AI SDR and BDR Interview Guide

An SDR interview is short on theory and long on performance. Somewhere in the loop, someone will play a buyer you've never met, and you'll have to call them or write to them, live or on a deadline. At an AI company, expect that buyer to be skeptical, busy, and possibly an engineer.

This guide shows you what each round is checking, how to structure the call and the email, and eight practice prompts to rehearse out loud.

None of the seven AI companies whose SDR and BDR postings were read for this guide publishes its interview loop. The rounds below are inferred from their requirement lines, opened on September 26, 2026. Ask the recruiter for the real order and format. Most will tell you.

The rounds to expect

Round	What it tests	What to bring
Recruiter screen	Motivation, communication, basic fit	A two-minute answer to "why sales, why AI, why us"
Hiring manager	Coachability, work habits, how you handle rejection	Two stories: a hard goal you hit, and feedback you acted on
Mock cold call	Opening, questioning, objection handling, asking for the meeting	A call plan for the company's likely buyer
Written exercise	Research and writing	A short email or sequence to one real account
Team or final round	Fit with the AEs you would support	Questions about how meetings are handed off and scored

Armand Farrokh of 30 Minutes to President's Club described a final "super day" that tests prospecting alongside fit in his AE interview process, published April 12, 2024. An SDR loop may fold the same test into an earlier round.

What the postings say they will test

Research before outreach

Fireworks' BDR posting asks for "personalized, tailored outreach through various research methods to prioritize the right accounts and prospects." Cohere's BDR posting names "LinkedIn Sales Navigator and CRM platforms to prioritise high-fit prospects."

Expect a written exercise that hands you a company and asks who you'd contact and why. The strongest answers name a specific, recent, public reason the account might care, and say where it came from. The research step in the AI SDR outbound playbook is good practice.

Technical comfort

Perplexity's BDR posting asks for "Comfort discussing enterprise software, APIs, and AI capabilities at a high level." Cohere asks for "a basic understanding of natural language processing (NLP) concepts."

You won't be asked to write code. You may be asked to explain the product to a CTO in two sentences, or to answer "how is this different from just using ChatGPT?" Use the product before the interview, read the overview page of its docs, and prepare a plain answer.

Qualification

Perplexity and Cohere both name BANT: budget, authority, need and timeline. Perplexity's BDR will "Qualify inbound enterprise prospects using BANT criteria to identify high-value opportunities for Account Executive handoff." Expect a question like "What makes a meeting worth handing to an AE?" Answer with the criteria, and say what you'd do with a lead that meets two of the four.

Volume with judgment

Cohere asks for "40+ outbound calls and personalised emails daily." Clay's ClayDR posting wants the "Ability to balance high-volume outreach with thoughtful experimentation and feedback." WRITER's SDR posting wants "an experimental mindset." A hiring manager will ask how you'd organize a day and how you'd know a message is working. Have a concrete answer: which numbers you'd track each week, and what you'd change first if replies dropped.

The mock cold call

Mock calls tend to be short. The interviewer plays a buyer and will push back at least twice. They're testing whether you stay calm, ask a real question, and still ask for the meeting.

Use a simple structure.

Opener. Your name, company, and an honest line on why you are calling this person. "This is a cold call, can I have thirty seconds to tell you why I picked you?" works because it is true.
Reason. One specific observation about their company, and the problem it suggests. Not a product pitch.
Question. Ask whether that problem is real for them. Then listen.
Objection. Acknowledge it, ask one question about it, and answer only if you understand it.
Ask. A specific meeting, with a specific person on your side, for a specific reason. "Would Tuesday at 2 work for 20 minutes with our AE, to see whether this fits your billing queue?"

The objections at an AI company are predictable. "We're building this ourselves." "We tried an AI tool and it made things up." "Legal won't let us send data out." Prepare one question for each instead of a comeback. "What did it get wrong?" moves the call forward. "Our model is more accurate" doesn't.

