Introduction
Abhishek, Fabric's co-founder and head of sales, had 40 demos on his calendar every week. Pipeline was strong. And we desperately needed an SDR to keep it fed.
The person who needed to evaluate SDR candidates was the same person who had no time to talk to them. Every hour spent interviewing was an hour pulled away from closing deals.
So we used our own product. Fabric screened 1,300 applicants, ran 127 AI interviews, and handed us a ranked shortlist of 10 candidates. We hired Taabish within a week. This post breaks down how we designed the interview, what we tested for, and what happened.
Why is hiring SDRs so painful for growing startups?
SDR hiring has a timing problem. You need SDRs most when your sales team is already stretched thin. The people best equipped to evaluate sales talent are buried in demos, pipeline reviews, and quota targets.
Series A through D startups feel this acutely. You have enough traction to need dedicated pipeline generation, but your sales leaders are still player-coaches doing everything themselves.
There is also an evaluation problem. SDRs are uniquely hard to assess in traditional interviews. Good candidates know the right things to say. They talk about quota attainment and outreach strategy without breaking a sweat. The gap between someone who talks a good game and someone who actually performs shows up months later, not in a 30-minute conversation.
What should you actually test when hiring an SDR?
We thought about what separates SDRs who ramp fast from those who flame out. Five things kept coming up.
1. Understanding of numbers
A strong SDR thinks in terms of funnels, conversion rates, quota, and ACV. Many candidates inflate their numbers during interviews. When you ask them to walk through their funnel math, the story starts falling apart. If someone claims they hit 120% of quota but can't explain how many calls it took to generate that pipeline, something doesn't add up.
2. Deep understanding of business and ICP pain points
The best SDRs don't just follow a script. They understand why their ICP cares, what problems keep them up at night, and how to connect those problems to a conversation. This requires genuine curiosity about the business they sold into previously.
3. Agency
SDRs who wait to be told what to do rarely succeed. You want people who figure things out, experiment with new channels, and take ownership of their number. This is hard to test in a standard interview, but it shows up in how candidates describe their past work.
4. Cold calling skills
You can coach messaging. You can refine talk tracks. But the fundamental ability to hold a conversation with a stranger, handle objections, and stay composed under pressure is something a candidate either has or doesn't.
5. Tasteful outreach
Cold emails that get replies look different from cold emails that get ignored. You want SDRs who understand personalization, brevity, and timing. The difference between a good email and a bad one is often subtle, but it compounds over hundreds of sends.
How did we design an AI interview that tests all five?
We built the Fabric interview around three sections, each designed to pressure-test different qualities without making the interview feel like an interrogation.
1. Experience deep-dive
The first question was simple: tell me about your last role. Fabric's AI then followed up based on the candidate's answers, digging into their understanding of the business, their ICP, how they hit quota, and how their funnel actually worked. This single question, combined with intelligent follow-ups, eliminated roughly 50% of candidates. The ones who had inflated their resumes couldn't maintain consistency when Fabric kept asking specifics.
2. Cold call simulation
Fabric is particularly strong at role-play scenarios. We set up a simulation where the candidate had to cold call a prospect matching our ICP. The AI played the prospect, complete with objections, questions, and realistic pushback. This tested whether candidates could think on their feet, handle rejection, and actually drive a conversation toward a meeting.
3. Follow-up email writing
After the cold call, we asked candidates to write a follow-up email to the same prospect. This closed the loop nicely. It tested their writing ability, their understanding of what happened in the call, and whether they could craft a concise, relevant message. Candidates who struggled with the cold call also tended to write generic follow-up emails, which confirmed the signal.
The entire interview took about 30 minutes per candidate and covered everything we needed to evaluate.
What happened when we ran 127 interviews in one week?
Fabric connected to our LinkedIn and Naukri job posts and started processing applications immediately.
1,300 resumes screened. Fabric filtered for relevant experience, skills, and role fit before anyone on our team touched a single application.
127 AI interviews conducted. Candidates who passed the resume screen went through the same three-section format with consistent evaluation criteria.
Top 10 candidates ranked. Fabric scored every interview based on the criteria we set, with specific notes on where each person excelled or fell short.
Taabish, our eventual hire, was ranked number 3. Abhishek only needed to speak with a handful of finalists before making the call. The whole process took one week. A typical SDR hiring cycle at a startup of our stage takes three to six weeks, assuming your sales leader actually has time to interview.
Here's the Linkedin post that shows a snippet from Taabish's interview.
How can you set up something similar for your SDR hiring?
If you are hiring SDRs without pulling your sales leaders out of revenue work, here's what we learned.
Define your evaluation criteria before you build anything. The five qualities we identified came from observing what made our best salespeople effective. Your criteria might differ depending on your market and sales cycle.
Design questions that create follow-up opportunities. Our first question seemed simple but opened dozens of paths for deeper evaluation. Avoid questions with yes-or-no answers.
Use simulations wherever possible. Asking an SDR to describe their cold calling approach is useful. Having them actually do a cold call is significantly more revealing.
Close the loop between sections. Our cold call into follow-up email structure forced candidates to build on their earlier performance. It tested consistency and adaptability in a way that disconnected questions cannot.
Conclusion
When your best sales leader is booked solid with demos, the worst thing you can do is pull them into weeks of interviews.
We needed an SDR fast. We got one in a week, without compromising on evaluation depth, and without taking a single hour away from Abhishek's demo calendar. The AI interview tested everything we cared about, often more rigorously than a human interviewer could in a single session.
FAQ
Can AI interviews actually evaluate soft skills like cold calling ability?
Yes. AI interviews that use simulation and role-play can assess how candidates handle objections, manage conversation flow, and respond under pressure. The key is designing scenarios that mirror real selling situations.
How long does a typical AI interview take for SDR candidates?
The interview we designed took about 30 minutes per candidate, covering an experience deep-dive, a cold call simulation, and a follow-up email exercise.
What is Fabric?
Fabric is an AI interview platform that screens resumes, conducts AI-powered interviews with follow-ups and simulations, and ranks candidates based on your evaluation criteria. It connects with job boards and handles the full top-of-funnel hiring process.
Does using AI interviews reduce the quality of hires compared to human interviews?
In our experience, the opposite happened. Fabric's interviews were more consistent, tested more dimensions in a single session, and eliminated the scheduling bottleneck that forces compromises.
Can AI interviews replace human interviews entirely?
For top-of-funnel screening, AI interviews handle the heavy lifting. We still had Abhishek speak with finalists before making an offer. The combination of AI screening with human final-round conversations gave us both efficiency and judgment.