How SalesMind AI Gets 42% Reply Rate on LinkedIn x Work Optional | Part 1

Transcript 0:03 Great. It's Saturday, June 7th. I'm here 0:06 with my friend Julian Gadilla. Very 0:08 excited because Julian, we've been 0:10 trying to book this uh conversation for 0:12 some time now. Uh so Julian Gadilla is 0:14 the uh CEO of a company called 0:17 Salesmind. Amazing background. He's 0:19 French but study and worked in China and 0:22 now he is having his uh fast growing 0:26 business out of Thailand. So uh welcome 0:29 Julian. Uh work optional mastermind is 0:32 really about uh learning and 0:34 interviewing folks such as yourself that 0:36 are building uh AI applications that are 0:39 going to transform the future of work. 0:41 So today if you if it's okay with you 0:44 love to find out more about your amazing 0:45 background uh want to pick your brains 0:48 about how you see AI is transforming 0:51 business and work uh sort of where the 0:53 trends are going. Uh then certainly uh 0:56 have you tell us what uh salesmind is 0:59 doing. I know that you you're getting 1:00 amazing traction. My understanding is 1:02 your your sales to has sent some 3 1:05 million messages on LinkedIn and your 1:07 response rate for your clients is more 1:09 than double the traditional 20% is what 1:11 I'm learning. So I'd love to learn more 1:12 about that. Um how does that sound? 1:15 We'll cover those three topics. 1:17 That sound amazing. Really excited to 1:19 get started. 1:21 Okay. So we're both in Bangkok, but 1:23 you're French and you study in China. 1:25 How tell us more about Julian. 1:28 Yes, sure. So Julian always wanted to 1:31 travel basically. So the the university 1:34 offer me a possibility to study in China 1:37 and I know that that country is really 1:39 fast growing. They have rapid innovation 1:41 and they are leading the innovation. So 1:45 doing computer science um as as study 1:48 was like going to China to study this 1:50 field was actually a very very wise 1:52 choice um looking back now. So I've been 1:56 studying China for for several years and 1:59 um and my background in France lies on 2:03 and fas on ethics data protection 2:06 regulation compl compliance and 2:09 everything. you know the the the 2:10 Europeans they like to put low before we 2:12 find use cases. Um so the the blend of 2:16 both of them is actually look like 2:18 Thailand to me. So I land in Thailand 2:20 two or three years ago 2:23 and um and then this is where I 2:25 identified the opportunity that um led 2:28 me to build salesman actually but I 2:30 guess we're going to get into that. 2:32 Where did you study in China and how 2:34 many years and was it a a master's 2:36 degree in comsai? 2:38 Yes, it was a master. I I did my third 2:41 and fourth and fifth year back in China. 2:44 It was in Tanin, Beijing. And I and I 2:47 finished my my time in China in 2020 2:50 right before COVID actually. And I was 2:52 in 2:52 Wow. Wow. And was it in Chinese or is an 2:56 English program? 2:58 No, it was in English, but I I do learn 3:01 Chinese. 3:05 Oh, okay. Okay. Fantastic. Did you 3:08 launch salesmind immediately or did you 3:10 do something else and how did you start 3:12 salesmind again? Then tell us about 3:13 salesmind. Yeah. 3:14 So no I haven't learned sales mind right 3:16 after my study basically I've worked in 3:19 several companies before big corporate 3:21 and always as a project or product 3:24 manager. My my skills and passion lies 3:28 in product themselves. So I've been 3:31 building products for businesses um for 3:34 the last five years before I started uh 3:36 salesmind basically. And then at some 3:39 point um it led me to identify uh a gap 3:44 and problem in the sales software 3:46 industry where the sales rep do still a 3:49 lot of things that is manual and where 3:52 the 3:53 automation where promising scale it just 3:56 bring actually spam. So we identified 3:58 that the AI will actually be much more 4:00 effective in some part of the sales 4:03 funnel and uh and in the lead generation 4:06 uh sphere basically where where it will 4:09 be the easiest part to leverage the 4:12 capacity of AI. So this is where we 4:14 started. We identify one specific use 4:16 case where people need to reply all 4:18 their messages and at some point we 4:20 added more features that that people 4:22 wanted right you you have to defend your 4:24 customer and just build the product they 4:26 they want. So we started this way and uh 4:29 we we keep uh we keep iterating and 4:31 improving the product on on a weekly 4:33 basis. 