Here's how that sounds in a made-up practice call. You're selling an AI support tool, and the buyer runs support at an online furniture store.

You: Hi Dana, it's Sam from Acme. This is a cold call. Can I have thirty seconds to tell you why I picked you?

Dana: Fine. Thirty seconds.

You: Your careers page has three openings for support agents, all mentioning delivery questions. That usually means "where's my order" tickets are piling up. Is that what's going on?

Dana: Sort of. But we tried an AI chatbot last year and it made things up.

You: What did it get wrong?

Dana: It promised delivery dates we couldn't hit.

You: That makes sense. Would Thursday at 10 work for 20 minutes with our AE, just to look at whether delivery tickets can be answered from your shipping data instead of guessed?

Notice what Sam didn't do. No pitch, no "our model is better." One observation, one question, and an ask tied to the exact thing Dana complained about.

The written email

The written exercise is usually one email, or a short series of emails, to a real person at a real company. Keep it short enough to read on a phone.

Subject. Plain and specific. No fake "Re:" and no clickbait.
First line. The observation, with its source.
Second line. Why it might matter to this person's job.
Third line. One sentence on how your company helps with that, with a proof point you can back up.
Ask. One low-effort next step.

Check every fact before you submit. A wrong claim about the prospect is worse than a generic email. If you used an AI tool to draft it, say so if asked, and explain how you checked it. The write-outbound and research-account skills in the open-source GTM skills are built around not inventing facts about a prospect, which is the habit this exercise rewards.

Original practice prompts

These are original practice prompts written for this guide, not questions from any employer's loop.

Prompt	What a strong answer shows
Cold call a VP of Support at a mid-size retailer. They answer, "We already have a chatbot."	Asks what it handles today and what it hands to people, finds a gap, and asks for a meeting tied to that gap.
Cold call a head of ML who says, "We self-host open models. Why would we pay you?"	Asks what is hard about self-hosting today, listens, and offers a meeting with an engineer on your side. Does not bluff on technical detail.
Write a first email to a named general counsel at a public company, using only public sources.	A specific, sourced observation, a relevant reason, a short ask. No invented pain.
The prospect says, "Send me some information."	Asks what they would look for in it, offers something specific, and books a time to follow up.
You booked 12 meetings last month and your AE accepted 3. What do you do?	Reviews why the other nine were rejected, tightens qualification, and talks to the AE before sending more.
Explain our product to your grandmother, then to a CTO.	Two clear versions with the same core claim at different depth.
Your reply rate halved in two weeks. Walk us through how you would find out why.	Checks deliverability, list quality and message changes one at a time, with the numbers each check would use.
Why do you want to sell AI, and not another kind of software?	A specific reason tied to the company's buyers, not "AI is the future."

Answer each aloud in under two minutes, then check against the right-hand column. For the call prompts, have a friend play the buyer and object twice.

How to prepare, in order

Read the posting line by line. Mark every tool, framework and target it names. Those are your interview topics.
Use the product. Where there is a free tier, spend an hour with it.
Pick three target accounts and research each as described in the outbound playbook. Bring your notes.
Record two mock calls. Count how often you asked a question versus pitched.
Write one email per account and check every fact in it.
Prepare the pay questions. Ask whether the posted number is base or OTE, the split, how quota is set, and how many SDRs hit it last quarter. The guide to OTE and sales pay plans and the AI sales pay calculator help you compare offers.

Questions to ask them

What is a qualified meeting here, and who decides?
How are leads researched and routed? Does a growth or GTM engineering team build lists, or do SDRs?
How many SDRs were promoted to AE in the last year, and on what criteria?
Which AEs would I support, and how do we review meetings together?

The SDR and BDR career guide explains why the promotion question matters: none of the postings read for it commits to a path.

Next steps

Build the research and sequence skills in the AI SDR outbound playbook.
Read what the role pays and owns in the AI SDR and BDR career guide.
See what happens after your meeting in the AI sales cycle playbook.
If you are changing careers, start with how to break into AI sales.