Fireflies Debuts AI Recruiter for Screening Calls (Full Transcript)

[00:00:00] Speaker 1: you you you you you you Hey everyone Looks like we have 36 people already We are expecting quite a few today. We'll start in just one more minute. All right, let's get started. Thank you so much, everyone, for joining us in this session, Build Your First AI Teammate Live With Us. I'm Nishant from Fireflies, my colleague here, Chan, with us. We are very excited to have you, and we're excited to tell you more about AI Teammates, which is the latest product on the Fireflies platform. And specifically for this audience, we're excited to tell you more about the AI Recruiter that's going to help you streamline your hiring processes and specifically automate your candidate screening calls. So yeah, super excited. Let's just introduce ourselves as well. I'm Nishant. I look after go-to-market and partnerships here at Fireflies. I've been at Fireflies for just over four years now, joined as a partnerships manager, and now I look after a bunch of different stuff, including go-to-market for this new product. My colleague here, Chan, is the senior product manager for integrations. So all of the cool integrations that you see on Fireflies, the 90-plus integrations, Chan looks after that, and then he also looks after AI Teammates. Chan?

[00:04:53] Speaker 2: Thanks, Nishant. Hey, guys. Excited to be here. Can't wait to show you all what we've built for you, especially the AI Recruiter. So yes, a quick intro about me. So Nishant already said, I'm a PM here at Fireflies, so we built something really cool. I've been here for almost like over a year. And yeah, super excited to get started and show you what we've built.

[00:05:14] Speaker 1: Awesome. Thanks, Chan. So what we're going to cover today, of course, a little introduction into Fireflies. I know a lot of you obviously already use Fireflies, you're already our customers. So we're happy to have you. We'll give you a brief introduction into Fireflies and why we built AI Teammates, what AI Teammates is and what it can help you automate. And then we'll go specifically into the AI Recruiter, what the capabilities are, what the features are, and how we can get you started with automating your candidate screening calls. And then we'll hand it over to Chan for a live demo to show you how to build the AI Recruiter within less than five minutes. And then we will show you a live demo of the AI Agent taking a screening call. We'll talk a little bit about what's next. And then we'll have a live Q&A session at the end. You can always leave questions here in the chat on the right side. We'll monitor those questions. We will answer them live if possible. We'll answer them in the chat if possible. If not, we will answer them during the live Q&A session, or we'll follow up afterwards with answers to those questions as well.

[00:06:25] Speaker 3: Awesome.

