How an Austin SaaS team learned to hire software engineers in 9 days
A 120-person B2B SaaS company in Austin needed to hire software engineers and used The Cognitive's live AI video interviews for its early rounds. It hired 18 engineers in 11 weeks. Customer-reported: time to hire fell from 38 to 9 days, engineering time on hiring from 22 to 4 hours a week, and offer acceptance rose from 62% to 91%.
What changed when this SaaS team moved its early rounds to AI?
The first 3 rows are the customer's before and after figures. The last 2 are our arithmetic on them, and the working is further down the page.
| Metric | Before | After | Change |
|---|---|---|---|
| Time to hire (customer-reported) | 38 days | 9 days | 29 days shorter, 76% less |
| Engineering hours on hiring per week (customer-reported) | 22 | 4 | 18 hours back each week |
| Offer acceptance rate (customer-reported) | 62% | 91% | Up 29 points |
| Offers needed for 18 hires (our arithmetic) | About 29 | About 20 | About 9 fewer offers |
| Engineering hours returned over the 11 weeks (our arithmetic) | 0 | 198 | 18 x 11 weeks |
Figures reported by the customer in our case study record dated 6 April 2026.
Who is the customer, and what were they trying to do?
The customer is a Series B, 120-person B2B SaaS company in Austin, Texas, named in our record only as Series B SaaS Company. It set out to double its engineering team inside one quarter, with a product launch waiting on those hires. 18 engineers is 15% of a 120-person company's headcount, added in one quarter and one function.
The hiring team was 2 recruiters, and both were already at capacity. Senior engineers filled the gap by running screening calls themselves, in some weeks for full days at a time. Across the team, engineering time on hiring had reached 22 hours a week, which is more than half of one engineer's working week spent on interviews.
Those 22 hours had a price. At the BLS 2025 median for software developers, QA analysts and testers, $64.44 an hour (BLS, read 5 October 2026), 22 hours is about $1,418 of engineering salary a week before benefits. Over the 11-week push that is 242 hours, or about $15,594, spent on screens instead of the launch the hires were meant to ship.
How do 9 days and 91% compare with typical engineering hiring?
Published 2026 benchmarks for technical roles, read 5 October 2026, set against the customer's reported figures.
| Measure | Typical for technical roles | This team before | This team after |
|---|---|---|---|
| Application to accepted offer | About 48 days (Ashby 2026) | 38 days | 9 days, about 39 days under the benchmark |
| Offer acceptance | 82% (Gem 2026) | 62%, 20 points below | 91%, 9 points above |
| Hires per recruiter | 3.8 a quarter for technical roles (Ashby 2026) | Not reported | 9 each in 11 weeks |
| Interview time per hire | 24.7 hours for engineering, all interviewers (Ashby 2026) | 22 engineering hours a week | 4 a week, about 2.4 per hire |
Per-hire engineering time is our arithmetic: 4 hours x 11 weeks = 44 hours / 18 hires.
Our engineers were spending more time interviewing than coding. Director of Engineering, Series B SaaS Company, Austin TX
Why did a 38-day hiring process break the plan?
A quarter is about 13 weeks. Their actual hiring window turned out to be 11 weeks, or 77 days. At 38 days from application to hire, one open seat fits only 2 full hiring cycles in that window. If an offer is declined late, a seat might get just 1 real attempt before the launch date.
Speed also cost them candidates. The record says applicants were dropping out while the process dragged on, and only 62% of offers were accepted. A strong engineer who waits 5 weeks for a decision has usually had other conversations by then, and the offer that arrives first tends to win.
The team first tried to work harder. Recruiters sent more reminders, engineers took on more screens and hiring managers argued over borderline candidates in long debriefs. The record calls the result motion, not capacity: the same people did more of the same work, and the calendar stayed full.
How did they run AI interviews for engineering roles?
The setup dates from April 2026 and the product has changed since. These steps describe how the same approach runs in The Cognitive today.
- Agree the bar with engineering leads Each engineering role gets a rubric of weighted criteria. The team wrote what a 1, a 3 and a 5 look like, then checked early reports against candidates it already knew well. Criteria we suggest for a product engineer: technical depth 30, debugging approach 25, system thinking 20, ownership 15, communication 10.
