates already scored on the role's rubric.
Which enterprise hiring teams does this approach fit?
Do
- Several departments hiring at once, each with its own idea of the bar.
- Dozens of hiring managers running first interviews on their own calendars.
- A compliance team asking for comparable evidence across teams.
- Roles mixing technical, sales and support hiring under one talent team.
Do not
- Executive searches run by a retained firm, one candidate at a time.
- Teams that want a single tool to replace their ATS.
- Departments that will not agree on written criteria before inviting anyone.
Start with 2 departments
Pick one technical role and one customer-facing role, write their rubrics with the managers who hire for them and compare the reports side by side. Start free at https://app.thecognitive.io/signup.
Where to go next
What is a structured interview · AI screening software and compliance · AI recruiting for tech hiring · AI recruiting for customer support · Pricing · All case studies
Sources and notes
Figures and quote reported by the customer in The Cognitive case study record for Enterprise SaaS Company, Austin TX; setup dated August 2026 (record dated 4 August 2026). Derived figures are our arithmetic and are labelled. The example criteria are illustrations, not the customer's own. The workflow is described as the product works today.
Benchmarks, all read 5 October 2026: SHRM 2026 Recruiting Benchmarking (39-day median time to fill, nonexecutive; requisitions per recruiter up 67% at extra-large organizations); Ashby 2026 Talent Trends, recruiter productivity (56 and 76 days to first fill); Gem 2026 Recruiting Benchmarks (13.4 open roles per recruiter; large-company funnel); BLS, human resources specialists ($36.51 median hourly pay).
Frequently asked questions
Where can I find an enterprise recruiting case study with real numbers?
This page is one: an Austin enterprise SaaS company with 6 departments and 40+ hiring managers that moved first interviews to The Cognitive's live AI interviews. It reported time to fill falling from 58 days to 21. Every figure is labelled as customer-reported or as our arithmetic.
What should AI recruiting software for enterprises do?
Apply one written bar per role across every team, and leave decisions with people. In this case study that meant a rubric per role in every department, a live AI interview for every candidate, and a report in the same format for each one. It also needs to work beside your ATS, through an integration or CSV.
How do you standardize interviews across 40+ hiring managers?
Move the rubric into the interview instead of a document managers are asked to follow. This company's same-rubric coverage went from about 20% to 100% because the AI interview applies the role's criteria to every candidate, whichever manager owns the role.
Can one AI interview platform handle engineering, sales and support hiring?
It did here. Engineering roles got technical rounds, sales roles got conversational role-plays and support roles got empathy and de-escalation scenarios, each with its own rubric. What stayed the same was the format: a live AI video interview and a report with 1 to 5 scores per criterion.
How do you reduce time to fill at enterprise scale?
Take the first interview off managers' calendars. This company's time to fill fell from 58 days to 21, a 64% cut, once candidates booked their own AI interview and managers decided from the reports who to meet. Recruiter hours per hire fell from 31 to 9.
What is a structured interview process?
A structured interview process assesses every candidate for a role against the same written criteria, with a set scale for scoring. With The Cognitive the criteria and weights are fixed per role while the questions adapt to each answer, and every criterion gets a 1 to 5 score.
Do departments lose the ability to assess what matters to them?
No. Each department writes the rubric for its own roles, with its own criteria and weights. What is standardized is that every candidate for a given role is assessed against that role's criteria, in the same interview format.
Does the AI make hiring decisions?
No. Each report suggests a verdict from the weighted score, and a hiring manager accepts or overrules it. Nothing is rejected automatically. In this company, managers kept every decision; what changed is that they all started from the same kind of evidence.
What is the average time to fill at large companies?
SHRM's 2026 Recruiting Benchmarking puts the median at 39 calendar days for nonexecutive roles, from data on over 4,600 organizations (read 5 October 2026). Ashby's 2026 data shows 56 days from opening to first hire for business roles and 76 for technical ones. This company went from 58 days to 21.
How many recruiter hours does an enterprise hire take?
This company reported 31 recruiter hours per hire before and 9 after. At the BLS median pay for human resources specialists, $36.51 an hour (read 5 October 2026), the 22 hours saved are worth about $803 per hire, or about $80,322 per 100 hires.
How many requisitions does an enterprise recruiter carry?
Gem's 2026 Recruiting Benchmarks report 13.4 open roles per recruiter, and SHRM found requisitions per recruiter up 67% at extra-large organizations in 2026 (both read 5 October 2026). Taking first interviews off recruiter and manager calendars is how this company handled that load with fewer hours per hire.
How do you roll out structured interviews across departments?
Start with 2 roles, one technical and one customer-facing. Have the managers who hire for them write 4 to 6 weighted criteria with a 1, 3 and 5 described, run a batch of AI interviews, and compare reports side by side before adding the next department.
Does The Cognitive integrate with enterprise ATSs?
Yes, through 60+ ATS integrations, including Workday, SAP SuccessFactors, Oracle Taleo, iCIMS and Greenhouse. Candidates imported from the ATS get a note with score, summary and report link. Stages never move on their own; your team moves them.
How much does AI recruiting software for enterprises cost?
The Cognitive publishes its prices: AI Interview and AI Sourcing each start at $99/month, on monthly plans you can cancel anytime. AI Interview plans include a monthly interview allowance; past it a team buys an interview top-up. Yearly billing is 17% off, and Team and Business plans add seats. Signup is self-serve at app.thecognitive.io/signup.
Explore: AI bias in hiring guide · How to train hiring managers · High volume hiring · Hire SDRs with AI · Customer support interview questions · Technical interview as a service case study