[00:06:26] Speaker 1: Let's get started with this webinar. So introduction to Fireflies. As I said, I know a lot of you are familiar with Fireflies. You use Fireflies for your note-taking, transcription, recording, et cetera, et cetera. Fireflies is a comprehensive AI platform. So far, we have been a note-taking platform, a platform that helps you record, transcribe, and take notes across all of your different meetings that you're having across different platforms, getting summaries, action items, as well as all the other good stuff that can also be sent into your integrations or integrated with your CRMs, your ATS systems, your project management tools, et cetera, et cetera. Now Fireflies is, of course, right now going through an evolution where it's not just going to be a note-taking platform, but a platform where you can deploy AI teammates to get work done for you, whether that is filling out your CRM, whether that is giving you notes, or whether in this case, it's going to be taking calls on your behalf. Over the past year, we've gone from Fireflies being just a note-taker or a passive participant in your meetings to an active participant in your meetings. Right now, you can talk to Fireflies in your meetings verbally by saying, hey, Fireflies, or you can talk to it in the chat going slash FF, asking any questions Fireflies during the meeting, making it an active participant. Right now, we're taking this one step forward and saying, hey, maybe it can also take your meetings for you. Now that's what Fireflies AI teammates is. AI teammates will be able to take redundant calls that you have on a regular basis, whether that is for you, recruiting calls, candidate screening calls, where you need to ask a set of questions and get some answers. You can, of course, review that transcript, review that call. You can get custom outputs from that, whether it's a scorecard, whether it's a recommendation, whether it's just a list of yes or no to the questions that you ask, et cetera, et cetera. Similarly, we have other AI teammates, such as the AI product marketer, the AI user researcher, AI sales rep, AI project manager. The sales rep can, of course, do qualification calls with your leads. The user researcher can do user interviews with your users. The project manager can do update calls with your team. Now here's where the interesting part comes. The AI recruiter is actually one of the first AI teammates that we've launched as a flagship product. And why we did that is we saw a huge gap in the recruiting segment. All of you are either working at a recruiting agency or you're a recruiter at your company. And you know the amount of time that is spent on admin, scheduling, screening, follow-ups. 52% of time is spent on admin. 70% of screening calls lead to rejection, but they still take 30 minutes, at least 10 to 15 minutes, up to 30 minutes each, sometimes even higher. Due to this, a lot of times top candidates drop off, or you're spending most of your time not evaluating the top candidates, but screening the candidates that will not go through the process. So less than 20% of recruiter time at the end of the day is spent actually evaluating the top candidates where you should be spending most of your time. And that's what we want to solve by automating your screening calls. So there has to be a better way, and that's where our Firefly's AI recruiter comes in. Now what is the AI recruiter? The AI recruiter is an AI teammate that will automate your first round of interviews or your screening calls that you have with your candidates. When you're just asking questions, learning more about the candidate, screening them to see if you can move them forward down the pipeline in your hiring pipeline. We can automate structured screening interviews. The structure is decided by you, but it also follows the best practices in the industry. So it will ask questions. If it receives a vague answer from the candidate, it will also ask follow-up questions or try to dig in more to get the information that you really need. In order to customize this to your company branding or how you would like it to act, you can customize the voice, the tone, the intro, however the agent should act on the screening call. Of course, the AI recruiter needs to have context about your company. So you can also add your company knowledge base as well as the job description that you're hiring for so that the agent has full knowledge and full context while it's conducting that screening and in case the candidate asks them any question, for example, what is the culture like at X company? Or what is the base pay? It can answer those questions according to how you want them to answer it. So that is the company knowledge. And then obviously you'll also be able to add the questions that you want to ask for each role. As I said, you can just add the questions and it will take care of the follow-ups and getting the answers by itself. After the call, you will instantly get summaries and other outputs. So just like you do on Fireflies, you'll get a summary, a transcript, a recording, as well as all the other AI filters and AI outputs, but you can also generate custom outputs using AI apps. AI apps on Fireflies allow you to create and generate custom outputs for your meetings. In this case, it could be a scorecard, it could be a recommendation, it could be a list of pros and cons of this candidate, et cetera, et cetera. And AI apps can help you get that custom output that you need in order to make the right decisions. So it's not just about taking the call, but also getting the insights and the intelligence from that call to enable you to make the right decision from that information. There is no scheduling required. So this saves you all the time that you spend on scheduling. All you have to do is share a link with candidates and candidates will be able to take the screening interview at their own convenience or at a convenient time. So this saves you not just the back and forth of trying to schedule or trying to do the screening interview, but it also means that now you are not capped to the number of interviews that you can do. We've actually seen all of our early customers are able to screen at least twice as many candidates in a week as they were able to do before. So there is just three steps in order to set up your own AI recruiter for the roles that you're hiring for. The first step is, of course, build and deploy the AI recruiter. All you need is three things, questions, knowledge, and instructions. We even give you a pre-filled list of roles that you can choose from. If you don't find the role that you're hiring for in that list, you will be able to enter a custom role and we will auto-generate that AI recruiter for you. So it will auto-fill the questions, the instructions, and the knowledge base for you. And Chan will show us this in a moment in the demo. And of course, then you can customize the questions, the knowledge, and the instructions depending on what you want the recruiter to ask, what you want the recruiter to act like, and what you want the recruiter to know. As soon as that is done, all you have to do is share the link and now your job is done. The candidate interviews will happen async. You don't need to follow up with them. They will take the screening at their own convenience. And as soon as that screening is done, you will get the insights, the summaries, the recording of that call, as well as the custom outputs that you want from that call. Now all you need to do is filter and decide from that list of screening interviews, not spend hours running those screening calls yourselves. Our early customers are already cutting time to screen by 70%. So what I mentioned before was you're able to screen at least twice as many candidates already. And right now you're seeing our early customers cutting the time it takes to take a good candidate from one stage to the next by 70% because of the automation that you're doing. All right. I'm going to hand it over to my colleague Chan here, and Chan will be able to walk us through a quick demo of how we can build this within a couple minutes and get started. Absolutely.