- Pick the interview length A recruiter sets each role to 10 or 20 minutes. The customer reported 30-minute interviews in its setup; today 20 minutes is the longer option and suits technical depth. Use 20 minutes for engineers. 10 minutes leaves room for one project story but not the follow-up on what broke and why.
- Bring applicants in and bulk invite them Import applicants through one of 60+ ATS integrations or a CSV, select the batch and invite everyone in one go. Candidates book their own slot without an account. Invite each day's applicants that same day, so nobody waits a week just to be screened.
- Let the AI run the conversation The AI interviews each candidate live over video, mixing technical and behavioral questions, and follows up when an answer stays vague. Candidates talk through their reasoning out loud; there is no code editor. More prompts by role: [software engineer interview questions](/interview-questions/software-engineer) and [backend developer interview questions](/interview-questions/backend-developer).
- Read the report before anyone meets the candidate Each report scores every criterion 1 to 5, adds a weighted score out of 100 and overall written feedback, and includes the transcript and recording. A review rule we suggest: recruiters read reports daily and pass any 4 or 5 on technical depth to an engineering lead, with the recording.
- Engineers meet only the candidates above the line A recruiter or engineering lead accepts or overrules the suggested verdict. Engineers spend their remaining hours on final rounds with candidates who already cleared the agreed bar. Make the offer within days of the final round. This team's 91% acceptance came with a 9-day process.
- Source when inbound runs thin For a senior or niche seat, describe the engineer in plain English, search ~900M public profiles, reveal verified emails and phone numbers, and invite shortlisted people to the same interview. A brief like: senior backend engineer, 5+ years of Go, payments experience, based in Texas or remote US.
How does the AI follow up when an engineer's answer is thin?
An illustration of interview technique: the next question is built from the last answer, so a vague story gets a specific follow-up.
The rubric is fixed for the role and the questions adapt live. When a candidate says they scaled a service, the AI asks what broke first, what they measured and what they would change. That is the probing senior engineers used to do on screening calls, now done before they join.
How did time to hire fall from 38 days to 9?
Every early round depended on a senior engineer's calendar: a free hour for the screen, then a debrief, then another slot for the next round. An interview the candidate books for themselves needs nobody on the team to be free at the same moment.
The result was 9 days from application to hire, against 38 before: 29 days shorter, or 76% less time. Over the 11-week push the company made 18 engineering hires, about 1.6 a week, with the same 2 recruiters. That works out to 9 hires per recruiter in under 3 months.
Put that next to the market. Ashby's 2026 data has technical recruiters at 3.8 hires a quarter, and about 48 days from application to accepted offer. At 3.8 a quarter, 2 recruiters would expect about 7 or 8 engineering hires in a 13-week quarter, so 18 in 11 weeks is more than double that pace. 9 days is under a fifth of the 48-day benchmark.
The working, line by line
Inputs come from the customer's record, plus BLS pay and Ashby's benchmark. Everything after an equals sign is our arithmetic.
TIME TO HIRE Before 38 days. After 9 days. 38 minus 9 = 29 days shorter. 29 / 38 = 76% less time. HIRING CYCLES THAT FIT IN THE 11 WEEK WINDOW (77 days) Before: 77 / 38 = about 2 After: 77 / 9 = about 8 ENGINEERING HOURS ON HIRING Before 22 a week. After 4 a week. 22 minus 4 = 18 hours back each week. 18 / 22 = 82% less. 18 hours x 11 weeks = 198 hours returned. 198 hours x $64.44 (BLS median hourly pay) = about $12,759 of engineering salary. Per hire after: 4 x 11 = 44 hours / 18 hires = about 2.4 hours. OFFERS NEEDED FOR 18 ACCEPTED HIRES Before: 18 / 0.62 = about 29 offers, about 11 declined After: 18 / 0.91 = about 20 offers, about 2 declined PACE 18 hires / 11 weeks = about 1.6 a week 18 hires / 2 recruiters = 9 each Benchmark: 3.8 a quarter x 2 recruiters = about 7.6
Why did offer acceptance rise from 62% to 91%?