4:35 And so what was the exact problem or 4:38 case study that you were that sort of 4:40 triggered you to start this company? 4:43 I think uh the trigger was really 4:45 strong. We were at one point I was 4:47 project manager for marketing agency. I 4:50 faced the same problem over and over and 4:52 over again with the client. They were 4:55 not able to handle the number of lead we 4:57 were sending them. So at some point we 4:59 found a strategy that were generating 5:01 thousand of lead per month for them. 5:04 Obviously not all of them were like 5:06 extremely excited but but they needed to 5:09 move in the funnel of conversion. And 5:11 that time it was a sport manufacturing 5:14 uh sport equipment manufacturer 5:17 the client that we had and we get him in 5:19 touch three times with the Olympic 5:22 organizer in Paris. So it was a client 5:23 in France and we get him in touch with 5:25 the organizer of the Olympic game in in 5:27 the same year and they were all super 5:30 excited about the opportunity that that 5:32 our client brought to them but our 5:35 client just never replied. So it mean 5:36 that that golden opportunity to actually 5:39 do something meaningful with the with 5:41 the the Olympics game. So we identify 5:44 that if the AI was able to identify 5:47 which conversation need immediate 5:49 attention and draft a reply or even 5:52 reply itself to make the lead move 5:54 forward and just offer simply either the 5:56 resources that the person is asking is 5:58 the meeting link for people to book is 6:01 just a simple reply to make the people 6:03 understand better. it will be extremely 6:06 easy for the AI to do this this part and 6:09 and that what the use case we 6:10 identified. So we started with just an 6:13 agent that was replying to the message 6:16 and then we add the qualification and 6:18 the real time and the trust connection 6:20 and so on. 6:22 you mentioned so you were working with 6:23 an agency and through their marketing or 6:27 lead genen they were they were getting 6:28 leads but the humans were not responding 6:32 fast enough or just didn't have the time 6:35 and so that was the problem you were 6:36 solving is to specifically build an 6:38 agent to to capture these leads is that 6:41 correct 6:42 okay 6:43 that's correct the time the time you 6:45 spend between the time the lead is 6:47 replying and the time you you have to 6:49 reply it cannot be You have to be 6:51 sure. It needs to be instant. It needs 6:53 instant. Okay. So, the agency was being 6:56 being paid by by the clients to to to do 6:59 marketing to do socials getting the 7:01 leads but not capturing enough. So, can 7:03 you talk about the product and I sort of 7:05 how that carry into LinkedIn and why 7:06 LinkedIn? So when we were providing the 7:08 go to market strategy, we tested all the 7:12 channel for B2B, B2C and so on and and 7:14 what we saw that has the most success is 7:17 LinkedIn because they have been able to 7:19 build a very comprehensive acquisition 7:21 funnel. Um so we automated the outbound 7:24 because it was the easiest way to scale 7:27 how many people you can get in touch 7:29 with. Now LinkedIn is like we automate 7:33 one part of the acquisition on LinkedIn 7:34 which is the outbound. Now if you create 7:37 content, if you engage with uh the the 7:40 content of your prospect, if you invite 7:44 them to like podcast like this one and 7:47 if you like basically do social warming, 7:50 you are able to really skyrocket the 7:53 number of leads that LinkedIn generates 7:55 for you because you use it as a very 7:57 comprehensive inbound, outbound, organic 8:01 and and paid. We never did paid any time 8:04 but because the organic is strong enough 8:07 there. 