[00:15:08] Speaker 2: Thanks, Nishan. Let me just quickly share my screen. There we go. Yeah. So before I get into the demo, I just wanted to give some quick background on... I just want to rewind the clock a little bit, a couple of months back when we decided to build this entire initiative. We spoke to a lot of folks from Recruiting, we spoke to our own talent acquisition team and a lot of customers these days, we are in touch with them on a daily basis. So initially we understood the problem that they have. We wanted to completely automate the screening rounds and a lot of customers even use it for the second and third rounds and even the final rounds in certain cases. But at the same time, what we did was before we launch any product, we have a habit of dogfooding it internally, meaning that we would test it with our internal team and only when they see value, we even surface it to an external customer, even if it's one customer. And what we did was we gave it to our talent acquisition team, we asked them to just start using it, give us good feedback on top of it. And it started off with them automating one low lift role that we currently had open at that time in Fireflys and over a week, it slowly became two roles. Then within a couple of weeks, we almost had all roles that are currently automated by the AI recruiter. So right now, every single role that we see in Fireflys that's open right now in our careers page is currently, the first round is currently automated by our very own AI recruiter. And that's when we started having a beta program, we started shipping it out to customers and we are seeing tremendous feedback. We have a lot of customers, you can see a lot of success stories coming up in social media real soon and the feedback that we've got from customers is amazing. So without further ado, let me just get into the live demo. So all of you, you might see this AI teammates section in your Fireflys app, so you can find it on the left nav, you can see AI teammates. So under AI teammates, you have multiple teammates that you can choose from. We'll soon bring an option where you can create a very own custom agent from scratch, but to get things easy for you, we have six agents that we pre-ship out of the box. You can create multiple agents from these templates. So starting with AI recruiter, which we will dive deeper, but you have colleagues in your company where you can also refer the other agents as well. We have a user researcher, a project manager, a product marketer, an AI manager, an AI sales rep that can completely automate the first round discovery calls as well. So that being said, let me just go into the setup of an AI recruiter. So let's say I have a role in Fireflys that I'm planning to automate using my AI teammate. All that I need to do is go inside my AI teammate, the AI recruiter template, and we have something that's an industry first, we make it super simple for you. As in, all that we do is tell us what role you're hiring for, and we'll set up an agent for you in less than a few seconds. That's something that's industry first. So here, what Fireflys automatically does is, based on the company email that you've signed up to Fireflys, it automatically fetches the open roles in your careers page. So you can see that all these roles are actually roles that are available in the Fireflys careers page. So like I said, I want to hire for the product manager platforms role, and all that I need to do is select that role and click autofill. So you can see there are three aspects of the agent, the questions, the knowledge base, and instructions. You can see that Fireflys is automatically cooking up the questions, the best practices of questions, the knowledge base, and instructions, and within just a few seconds, you can see that it's automatically filling up that data for me. So all that we do is select the role in Fireflys and create the questions that it needs to ask. Basically, at the names, I guess, questions is basically the questions that the agent will ask the candidate. Knowledge base is basically company information on your JD. So Fireflys will automatically combine information about your company and the job role and add this knowledge base. So let's say the candidate asks the agent a specific question, for example, hey, how does my PTO policy look like? And Fireflys will use this information to respond accurately on top of that. We also have instructions. So Fireflys AI agents or AI teammates are already hard-coded with the best practices in industry for every single vertical, like recruiting, sales, et cetera. But let's say you have very specific instructions that you need the agent to follow. All they need to do is specify it here and Fireflys AI teammates will follow it right for you. For example, hey, greet the candidate in such a way, end the call in such a way, or anything of that sort. So here we go. I have all the details that I have right now to create my AI teammate or an AI recruiter. I can just click create and I'm pretty much done. But let's say you need to modify it, or let's say you did not find the role that you're hiring for over here, we ship 200 plus common roles right out of the box here so you can choose from. And even then, even then, let's say you cannot find the role that you're hiring for, all that you need to do is just type it out, let's say, technical account manager, for example, let me just go from scratch, just type it out, technical account manager, and Fireflys can automatically, you can see that it says can't find this guy, I can say technical account manager, and I can click autofill and Fireflys will automatically