Offer acceptance moved 29 points, and the Director of Engineering ties it to speed. The record's own lesson is that the offer which arrives first is usually the one accepted. At 9 days, most candidates were still early in other processes when this company made its offer.
The arithmetic shows why this matters as much as the time saved. Reaching 18 hires at 62% acceptance needs about 29 offers, so about 11 of them come back declined. At 91% it takes about 20 offers, with about 2 declined. Every declined offer reopens a seat and sends engineers back into final rounds for it.
The launch is the outcome the team cared about. In the Director of Engineering's words: "We shipped our launch on time because of this." Gem's 2026 benchmarks put offer acceptance at 82%, its highest since 2021, so this team went from 20 points below the market to 9 points above it.
What did a senior engineer's week look like before and after?
Built from the reported hours and the workflow the team described. Hours are team totals per week.
| Part of the week | Before (38-day process) | After (9-day process) |
|---|---|---|
| Engineering hours on hiring | 22 a week across the team | 4 a week |
| Screening calls | Senior engineers, some weeks for full days | Run by the AI at slots candidates book |
| Debriefs | Long arguments over borderline candidates | A written bar per role and a 1 to 5 report to argue from |
| What engineers read | Their own call notes | Overall feedback, scores, and the recording where a score looks borderline |
| Final rounds | Squeezed between screens | The main use of the 4 hours |
| Declined offers | About 11 for 18 hires | About 2 for 18 hires |
Declined offers are our arithmetic on 62% and 91% acceptance.
What did engineers review instead of running screens?
An example interview report for an engineering role: a score per criterion, a weighted score out of 100 and resume claim checks.
An engineer opens the report, reads the overall feedback and the scores, and jumps into the recording only where a score looks borderline. The report suggests a verdict, and a person makes the call. Nothing is rejected automatically, and the ATS gets a note while stages stay where the team puts them.
What does this case study not show?
- One company, one hiring push, one role family, so use it to plan your own push.
- The AI interview covers early rounds only. Final rounds, offers and closing stayed with the team.
- There is no code editor in the interview; add a coding exercise later if a role needs one.
- The setup dates from April 2026. The product has changed since, and this page describes today's workflow.
What still needed engineers and recruiters?
Final rounds stayed with the engineering team, and so did every hiring decision. The 4 hours a week left on hiring went on candidates who had already shown the work. Offers, references and closing stayed with the recruiters, because The Cognitive is not an ATS and does not send offers or schedule human interviews.
Trust took work too. In the first weeks some engineers wanted to screen candidates again anyway. The team had to agree in writing what a pass meant for each role before engineers stopped double-checking the reports. If you copy this, budget a week for that conversation before the first invite goes out.
Which engineering teams does this approach fit?
Do
- Teams hiring several engineers at once, with senior engineers running first screens.
- Roles with steady inbound applicants, imported from an ATS or a CSV.
- A launch or roadmap date that a slow hiring process puts at risk.
- Teams willing to write down what a 1, 3 and 5 look like.
Do not
- One senior hire filled from your own network in a few calls.
- Roles where a hands-on coding exercise must be the first filter; add one as a final round.
- Teams that will not let a written bar replace a manager's gut feel.
Try it on one engineering role
Set up one role, agree the rubric with an engineering lead, bulk invite this week's applicants and compare the reports with your own screens. Start free at https://app.thecognitive.io/signup.
Where to go next
Hire software engineers with AI · AI recruiting for tech hiring · Software engineer interview questions · How the AI interviewer works · Pricing · All case studies
Sources and notes
Figures and quotes reported by the customer in The Cognitive case study record for Series B SaaS Company, Austin TX; setup dated April 2026 (record dated 6 April 2026). Derived figures are our arithmetic and are labelled. The quote is shortened to the sentences that match how the product works today, and the workflow is described in today's terms.
Benchmarks, all read 5 October 2026: Ashby 2026 Talent Trends, recruiter productivity (48 days application to offer and 76 days to first fill for technical roles, 3.8 technical hires per recruiter a quarter, 24.7 interview hours per engineering hire); Gem 2026 Recruiting Benchmarks (82% offer acceptance); BLS Occupational Outlook, software developers ($64.44 median hourly pay in 2025 for software developers, QA analysts and testers).
