How AI Is Helping Us Hire

Blog Post
Deals5

AI Hiring

AI Hiring

Hiring has been kicking our butt lately. We've been rebuilding our editing process and, honestly, it's taken way more time (and energy) than we expected 😅

But we're bringing you along for the ride and sharing the tools (cough cough, AI) that have been keeping us afloat.

If you've ever felt stuck trying to find the right person for your team, you'll probably find a few of these tips helpful
 or at least fascinating 🙃? Well, you tell us!

Phase 1: Faster Applicant Processing

Phase 1: Faster Applicant Processing

We posted the role on LinkedIn, Indeed, and a few niche job boards. We got 450+ applicants, and it only cost us $178.75 in ads. Not bad!

Embedded Image

LinkedIn Ad Spend

To handle the flood, we built a 50-question application (skills + culture fit) using ChatGPT to draft the questions and Typeform to host it (we already had a subscription, so why not).

We created a custom GPT to help score candidates, but ChatGPT actually kept hallucinating 😏 so we ended up turning to Claude which did a better job.

The whole system took one day to set up with AI đŸ”„

Interesting Insights

Interesting Insights

One of the coolest parts about running this questionnaire is the data we get back. A couple of things really stood out:

High Failure Rate

90% of applicants didn't pass the questionnaire, which saved us so much time and brainpower. We could easily spot problem areas before even hopping on a call.

AI Adoption

Even though more and more companies are openly saying "learn AI or lose your job," about 30% of applicants admitted they still don’t use AI regularly (or at all!).

Embedded Image

30% of applicants said they still don’t use AI regularly/fully trust it.

Accountability vs "No Biggie" Mentality

We asked: "What do you do if your manager says your work isn't what they wanted?" Only 40% said they take ownership early and reflect on what they could've done differently.

Nearly 50% went with the “no biggie, I’ll just re-do with feedback” option.

We’ve found the first group tends to improve faster (our motto: “questions up front = 10x better results”) so this narrowed the pool significantly.

Embedded Image

More than 50% of candidates said they're happier to re-do work rather than "take ownership and ask more questions up-front".

Candidates Thanked Us

And a fun surprise: a bunch of applicants actually thanked us for the process! Some even said they learned something new about themselves just by filling it out.

Phase 2: No "Gut Feeling" Interviews

Phase 2: No "Gut Feeling" Interviews

We've found the sweet spot is usually 2–3 interviews:

  1. A quick screening call.
  2. A longer in-person (or video) interview.
  3. A shorter wrap-up call if we're still seriously considering them.

Pretty standard
 except we've added one huge upgrade: Motion's Meeting Coach (our AI employee đŸ§‘đŸ’»).

If you're using Motion and haven't turned this on yet, dooo it! The Meeting Coach listens in, analyzes the interview (or any call), and then gives us feedback afterwards.

That "hmm, I think that went well?" feeling? It's now replaced with an actual score and breakdown.

For example, here's what it told us after a recent interview:

Embedded Image

We’ve found the feedback has been super helpful. We both find ourselves looking forward to the Meet Coach review, and and we learn something each time. It's actually making us better at our job.

Phase 3: Paid Trial (Resume = Reality?)

Phase 3: Paid Trial (Resume = Reality?)

Once we narrow down to the most promising candidates, we invite them to do a 3-hour paid trial project.

ChatGPT helped us draft the instructions, and now it lives in Superhuman Mail as a saved snippet (keyboard shortcut: ;trial). It takes two seconds to send out!

But here's the key: if someone performs well and we want to continue the process, we don't just look at the work. We also give clear, structured feedback and see how they respond. That's often even more telling than the task itself.

Here's a snippet of feedback we recently shared with a candidate:

Embedded Image

We give candidates clear feedback to see how they respond.

What we've learned:

  • Some candidates who looked incredible on paper actually struggled with tasks we thought they'd nail. HUGE reality check! It's shaped how we think about compensation.
  • Having multiple candidates complete the exact same task gives us a true apples-to-apples comparison.
  • How a candidate responds to feedback is just as telling than the work itself.

Phase 4: Backchannel References

Phase 4: Backchannel References

My fave part! Wade Foster (CEO of Zapier) recently shared on X that he does 20 reference checks for every executive hire, including backchannel references.

Now, we're not that intense (yet 😉), but his post inspired us. The most we've done so far is 9 references for a single candidate, which is still way more than we ever used to do đŸ©.

Here's how we run it:

  • Start with the 2–3 references a candidate provides.
  • Let them know upfront:
We also do backchannel references, which means reaching out to people in our network or LinkedIn who may have worked with you. ‍I wanted to be upfront about this and also check if there’s anyone you’d prefer we not speak to (for example, a current client where it could create an awkward situation). Please let me know and we will avoid reaching out to them.”
  • Then we dig into mutual LinkedIn connections or past managers for real insight.

What we learned:

  • Reference calls are time-consuming, but the payoff is huge. In a couple of hours, we learn more than we would in months of working with someone.
  • One backchannel call even revealed a candidate had fabricated his entire experience (and convinced coworkers to back him up 😯). We literally had an offer drafted, and a single 15-minute backchannel reference saved us big time.

Where We're At Now

Where We're At Now

We still haven’t found our person yet. But the process is doing its job: a few times we've gotten really close, only to realize (thanks to the steps above) that the candidate wasn't the right fit. Better to find that out now.

As a small team, this has definitely pulled me away from content and into full-on hiring manager mode. But this next hire matters too much to rush!

So thanks for sticking with us 🙏 Long-form video content is taking a bit longer, but as one of my favorite quotes says: “Sometimes you need to slow down to speed up.” That's exactly what we’re doing.

💃💃💃

Deal of the Month

Deal of the Month

Sign-up for this Ramp offer and get get $250!

We're so adamant that everyone on our team needs to be using AI that we give our team a monthly AI Budget to test new tools that help speed up their workflow.

How do we do that? We issue relevant team members a Ramp card with a $50–$150 budget (depending on their role) that they can spend freely. No reimbursements or headaches, Ramp makes it incredibly easy.

Embedded Image

We switched to Ramp six months ago as our business grew and our budget limits increased, and we've been blown away by their tech and product ever since. Let me tell you...the user experience is SOOOO good, truly best in class for expense management! Read our full Ramp review.