use the role that you just inputted to Fireflys and also take information from your company, search internet, and automatically create that AI teammate for you, you can see that they're automatically creating it. So you can do either of these three things, or let's say you need to modify it from scratch, let's say you don't want the autofill feature, you want to build your AI teammate from scratch, you can pretty much enter all of these manually, modify it, totally up to you. So it's highly customizable. And the other aspect that I wanted to talk to you about is the customization that we provide for the AI teammates. We have an option where you can upload your own company logo, a lot of recruiters prefer that, you can give your AI recruiter a custom name, let me just add something here, and you can also modify the company name, for example, you can add your company name, for example, at me and when users or candidates see the link to talk to the agent, they can see your personal branding. And here's the special part, we have tons of customization options with respect to the voice. So we have seven different voice options and 74 different languages that you can choose from, 74 different languages. So let's say you're hiring all over the world, you can just basically create an AI recruiter for every single language and you can have it sent across to candidates across the world. So not only that, we also have accent and dialect options for the English language, which is pretty amazing, you should try it out. When you click preview, you can actually check how the voice sounds so that you ensure that the best voice is suited for the candidates so that they can get the best experience going forward. And here's the best part. So you have the session duration, you can control the duration of the meetings. It ranges from five minutes to 40 minutes. So mostly from what we've learned from customers is that the recruiting calls take not more than 25 or 30 minutes. So I can just specify 30 minutes here and the AI recruiter would gracefully end the call. It's not going to just abruptly say, it's time, I'm going to end the call. No, it's going to say, Hey, it's almost time for us to wrap up. Do you have anything else or can we just end the call? So it's going to be as human and as natural as possible, which you will see soon. And yeah, and you can also choose the tone, basically set the tone for the agent. It's something that we have such as casual, professional or talkative, we'll add more tones as in when we see more demand, but you can specify the tone for the AI teammate and you're pretty much done. We also have some additional settings such as you can actually have your entire HR team in Fireflies take a look at this agent by enabling the visibility, set meeting privacy options and also set an email recap setting and pretty much you're started. Just click create and you are done. So what happens is right after you click create, you get a link to the agent. So this is the link. So like I said, this is like Nishant rightly said, that it is going to save 70% of your screening time. But at the same time, it also gives your team a superpower. Basically, this is like Batman for recruiters. So the agent is available 24 seven. So just have this link sent across to your candidates via email, slack or any medium of your choice. And anytime that they want to have a call with this recruiter, they can just go to the link into the details and just have a call just like a normal person would. So you can see that I have two options here, which is copy link and preview. So preview is basically going to take you to the entire call experience so you can go through it first hand so you can decide if things need some tweaking or things like that. But I'm just going to copy the link and I'm going to show you how does the candidate experience look like. So there you go. You can see the logo, the company name and the agent name and all of those details here. So we made it super seamless for candidates to join the call. As in, all they need to do is enter the name and their email. And what we do is we send a one time password of six digit code so that they can authenticate themselves. So we figured out that this is the most secure way rather than asking them to log into a platform that they don't even know about or something that they may feel that it might be weird. So it's good to have them securely authenticate, which is basically to enter the name and email. So the moment I click continue, I get a one time password to my email. So just checking for my email. Yeah, there we go. So after I authenticate, that's pretty much it. All that I need to do is, it takes a few seconds for Firefly to initialize the agent, get your agent ready and bring it to a meeting. So that takes just less than five to six seconds. And right after it's done, when you click start meeting, you can immediately start talking to the agent. So yeah. So the agent joins the call right now in Google Meet, we're also expanding to other platforms and also our proprietary audio platform, which will come real soon. But right now it's available on Google Meet, which is a platform that everybody knows and a lot of people are familiar with the interface and also delivers a much superior experience in Google Meet as well. So let me click start meeting. And you can see that the agent is already in the call. So it's ready for me to join and talk to it. So let me just quickly re-share my screen so that you can hear my tab audio as well. Sharing my screen. There we go. So yes, let's take a call, a live call with the AI teammate and see how that goes.