Frequently asked questions
How do you hire software engineers faster?
Remove the scheduling waits in the early rounds. This Austin SaaS company moved its early interviews to The Cognitive's live AI video interview, which candidates book for themselves, and reported time to hire falling from 38 days to 9. Engineers then met only candidates who cleared a bar they had agreed in writing.
Where can I find a SaaS hiring case study with real numbers?
This page is one: a 120-person B2B SaaS company that hired 18 engineers in 11 weeks, with every figure labelled as customer-reported or as our arithmetic. More customer stories are at thecognitive.io/case-studies. In any case study, check who reported the numbers and over which period.
How do you reduce the time engineers spend on interviews?
Take first screens off their calendars and give them a report to read instead. This team cut engineering time on hiring from 22 hours a week to 4. The 4 hours that remained went on final rounds with candidates who had already shown the work in an AI interview.
Does a faster hiring process improve offer acceptance rate?
It did here. Offer acceptance rose from 62% to 91% as time to hire fell from 38 days to 9. The pattern is common: the first offer a strong engineer receives is often the one they take, and a 9-day process gets yours there first.
Can AI interview software engineers?
Yes, as a spoken technical interview. The Cognitive's AI asks technical and behavioral questions live over video, follows up on vague answers and probes up to 5 resume claims. Candidates explain their reasoning out loud. There is no code editor, so pair it with a coding exercise in a later round if you need one.
Does the AI replace the final engineering interview?
No. In this case study engineers still ran the final rounds and made every hiring decision. The AI interview covers the early rounds, and each report suggests a verdict that a recruiter or engineering lead accepts or overrules.
How long is the AI interview for an engineering role?
A recruiter sets each role to 10 or 20 minutes. This customer described 30-minute interviews when it reported its results in April 2026; the product has changed since. 20 minutes gives the AI room for follow-ups on technical answers.
Does the AI reject engineering candidates automatically?
No. Nothing is rejected automatically. The report gives a 1 to 5 score per criterion, a weighted score out of 100 and a suggested verdict, and a person decides who moves on.
Does The Cognitive connect to our ATS?
Yes, through 60+ ATS integrations, including Greenhouse, Lever, Ashby, Workday and iCIMS. Applicants imported from the ATS get a note with the score, summary and report link after the interview. Stages never move on their own; your team moves them.
How much does an AI interviewer for engineering hiring cost?
AI Interview starts at $99/month, on monthly plans you can cancel anytime. Each plan includes a monthly interview allowance, and extra interviews are billed at the plan's overage rate. AI Sourcing starts at $49/month if you also need to find engineers who have not applied. Signup is self-serve at app.thecognitive.io/signup.
How many engineers can 2 recruiters hire in a quarter?
This team's 2 recruiters made 18 engineering hires in 11 weeks, about 9 each, once early interviews no longer needed their time or engineers' calendars. Ashby's 2026 benchmark for technical recruiters is 3.8 hires a quarter each, read 5 October 2026.
What is the average time to hire a software engineer?
About 48 days from application to accepted offer for technical roles, per Ashby's 2026 Talent Trends data, and a median of 76 days from opening a technical role to its first hire (read 5 October 2026). This Austin team went from 38 days to 9.
What is a good offer acceptance rate for engineering roles?
Gem's 2026 Recruiting Benchmarks put offer acceptance at 82% across roles, the highest since 2021 (read 5 October 2026). This team moved from 62%, well below that, to 91%, above it, after cutting time to hire to 9 days.
How many interview hours does one engineering hire take?
Ashby's 2026 data shows 24.7 hours of total interview time per engineering hire, counting every interviewer (read 5 October 2026). This team spent 4 engineering hours a week over 11 weeks for 18 hires, about 2.4 hours per hire by our arithmetic.
What should an AI screening rubric for software engineers include?
4 to 6 weighted criteria a senior engineer would defend. Our suggested set: technical depth, debugging approach, system thinking, ownership and communication. Write what a 1, a 3 and a 5 sound like for each before the first invite, as this team did.
Explore: AI interviewer for software engineers · AI screening and time to hire · Time to hire compression research · Technical interview software · Interview scorecard reports · BPO hiring case study