Symphony is your Conversational AI Recruiter (and my test interview with it)

Ewan MacLeod

Nov 8, 2024 — 6 min read

Now then, when it comes to using Conversational AI to manage and triage the candidate screening process, I think it's fair to say I wasn't necessarily sold on the concept.

The idea, yes.

The use of technology, yes.

But the candidate experience?

Are we at risk of making things even more frustrating?

Well... Not necessarily.

I say this because I just did an interview for a Customer Support Representative with Symphony's AI recruiter. This is the company's sample experience that enables you to try out the system from the candidate's standpoint.

I was very, very impressed.

I didn't think I would be. In fact, I mentally had my arms crossed across my chest as I pressed the 'try Sample interview' link on the company's website... but I thought I'd do my best to keep an open mind.

And I was surprised.

I enjoyed it.

I felt listened to.

No, really.

It's very smart what the team has built.

The system dynamically responded to what I was saying and then prompted me for questions that felt natural and easy.

I began to realise what the founders were meaning in their ycombinator interview thus:

Candidates tell us they feel more comfortable talking with our voice AI interviewer than humans. They are less nervous and can more openly share their experiences.

This reminds me very much of the Incio AI approach to using Conversational AI to discuss the sensitive topic of debt collection with customers who are behind in their payments – the Incio AI team highlights that customers prefer speaking to the AI as they don't feel as judged as they would be when speaking to another human, irrespective of how 'nice' that human is.

I got that when I was talking to Symphony AI. I felt listened to. Yes I know it's a system, but I was pretty excited to know that it was listening to me and transcribing what I was saying – so that a human (at some point) could look through my answers and potentially choose to go ahead with me. I was also hopeful that the answers I was giving would be looking good compared to others!

I liked the idea that there is no bias. It was listening, transcribing and then reacting to what I was saying – and giving me the chance to say my piece patiently.

I can imagine it being seriously frustrating if you had to do 10 of these a day when job hunting. But, again, the ability to essentially be able to get a fair chance at pitching myself? That is positive.

What's to prevent me hooking up ChatGPT to Symphony and having ChatGPT respond? Well, I do wonder about that: Are we at risk of having a rather crazy situation whereby your AI ends up speaking to my AI?

My simulated Interview with Symphony's "Eve"

Here's an overview through a few screenshots of my experience playing with the demo interview process. I decided to lean in and treat it as though I was applying for the Customer Support role!

So Symphony's questions appear on the right – they've chosen a clear, friendly US-American female voice (name Eve).

I got started by banging on about my most Chief Transformation Officer role for a few moments.

I took the above screenshot whilst it was still throwing up the next question. It's done what I'd expect and listened to my answer, distilling it down 'ChatGPT-style' to 'background in leadership and digital transformation'. Bonus points for being able to spell my name correctly.

It continued things by asking me specifically about my customer relationship experience.

That function reported to me in my last role ("Customer Experience") so I told it all about that – in Screenshot 3 above, you see it asking me for a bit more detail.

Here in Screenshot 4 above, it's asking me for a bit more detail in the context of the "ACME" company role. I was having fun talking away about myself and my experience as I tested it out!

I liked its immediate and positive responses – "that sounds like valuable experience" it tells me. Then it moved to ask me about software tool experience. I mentioned something about Microsoft Power BI.

I'm not too sure it was impressed by that one though:

Here it's asking if I'd be comfortable learning new tools – presumably the ones I failed to mention in the previous question!

Above in Screenshot 7, it's asking me about adaptability.

Again it's complimenting me. I like it!

Then it's moved to ask me about managing a 'difficult customer issue'. I talked about the CFO of one of the bank's clients that needed some immediate help.

Remarkable, eh? Thank you Symphony. I realised that I didn't actually explain the steps I took to resolve things - I'd rushed ahead, so notice how it's prompted me above to give a bit more detail.

I like how it's playing things back to move above - "sounds like" and so on. "That's a great approach' ("thank you").

Now for the famous 'where do you see your self in 5 years' question. I told it I'd want to be running the department...

I'm not too sure Symphony liked that though... am I reading too much into the briefer reply? ;-)

As for future goals, I explained how I was obsessed with great customer experiences - something like that, so above you can see it appeared happy with that.

Now, I realise I should have asked some questions about the role but whilst I was playing, I forgot to do this. It would have been useful to see what Symphony could have told me about the ACME (test) company and it's on-going goals and objectives.

And there we have it. Interview complete.

I wonder what my answers look like when they're 'written up'? How much editorialising does Symphony do? Or has it simply transcribed what I said and left it for the human to review?

Here's how the process works with Symphony:

We evaluate all inbound applications against the custom criteria for the role to surface the best matches. Our AI interviewer, Eve, then conducts live voice screening calls with top candidates - saving recruiters from spending their time on initial screens while removing human bias from the process. These 15-20 minute conversations bring resumes to life, highlighting important traits like communication skills and EQ (both of which are critical for customer-facing roles).

You can see a demo of the system in use here (including the back-end system interface):