[00:26:44] Speaker 3: Hi Chantanu, it's really nice to meet you.

[00:26:54] Speaker 1: How's your day going so far?

[00:26:56] Speaker 2: My day is going fantastic, how about you?

[00:27:00] Speaker 1: I'm doing well. Thank you.

[00:27:02] Speaker 3: This will be a structured conversation where I ask a few questions to understand your experience and fit for the role. Feel free to take your time. Just let me know if anything's unclear. Sound good? Yes. Describe a time you had to manage a technically complex issue for a client that was causing

[00:27:21] Speaker 1: significant disruption.

[00:27:23] Speaker 3: What steps did you take to diagnose, resolve, and communicate the resolution to the client and internal stakeholders? Okay.

[00:27:30] Speaker 2: So there was this one time where there was an issue with the client with the Firefly HubSpot integration. So the integration was not creating tasks as expected in HubSpot right after a meeting is done. So we have a feature where after a meeting is done, Firefly automatically creates action items for the users in HubSpot. Then what we did was we did a quick troubleshooting and found that they had enabled a team integration and somebody else in the team had disabled that feature. So what we did was we had this communication.

[00:28:02] Speaker 3: Could you elaborate on the specific steps you took to resolve the integration issue and how you communicated that back to the client and your team?

[00:28:09] Speaker 2: Yes. So, yeah, like I said, that's pretty much it. So we immediately troubleshooted it and, you know, we had that communicated with the client and that was some of the good experience that we've had and it was all good right after

[00:28:25] Speaker 1: that.

[00:28:26] Speaker 3: Got it. How do you stay current with the rapidly evolving AI landscape and new product developments, particularly in areas like natural language processing and meeting intelligence? Got it.

[00:28:37] Speaker 2: So I listen to a lot of podcasts in this area, especially in AI, and I also subscribe to a lot of newsletters. I read a lot of articles to keep myself updated and I'm always on top of the news. So yeah, these are some of the mediums where I use myself to keep myself updated.

[00:28:57] Speaker 3: Can you share a bit more detail on a specific podcast or newsletter that's been particularly valuable to you and how you apply what you learn from it?

[00:29:04] Speaker 2: Yeah, got it. So one podcast or a newsletter that I love is Lenny's newsletter. So I read a lot from Lenny's content. So a lot of growth-related mindset that I currently have is from Lenny's and which is something that I really love. So I'd probably say Lenny's for sure.

[00:29:27] Speaker 3: Imagine a key client is experiencing performance issues with our AI transcription service and initial troubleshooting hasn't identified the root cause. Walk me through your process for escalating and resolving this, ensuring client satisfaction.

[00:29:39] Speaker 2: Okay, so 11.1 is something that I would try to troubleshoot. But at the same time, if I'm something that I'm not able to do it, I'll probably look at my engineering team. But at the end of the day, we have a very strict SLA for our clients where we want to get them resolved and up and running as soon as possible. So it's less than a couple of hours. But yes, we want to ensure that the communication is spot on and also the time and SLA are respected. But hey, it was great talking to you. That's all I had for you today. But yes, thank you for your time.

[00:30:17] Speaker 3: I'd like to dive a bit deeper. Could you walk me through the specific steps you take during the escalation process with the engineering team and how you'd keep the client informed?

[00:30:26] Speaker 2: Yeah, so this was just a demo and I'm going to drop off, but it was great talking to you.

[00:30:38] Speaker 1: Thanks.

[00:30:40] Speaker 3: Understood. Just to confirm, before we wrap up, do you have any final questions or anything else you'd like to cover?

[00:30:47] Speaker 2: No, that would be all. Thank you so much. Bye. Bye.

[00:30:52] Speaker 3: Perfect. Thank you so much for your time today. It was a pleasure speaking with you. I'm ending the call in a moment. Take care.

[00:31:03] Speaker 1: Yes.