8:08 I I know you've got many clients. I'm 8:10 not sure if you're eligibility to 8:11 disclose how many but I know you're 8:13 growing fast. Julian um maybe describe 8:15 the typical client. Is it direct to 8:19 client to the brand or are you working 8:21 through intermediaries or agencies for 8:23 for sales mind? 8:24 Yes. So I think initially we built the 8:28 product for agencies uh lead generation 8:30 agency B2B marketing agency because that 8:33 was where we were and what we what we 8:35 the problem we identified we wanted to 8:37 solve. U but at some point for the first 8:39 year we understood that it was way too 8:42 early for a product like sales mine to 8:44 actually uh fulfill all the requirements 8:47 of the lead generation agency. So we 8:49 started to target and business users. So 8:52 small businesses uh founders, small uh 8:55 sales team leaders as well. So five 10 8:58 15 people and and that was working 9:00 really well because we automate most of 9:02 the path in the lead generation, the 9:04 capture and the qualification. So that 9:06 was bringing up a lot of their time. 9:10 They were even replacing the VA. they 9:12 were hiring in the in the Philippine and 9:14 so on because uh they were able to 9:17 manage all the outreach by themselves 9:19 because the AI was making that so much 9:21 easier for them. 9:23 I know you you're using sales mind to 9:25 prospect your own clients within 9:27 LinkedIn. I love that. So what's 9:31 working? Is there certain vertical 9:32 certain you know client size or and 9:35 what's this most most popular use case 9:37 if you will? 9:40 I would say we are not uh we're kind of 9:42 industry agnostic. Uh we can serve most 9:45 of the industry. LinkedIn have 1 billion 9:47 users so far. So if you want to reach 9:50 someone in certain industry, the sense 9:52 that this person is on LinkedIn and 9:54 active on LinkedIn is really high. Now 9:56 what we see is there is certain industry 9:59 we cannot target like the manufacturing, 10:01 agriculture, 10:03 uh nurses, hospitality. Yeah. All those 10:06 industry are not really active on 10:09 LinkedIn. So it will be very challenging 10:11 to scale. So the best use case we're 10:14 seeing is with the SAS 10:16 we have been proving it with salesman. 10:18 We use salesman and that generates 60 10:20 70% of the leads that we're generating 10:22 come from salesmind. 10:23 I love that. I love that. 10:26 I was talking with someone with a VC 10:28 this week and he told me why don't you 10:30 eat your dog own dog food and and I told 10:32 him that exactly what we are doing right 10:34 because sales mind is generating the 10:35 lead for testine itself so at some point 10:38 we were able to send so much traffic 10:40 into the application itself what was 10:42 funny is at some point we were seeing 10:44 people we were messaging them we were 10:46 getting in touch with them they were not 10:48 replying but they were complaining in 10:50 the support for something because they 10:52 were signing up into the application 10:53 even though they didn't reply to to the 10:56 engagement we are creating. So the SAS 10:58 work really well, IT services consulting 11:01 works really well, cyber security works 11:03 really well, financial services works 11:06 really well and that the primary uh type 11:09 of client that that we had now we are 11:12 leaning uh a little bit more toward the 11:14 agency that have larger accounts stick a 11:18 little bit more longer engagement and so 11:20 on. So this is the market that now we're 11:22 tackling the the agency for that that is 11:25 selling to people. 11:26 Okay. Can you give examples of I mean 11:28 you don't have to give names of clients 11:30 in those two verticals that are actually 11:32 using salesmind and what are they 11:34 actually doing? 