Job Interviews Step Into the Future with AI That Talks Back

The job interview is undergoing a fundamental shift: AI-powered screeners with lifelike synthetic voices are now conducting live, two-way interviews—no human recruiter required.

Startups such as Apriora, HeyMilo AI, and Ribbon are at the forefront of this transformation, reporting rapid adoption of their platforms for real-time, AI-driven video interviews.

These systems simulate a human recruiter, asking follow-up questions, evaluating key competencies, and generating structured feedback for hiring managers.

The goal is twofold: streamline the hiring process for employers and give candidates more flexibility—especially those in industries like nursing or trucking, where odd-hour availability can be a major advantage.

For companies hiring at scale, the appeal is obvious.

Running hundreds of interviews a day becomes feasible without overloading human teams.

Ribbon CEO Arsham Ghahramani, whose Toronto-based company raised $8.2 million in a round led by Radical Ventures, acknowledges the cultural shift:

“A year ago this idea seemed insane. Now it's quite normalised.”

At Canadian nonprofit Propel Impact, which plans to hire over 300 fellows this year, the move to AI interviews was driven by necessity.

The traditional approach—written applications and alumni-led interviews—simply did not scale.

Worse, ChatGPT was blurring the authenticity of applications.

Cheralyn Chok, Propel’s co-founder and executive director, explained:

“They were all the same. Same syntax, same patterns.”

The technology underpinning this shift has been in development for over a decade.

Platforms like HireVue introduced one-way video interviews in the early 2010s and later added automated scoring using facial recognition and language analysis—though visual analysis was rolled back in 2020 after criticism.

These systems, however, remained largely static.

Candidates recorded responses for later review, but real interaction was missing.

That changed with the rise of large language models like ChatGPT in late 2022.

Developers began building more dynamic systems that could actually converse—in real time.

Ribbon launched in 2023 and signed nearly 400 customers within eight months.

HeyMilo and Apriora followed closely, each experiencing fast growth, though specific customer numbers remain undisclosed.

HeyMilo CEO Sabashan Ragavan noted:

“The first year ChatGPT came out, recruiters weren’t really down for this. But the technology has gotten a lot better as time has gone on.”

Glitches Inevitable

Despite growing adoption, the rollout of AI interviewers has not been without hiccups.

A few viral TikTok clips have captured awkward moments—like bots repeating phrases or misunderstanding basic responses.

One particularly shared video showed an Apriora AI interviewer inexplicably repeating “vertical bar Pilates,” a glitch CEO and co-founder Aaron Wang attributed to a voice model misreading the word Pilates.

Wang noted the issue was quickly resolved and emphasized that such glitches are rare:

“We’re not going to get it right every single time. The incident rate is well under 0.001%.”

Propel Impact’s Cheralyn Chok also reported encountering minor issues during interviews, though it was not always clear whether the problems were caused by Ribbon’s software or unstable internet connections on the candidate’s end.

In those cases, restarting the session usually fixed the problem.

Meanwhile, Braden Dennis, founder of AI-powered investment platform FinChat, observed that AI systems can struggle when candidates ask nuanced or off-script follow-up questions—highlighting a current limitation in conversational flexibility.

He pointed out:

“It is definitely a very one-sided conversation. Especially when the candidate asks questions about the role. Those can be tricky to field from the AI.”

To address reliability concerns, companies behind these tools are investing in active monitoring and real-time support.

HeyMilo runs a 24/7 support team and uses automated alerts to catch issues such as dropped connections or missed prompts.

Ragavan stated:

“Technology can fail but we’ve built systems to catch those corner cases.”

Ribbon has a similar system in place—every time a candidate hits the support button, a notification is sent directly to the CEO.

Ghahramani said:

“Interviews are high stakes. We take those issues really seriously.”

While videos showcasing AI fumbles might seem damaging, Ribbon CEO Arsham Ghahramani views them differently.

He sees the mockery as a milestone: the moment AI interviews moved from novelty to cultural familiarity.

Getting Potential Employees Ready

FinChat, which uses Ribbon to conduct initial interviews, informs candidates upfront that they will be speaking with an AI—acknowledging that the experience may feel impersonal.

Dennis noted:

“We let them know when we send them the link to complete it that we know it is a bit dystopian and takes the ‘human’ out of human resources. That part is not lost on us.”

Yet, the asynchronous nature of the platform has clear advantages, including expanding access to a broader talent pool and reducing the risk of overlooking qualified applicants.

He added:

“We have had a few folks drop out of the running once I sent them the AI link. At the end of the day, we are an AI company as well, so if that is a strong deterrent then that’s OK.”

Propel Impact takes a similarly transparent approach, clearly explaining why AI is used in its hiring process and offering live, human-led information sessions to preserve a sense of connection and trust with candidates.

Chok expressed:

“As long as companies continue to offer human touch points along the way, these tools are going to be seen far more frequently.”

At the same time, regulators are paying closer attention.

While AI tools are often marketed as objective and equitable, concerns are growing over how these systems evaluate candidates—and whether they may unintentionally reinforce bias at scale.

Illinois now requires companies to disclose if AI is analysing interview recordings and to obtain candidate consent, while New York City mandates annual bias audits for any automated hiring systems used by employers.

These developments signal that increased scrutiny of AI in recruitment is not just inevitable—it is already underway.

Moving Past Basic Screening Calls

While AI interviewing tools are still primarily used for early-stage screenings, that is starting to change.

According to Ghahramani, 15% of interviews on the platform now occur beyond the initial screening—up from just 1% a few months ago.

This shift points to a growing willingness among employers to explore new use cases for AI in hiring.

Some companies are testing AI-led interviews to gather sensitive information, such as compensation expectations or feedback on the interview process—interactions that can feel awkward for both candidates and hiring managers, but may be easier when handled by a bot.

In certain instances, AI is even being deployed for technical assessments or to replace second-round interviews altogether.

Wang stated:

“You can actually compress stages. That first AI conversation can cover everything from ‘Are you authorized to work here?’ to fairly technical, domain-specific questions.”

Still, most vendors see these tools as decision-support systems rather than decision-makers.

For now, AI’s role remains focused on collecting structured insights, not making the final hiring call.

Ragavan concluded:

“We don’t believe that AI should be making the hiring decision. It should just collect data to support that decision.”