[00:31:04] Speaker 2: So let me just quickly share my screen again. So if you had noticed, when I was able to not evade the answer at the end, it was able to drill deeper, just like a real recruiter would. These are things that you can automatically configure and tweak as part of your instructions. So I'm going to quickly share my screen again, and here we go. So we've seen the part where you can create and also a candidate experience on the specific of missing users, having problems listening to the agent. Is it just Stella or everybody else can hear the agent? Conversation up to this point. Okay, got it. No worries, Stella, after the call is done, we will share the recording with you. But anyways, okay, folks. So we saw the experience where we would create the agent and also experience where users can talk to the agent. And what's most important is what happens after this call. So that's where the decision making process lies in. So as you can see, right after a call is done, you get a complete summary of that specific meeting, just like how you get your Firefly's notes. You can see that the template right now is generic hiring. So the meeting that you're seeing here is an AI teammate meeting where the AI teammate has conducted the interview. I get a neat summary, the entire transcript, and you can also use AskRed to ask specific insights about that meeting. And all of these fall into your Firefly system neatly so that your post-meeting workflow remains intact. I wanted to show you something really exciting, which is AI apps. So here is an actual call that we have here in Firefly from a real candidate. I can't scroll up and show you the name, but I have the AI apps here where we use a scorecard. So the scorecard is to give us an insight on how did the candidate perform in that specific screening round. So you have a scorecard that you can create. You can create custom scorecards for very specific for your company. And you can see that it's created scores for every single aspect. For example, this is a screening agent for a designer role, and you can see that it's given a specific score for every single aspect such as communication and delivery, how is the craft and precision for the specific candidate, and also gives you an overall score at the end as well. So here's the best part. We have tons of AI apps that you can choose from. So you can see that in Discover, you have multiple categories such as sales, recruiting, hiring and whatsoever it may be. For example, let me just select recruiting and tell you that we have tons of apps for recruiting as well. We have interview screening, intro scorecard, sentiment, insights, behavioral questions, scripted response. We also have cheat detection, which is a very core use case that our customers ask for. It's a very popular thing that customers ask for. So we have AI apps to detect if the candidate is giving you a scripted response. We have 10 plus apps for cheat detection, prohibited words, vague answers, contradicting answers, LLM responses, claim credibility, and a lot of other things as well that you can automatically enable right after for your AI teammate calls. And finally, the icing on top of the cake is integrations. So as most of you know, Firefly has 95 plus integrations that includes ATS integrations as well. We connect with the top four popular ATSs, Ashby, Bamboo, HR, Greenhouse, and Lever. But at the same time, we understand that the ATS market space is huge. There is always something that we will not integrate with. But at the same time, if there's something like that, we also have for AI teammates, what we do is, if they're willing to buy AI teammates in good volume, we always offer them integration with that specific ATS out of the box. So something that we can build for you in less than two weeks. But at the same time, if you want to do the automation yourself, we have integration with a lot of connectivity platforms, automation platforms like Zapier, Make, where you can plug and play right away starting now. For example, we have Relay, we have Make, we have Lindy, we have Zapier, and whatsoever that you might need. So right after a call is done, you can get your notes. So if you take the native integration that we have for ATSs, Firefly acts in an agentic way where it creates candidates for you in your ATS, maps the scorecard for you from your AI apps, and also sends the notes for you in that specific candidate profile as well. So you can see that with the AI teammates flow, you're almost automating the entire process. All that you need to do is send the link to your candidates and you're pretty much sorted. So here you go. Right after you send the link to a candidate, the candidate takes the call of the AI teammate, they have the best experience, and you get the notes, summary, etc. And then you have an integration connected where it creates a candidate for you, maps the notes as well. So this is what I had to show you today. I can't wait for all of you to try it out and let me know what you think. But before I quickly wrap up, I want to show you something, which is a quick wall of love that we recently made, where we have captured all the experience from our customers and also our internal candidates as well. So you can see that 95% of the candidates have responded positively. Some of them have even gone one step further and they've posted in social media, as you can see from the first one, that they've had an amazing experience. So this is going to give a wow factor to your candidates. It's fairly new tech, but you can see that we have five star ratings for most of the interviews that we've had and customers have had. So what you're seeing is actual responses that we have received from candidates. But yes, you get the image. So looking forward to seeing you all try AI Team Edge for your very specific use cases. But yes, let me know if you have any questions. That's all from my end. Thank you.

[00:37:54] Speaker 1: Thanks so much, Chen. It's super exciting. Awesome. Well, that was a demo of how you can create the AI Recruiter, as well as how you can send it to candidates, what the candidates, the experience that the candidates will see, and then what you get afterwards. If you want to try this out right now, it's available on your Fireflies on app.fireflies.ai slash agents. When you go on app.fireflies.ai slash agents, you'll see that there is an option for AI Recruiter. So if you go to AI Recruiter, you can either select a job role that already exists, or you can create a custom job role and have it create a AI teammate or an AI recruiter for you. You can reuse it, you can customize it, and then you can test it out for yourselves. Most of our customers like to test it out for themselves, and then they roll it out to at least one job role that they're hiring for. And then of course, they learn from that and then scale up from there and roll it out to

[00:38:56] Speaker 3: multiple job roles that they are hiring for.

[00:38:59] Speaker 1: If you have any questions, please feel free to reach out to us. You can also reach out to Chan or me directly. We'll leave our emails here. And obviously, we'll also send you a follow up email tomorrow with the recording, as well as other materials for you to get started. Now let's just go through some of the live Q&A that we've had. So Chan, can we add PDF documents under the knowledge base? Fantastic.

[00:39:32] Speaker 2: So this is a feature that we were just building, and we are shipping it later this week. So not only just PDF, we'll also support text documents, we'll also support Word documents. And the best part is we'll also support links. So let's say you have a knowledge base from your company, you'll be able to upload that to your AI recruiter, so that it always has the top context when it's responding to candidates shipping this week, later this week.

[00:40:00] Speaker 1: Yep. So right now you can copy paste as much information as you want. We have PDF uploads and links coming very soon. All right, next question is from Jody. If we have an initial call with the hiring manager, and you decide with your hiring manager that these are the nice-to-haves and must-haves for the person that you're about to hire, can that be included as something to be reviewed in determining the qualifying questions? So when we generate those questions, can we use that information?