11:36 We have a a client called Zebra in 11:38 France. Uh he's is happy to to share the 11:42 the case study with us. He have been 11:44 satisfied with the product really early 11:46 adopters. So he faced a lot of 11:48 challenges that we that we let let him 11:50 face. But basically what what what he's 11:53 doing is helping company to drive 11:55 change. So it's a change management 11:57 software. He helped to organize all the 11:59 change in the in the company and uh and 12:02 basically we generate a lot of meeting 12:05 for him because he have a sign up free 12:07 signup page. So we just send constant 12:10 flow of uh of lead into his page and and 12:13 people sign up and then after we do not 12:15 have so much effort after the people 12:17 sign up right we stop more or less there 12:20 for the SAS and it's only to the product 12:22 to to see how much uh the conversion 12:25 rate of the product is and and then that 12:27 how we calculate the ROI. Now for the IT 12:30 services, consulting, cyber security, um 12:33 people are are offering like services, 12:35 development, market development, design 12:38 and so on. Uh and and that what works 12:40 really well because those service uh 12:44 companies they need a regular flow or 12:46 client. They need a predictable pipeline 12:48 and that what we're providing them. we 12:50 able to understand what was the the 12:52 success we had before and improve the 12:55 machine to ensure that we we understand 12:57 what are the the next the next batch of 13:00 lead that we are bringing. 13:02 Gotcha. Gotcha. Um 13:05 every business, every entrepreneur, 13:08 every investor needs to build brand and 13:10 sales. So maybe I'll toss out a few. Uh 13:14 so on work optional I've got a partner 13:16 his name is Michael Block and uh we are 13:19 looking to 13:22 acquire property management companies in 13:24 the USA right so uh specifically DMC's 13:29 that manage multif family apartments and 13:32 there's 326,000 13:34 of them in America and so we want to 13:37 reach a certain criteria maybe 10,000 of 13:40 them okay if they have a certain 13:43 financial standing whatnot. Um, now some 13:45 of these can be found in LinkedIn, some 13:47 cannot. Um, but if we were to then 13:50 engage salesmind to help us uh identify 13:54 the targets and then prospect the 13:56 targets, that's something that you can 13:58 do within LinkedIn, right? Can you walk 13:59 me through the process of onboarding 14:01 this project if we were to use salesmind 14:04 for that? 14:06 Yeah, sure. So, I will I will start by 14:07 asking a quick question. And do you have 14:09 that list of 10,000 people that you want 14:11 to reach or you you want us to find out 14:13 for you? 14:15 Uh we have 2,000 on a CRM system. Um and 14:21 then we can use uh services like Apollo, 14:24 right? Apollo.io to get them or we buy 14:27 lists from from property management 14:29 associations and obviously each of these 14:32 have certain quality of prospect and 14:34 certain cost of all of that. So ideally 14:36 we would give you the criteria right we 14:39 want a certain size whatever it is a 14:41 certain region you would find it first 14:43 and then I would imagine this will be 14:45 some kind of training of salesmind as 14:47 agent and you would the prospect right 14:50 like so again walk me through this is a 14:52 problem we need to solve uh I would 14:55 prefer to automate it versus having a 14:58 analyst call which is difficult or email 15:01 yeah 15:02 yes definitely so that that actually 15:04 works if you have past data of the 15:07 profile that you're targeting that's 15:08 even better. So we can rely on on those 15:11 profile to make a lookalike audience. Um 15:14 but basically we will take your 15:15 criteria. We have the agent that is 15:17 filtering the lead based on statistics 15:20 criteria for job title, industry, 15:23 skills, gender and BTI. I mean a very 15:25 comprehensive set of of criteria to 15:28 assess if the person is the right fit 15:30 for you or not. and and then basically 15:33 the you will we will train the AI for it 15:37 to actually remain compliant with your 15:40 brand value guideline and and tone of 15:42 voice. Um so basically reading your 15:44 website, reading your inbox help us to 15:47 to do that. We will ask you to validate 15:50 and and I think that what the AI has 15:52 brought in the last in the last few 15:54 years. It moves human from doer to 15:58 validator and then basically we'll ask 16:00 you to validate what the AI understood 16:02 out of your business your audience and 16:04 the people that it it gather for you and 16:07 then the engagement will start rolling 16:09 automatically. 