[00:40:36] Speaker 2: Chan? Yes. So you can do anything that you might want with the instructions that we have. So it's totally customizable, like I said. So for example, like I said, whatever you feed into the agent is something that it is what it will use to behave. So all of these are totally possible. And all that you need to do is decide the way that you want to input that as a text or a data. And you can see the agent can behave in that way as expected.

[00:41:09] Speaker 1: Yeah. And Jody, you can also reach out to us. If you have a list of nice-to-haves and must-haves for a specific role, we can help you turn that into an agent with the right questions and the right instructions as well. So feel free to reach out to us, and we'll be happy to help you there as well.

[00:41:27] Speaker 2: Awesome. Next question.

[00:41:32] Speaker 1: Can the agent pick up details like pronunciation? We interview candidates who are not native speakers, and we also need to assess their English skills. So it cannot pick up pronunciation per se, Sebastian. It can pick up communication. So a lot of the times we're also interviewing candidates, not just the pronunciation, but how they communicate in general. If you look at their English skills, then it will be able to pick up details on how their verbal communication is through the transcript itself. We don't do voice processing right now, but we do process the transcript, which will give the AI a sense of the communication skills. In fact, we also include this in a lot of our own hiring, where we ask it to assess their communication skills.

[00:42:20] Speaker 3: And most of the time, it's very accurate.

[00:42:24] Speaker 1: All right. What if the candidate doesn't understand the question? Will the AI recruiter rephrase it? If the candidate goes off and starts answering something else, the AI recruiter will go back to the question and restart that. But if the candidate says, hey, can you repeat the question? Can you rephrase the question? It will act just like a human recruiter, and it will either repeat or rephrase or ask the same question in different terms. So yes, in that way, it does behave like very human-like.

[00:42:58] Speaker 2: Yeah, so just adding on top of what Nishant said, the experience is as human as possible. Let's say the candidate says, hey, just give me a minute. I'm just going to grab something. Somebody has the door or something like that. The AI teammate would wait until the candidate is back so that it's as natural and human as possible. So every day, we keep pushing the engineering team to ensure that this experience is as human with near zero latency as possible. So it's only going to get better. This is new tech, as you can see, and it's just going to get better and as human as possible. But 95 percent, it's already there to give your candidates the best experience.

[00:43:34] Speaker 1: Awesome. Does AI Recruiter integrate with Rippling? So today we, yeah, go ahead, Shaan.

[00:43:41] Speaker 2: Oh, sorry. So not natively, we don't have an integration with Rippling right now, but that's something that we've explored. But like I said, there are two options for you. You can use Zapier, or if you're willing to try out AI teammates as a customer, we'd be happy to build an integration for you in less than two weeks.

[00:43:56] Speaker 1: Yeah, if you end up liking the AI Recruiter and you say, hey, I want to use this for X number of roles or X number of interviews, you can always reach out to us and we'll help build that integration. We are constantly shipping integrations at Fireflies. Right now, we have four ATS platforms that are supported and a number of others in the in the roadmap as well. So, yes, and you can definitely reach out to us if you're interested in integrating with Rippling. Where can we find the pricing for the recruiting agent? So, yes, we did not touch on pricing, but basically the AI Recruiter as a product is included in your Fireflies platform or your Fireflies account, but it's charged on a usage basis. So we don't want to overcharge you, undercharge you, nothing like that. It will charge you exactly how much you're using the AI Recruiter and it's charged on a permanent basis. So however long that interaction between the agent and the candidate is, it will charge you AI credits. Some of you may be familiar with AI credits if you've used things like Ask Fred or used AI apps on Fireflies. So AI credits will be used here as well. There's no separate subscription. The AI credit subscription will be used. The pricing is one AI credit per minute, which comes out to be at around six dollars an hour. So let's say one recruiting call lasts 10 to 15 minutes. So that's around one, one point five dollars per screening call that you're doing, which is, of course, much more affordable than spending your time, spending one person's time doing those screening interviews as well. You'll find AI credits pricing on our website as well as on our helpdesk guide, guide.fireflies.ai. When will it work with MS Teams? So we have decided right now to have this on Google Meet as a native experience because it's very easy to use. Most people are familiar with Google Meet and it has much less restrictions than platforms like MS Teams. However, we are looking at deploying this on other platforms in the future as well. Shan, anything to add there?