16:11 We start engaging up to 200 300 16:14 activities per day and and let you 16:16 handle the the rest. And I mentioned 16:19 this use case because acquiring a 16:22 company 16:24 and co- calling them is obviously harder 16:27 to do than selling a SAS service or 16:30 selling a a a health product. The 16:32 prospecting and the training uh part you 16:36 mentioned, how does that actually work? 16:38 Does your agent ask us questions and 16:42 then we answer them? Do we create some 16:45 kind of dossier to then to then you then 16:48 ingest into our specific agent? 16:51 Yes. So basically the agent is not yet 16:54 asking you question to validate what he 16:56 understood. So you still need to 16:58 navigate a little bit through the 16:59 through the interface that we build for 17:02 you to actually validate what the AI 17:04 generated. Um so the way how we will 17:07 tackle this problem and and I think uh 17:10 what you mentioned also there there is a 17:12 big component of the timing in in what 17:14 you're providing like if it's not the 17:16 right timing people will just not not do 17:18 it right. 17:19 Yeah. 17:20 Yeah. 17:21 So this this is going going to be a 17:23 little bit uh the tricky part I would 17:25 say because identifying the right 17:27 trigger might might be the challenging 17:30 if we mon channel on on LinkedIn. Um now 17:33 the way how we will ask you to to to do 17:36 that will be to um just go through the 17:40 different things that the AI have done 17:42 for you. The lead that it gathered we we 17:45 give you access to the list. We give you 17:47 access to the entire knowledge base. So 17:49 you just like read online uh document 17:52 everything that we can find about you 17:54 online. We do the same for the for every 17:56 single and it okay 17:58 and we make it match. 18:00 Okay. So, just one more question on this 18:02 use case, and I may have one more for 18:04 you. So, so let's say we gave you 2,000 18:06 and we've got emails and and phone 18:08 numbers. Uh, and then we say, "Hey, 18:11 Julian, we want to send to 10,000, 18:12 right?" So, so then I would imagine your 18:16 sales mind would help us prospect within 18:17 LinkedIn. But let's say the 2,000, some 18:20 of them are not on LinkedIn, some are, 18:23 but you're starting with 2,000 from a 18:24 database and 8,000 that you prospect. 18:26 That's 10,000. 18:28 um do you 18:31 reach out to the prospects on LinkedIn 18:34 or do you also go outside of LinkedIn 18:36 let's say email SMS 18:38 that's a very good question uh we 18:41 currently only do LinkedIn and that 18:43 actually one of the reason why we're 18:45 trying to raise why we're raising 18:47 basically that to open more channel 18:49 where we see the most significant 18:51 traction is in the calling like the 18:53 email is really crowded now so doesn't 18:56 really matter how good the email is 18:59 people end up with thousand of email per 19:01 day while calling will be having a 19:04 significant impact if you can if you can 19:06 have that that scale and and people are 19:08 asking for that. 19:10 to voice voice. Okay, gotcha. Out of 19:13 2,000, let's say 1,000 you sort of match 19:16 LinkedIn and then you then we purchase 19:18 another 9,000 from you. There's 9,000. 19:20 Uh then you would uh ask for more 19:23 information. We would train it and then 19:25 you would then contact them over 19:26 LinkedIn. When when is the handover to 19:29 us like because if you if you send 19:32 initial messages and intros to 9,000 19:36 maybe 10,000 is interested to learn 19:38 more, right? Does your service stop at 19:40 saying ah now I'm going to I'm going to 19:42 bring in my client so that maybe we have 19:44 a conversation or it bring this email or 19:48 does it all happen within LinkedIn like 19:49 how far can it go? 19:51 Yes, currently where we so the all the 19:55 automation stops when someone reply then 19:58 the AI take over 19:59 and draft for you the the next message 20:02 for you to reply. So we make it very 20:04 easy for you to go through