[00:46:23] Speaker 2: Yes. So we want to make the entire experience platform agnostic. So Google Meet is our initial B0 or an MVP version. We are looking to have a native Fireflies version real soon, which is going to offer a much superior experience with respect to voice. But yes, Google Meet will have its advantages with respect to video and other video conferencing platforms like Zoom and MS Teams are on the cards. But as you might know, there are rate limitings and API limitations that we have to come across when we want to deploy our agents into these platforms. So Google Meet so far has been pretty friendly for us with respect to the API policy. So that's why we decided Google Meet is the right way forward for an initial version. But as and when we navigate through demand, we'll see how we can deploy this to other video conferencing platforms like Teams and Zoom.

[00:47:16] Speaker 1: Yeah, we would suggest start using it on Google Meet itself. It's super easy for yourself as well as for the candidates. But with MS Teams and other platforms, they will definitely come up soon. All right, next question from Grayson. If we include in the instructions to wait before the agent responds, will the agent abide by that? So if the candidate responds, can we ask the agent to wait five seconds before responding? It should follow the instructions. We haven't tested it out in that specific scenario. Right now, as Chan said, we want the most human like experience. So we've reduced the latency as much as possible so that there's no waiting in between because that degrades the experience a little bit. But this is if you add it to the instructions, it should be supported. I think, Grayson, your second question was about Notion pages. Right now, we'll just need you to copy paste them. But as Chan said, you'll very soon have the ability to either upload them as a PDF, as a document or add them as a link in the knowledge base as well.

[00:48:31] Speaker 2: So here's a quick tip for you. Let's say you make a Notion page public and add it as a link. You will be able to do that. So but if you don't want to make it public, you'll have to copy paste that for you now. Or you can just export a Notion page as a PDF and just have it uploaded for now.

[00:48:50] Speaker 1: Yeah, the best bet would be to copy paste or upload. Yes. If we upload documents for the knowledge base for one role, will it apply to other roles too? No. So you'll be basically creating one AI recruiter for each role and each AI recruiter will have its own knowledge base. And that is because obviously for each role, you'll have a different job description, maybe different requirements, maybe different questions and answers for the candidate and the agent. So there is a separate knowledge base for each of them. You can, of course, include the same information in both. But we do have a separate knowledge base that Chan is going to tell us about that can connect to both both AI teammates or AI recruiters or multiple AI recruiters.

[00:49:34] Speaker 2: Yeah. So even now today, you can find something called knowledge base in your settings. So what knowledge base does it let you create templates for different use cases? For example, sales, recruiting, you can create one single knowledge base for all your roles or create specific ones dedicated to a specific recruiter and feed information to it. It's totally up to you. We have the flexibility to do that. So if you want one knowledge base to be shared across to your all agents that are currently active for you, can be done. If you want individual ones that needs to be connected to specific agents, can be done.

[00:50:16] Speaker 1: Awesome. Yeah, so there is a separate knowledge base that you can connect and that we would, again, suggest only to keep the company information, the general information that applies for all roles in that general knowledge base, but for the knowledge in the agent itself, that should be role specific.

[00:50:36] Speaker 2: OK, I'm going to take the next question, which says, I also use Fireplace for note taking during the vacancy intake meeting with the hiring manager. I can also be the recruiter. I think we also already answered this question. I'm just going to click finish. Next one, apart from the usual interview questions, is it possible to do technical assessments such as tools navigation wherein candidates need to share their screen and navigate based on instructions provided? OK, so we really don't have that level of control for now, but you can add that as part of the instruction and you can review the video later. So we have video detection at the moment. So you can have, you know, basically understand just binary at the moment. So where the AIT can understand if the video is turned on or not and not decide to proceed if the video is not being turned on. So that's pretty much the capability that we have with video. But immediately, what we are trying to do is to have cheat detection and video analysis. So once we have video analysis in place, you'll be able to satisfy that use case for sure.

[00:51:46] Speaker 1: Awesome, I think we covered most of the questions. Of course, if you have any other questions, please feel free to reach out to us. We hope you enjoy using AI teammates. We're very excited to see how you use the AI recruiter, how you like using the AI recruiter with your candidates, which screening calls you're running, et cetera, et cetera. We will, of course, drop you a follow up email with the demo recording, as well as some questions and answers, and then we'll take it from there. Thank you very much for joining us on this session, and we hope you have a great day. That's a wrap.

[00:52:30] Speaker 2: Awesome, fantastic talking to you all. Thank you so much. Excited to see what you do.