the 20:05 conversation and and make the lead move 20:07 forward in the funnel. So we drop the 20:09 reply and we ask you to validate the AI 20:12 replies before it's being sent early 20:14 July we're going to release the entire 20:16 autopilot that is actually conversing 20:18 within minutes and replying within 20:21 minutes to every single project. Okay. 20:24 And all within LinkedIn, right? Okay. 20:26 All within. So, so again in this case, 20:28 and again I picked a more complex 20:29 situation, we would train salesminds AI 20:33 to say, "Hey, we're looking to buy 20:35 PMC's. Are you interested?" Most will 20:37 say no. But they're interested yes. Then 20:39 within the LinkedIn messaging, we would 20:41 reply. But at some point, at some point, 20:43 we need to take this off LinkedIn, 20:45 right? Do a a call. So salesman would 20:48 stop once we get that to that stage. Is 20:51 that correct? 20:53 at that moment 20:54 you hand it over. Yeah. Okay. Okay. But 20:56 maybe you could explain to the audience 20:57 how you guys charge what we pay as a 20:59 client. 21:00 Yeah. Sure. So the the business model is 21:02 very straightforward. We have a monthly 21:04 subscription fee and uh that I'm sure 21:06 you have access to the entire 21:08 application which is one fee per 21:10 LinkedIn account that we supercharge and 21:13 we have planned for unlimited uh 21:16 LinkedIn account for agency that are 21:18 really loud basically. 21:18 Okay. Okay. very straightforward 21:21 and if I recall it's one fee monthly and 21:24 there's a LinkedIn limit to the number 21:26 of prospects but but not a salesman 21:28 limit and what was the limit per month 21:30 for LinkedIn? 21:32 Sure. So the the only hard cap that 21:35 LinkedIn is giving is actually 200 21:38 connection request per week. Now it 21:40 doesn't really matter how many messages 21:42 you send per day. Uh we have users that 21:45 send up to 150 messages per day. But the 21:48 the really important thing is the delta 21:50 how many messages and activity you done 21:52 yesterday and how many you done today 21:54 and and that what make raise flag from 21:56 LinkedIn. 21:58 So uh how many unique contacts can can 22:00 we do per day or per week or per month? 22:03 It's around 190 per day. 22:06 What's what's the fee again? 22:08 Yeah, we start at $89 per month for a 22:11 year commitment. 22:12 Gotcha. Gotcha. So, so 89 a month and I 22:16 think you you're getting a 42% reply 22:18 rate on your CL existing clients. 22:22 Yes, correct. Over the the the 3 million 22:26 messages that we send that that what we 22:28 observed. Yes. 22:29 Okay. Just to clarify, is it reply 22:34 including don't spam me or is it more 22:35 replying saying yes, I want to talk to 22:37 you? 22:38 No. 22:39 it's the entire the entire reply reply 22:42 pool that we're getting. Now what what's 22:45 really important and uh to to monitor 22:47 and and I think that was one of the 22:49 differentiator of salesmind is we 22:51 identify the opportunity. So people who 22:53 are interested people who are uh warm 22:56 people who are on open deal and we 22:59 identify that now what we see is 23:01 realistically between 25 and 35% of 23:05 interest rate which is a much better 23:07 metric to try. M your your typical 23:11 client typically buys monthly or six 23:13 months a year and what's the renewal 23:15 rate just so we get a sense of the 23:16 satisfaction level? 23:19 Sure. So I think most of the users are 23:21 six and 12 months package. Um it doesn't 23:24 really make sense for a business to that 23:26 have actually a sales cycle of six to 23:29 eight months to test a lead generation 23:30 tool and say okay it doesn't work after 23:32 one month. So people are usually 23:34 committed for for longer term because 23:36 they know that it takes time for them to 23:38 close the deal. U now I think the 23:41 retention rate we must be around 50 to 23:44 60% since we started. Um we we had uh 23:48 from the entire uh user base 23:51 that's fantastic. Those are good 23:52 numbers. 23:54 It it sounds to me actually yet I would 23:56 like to increase that to 100 but I know 23:59 I cannot serve anyone yet.