How SalesMind AI Uses AI Memory to Achieve $9 CAC x Work Optional (Part 2)

uh for salesmind how often does does your client update the package if you will you know the training data

unit and and is it is it self-learning with memory or is it clients coming in who who does that

so before it was the client that had to update and validate what we what the the AI was generated um now we have started to make the model learn by itself and improve the machine over time based on the result that it gets itself. This is very the very beginning of that. Uh the product needs still improvement and iteration based on that and how we can improve the the photo optimization but this is very exciting to to see that now based on the result the AI can can get better by itself.

Yeah. Yeah. Okay.

So now we get to sort of the technology. This is the that I think in in my view for for any any any sales to any agent, the ability for the AI to have memory, right? And and if you're through sales man, if you're doing 150 200 outreaches per day, you're getting objections, you're getting yeses, the ability to sort of learn from this and then propose to the client how to improve your product or your sales pitch, that is just amazing, right?

and and and and again with the memory it will become infinitely better than a human just because humans cannot process 190 every single day. So talk to us about that uh maybe maybe as you do that in terms of as well how how your agent is learning and improving if not yet how you plan to do that and then how does the text work. Basically we build rag uh for some of the some of the purpose and then we increase and add the objection database of uh the knowledge base of of the the agent that have access to it for getting new objection and then basically each reply we get is a reassessing.

Do we need to reassess the strategy based on that reply? And then the AI decide by itself, yes, it is useful for me to understand better the products, understand better the the prospect, understand better the lead itself or the segment that it lead is belonging to and and then basically it improves the knowledge base that will have an impact on on the next conversation and the next activities basically.

And is that live?

We we at the very beginning of it. So we have started to work on that. Uh this is not released yet on production.

This is still on on beta testing from our end. But we have started to work on that yet.

So at this point the current product if I if I'm talking if I'm prospecting 100 200 a day Julian uh am I getting are you giving feedback to me as a client? Because you you have the transcript you have the transcripts you have you have the conversations right. So, so if it doesn't automatically give me feedback, are you able to download that into some kind of Excel, feed it to some kind of agent and then summarize it on a weekly basis?

Are you are you offering that service right now or is that eventually going to be offered as an autom automated feedback system?

No, we do. What kind of feedback do do you have in mind? For example, if if salesmind is talk prospecting 100 people and 50 came back saying they they like this and they don't like this and 20 says you know the price is too high whatever right I don't really have time as a client to look at 100 transcripts but you can summarize that for me right you can say out of the 100 20% like love the product but they would wish for an average 20% discount or whatever right I I as a client I would want that and it seems to me the the the the the the data can can be digested easily by your agent to give me that feedback.

I'm just curious if that's offered whether it's manual or it's coming.

So we do not have that precision yet. We are able to tell you okay you have like 10 10 20 people that are interested 70 people that are skeptical and and whatever. So we are able at that point to do that.

But the the feedback about

of the out of the interested people like 5% were expecting a lower price them expect we're not having that yet. But I take the note and as a and as a a good product manager I'm going to add that to the road map at some point.

Okay. Okay. I mean it just seems to make sense, right?

Especially as you scale this up. Um. Gotcha.

the way how we plan to achieve that is by starting getting integrated in people workflow and and that why we're switching slightly the audience that we are targeting is because we know that agency B2B marketing agency and lead generation identity they will start putting workflow in their client uh processes and that where we want sales mind to to start being not only an interface, not only the agent that can perform the task on your team, but being integrating on workflow and interacting with the application and the knowledge base that we have about people in the company. So the technology that we're using is fairly simple, right? Node GS and react to to build the thing.

Uh but my CTO can can talk much much more about that. So I I don't have a very comprehensive overview of what are all the tools that we're using all the the languages that we are using for all the application we we have among 10 different microservices and so on. The the tech team is is really working hard on that.

How easily replicable is it you know if somebody were to basically come out with a competing product? what what what's in your sauce technology or distribution or something that would prevent somebody else from offering something that's 49 a month.

Um the way how we see that is by having a very comprehensive solution like

we see so many different use cases. We even have like now client and and people on gold that trying to ask ask us what are the difference between myself and and their self doing workflow and n and and whatever and the answer is fairly simple either you want to build and and spend time to actually build your own workflow and and put resources out there. So it's actually very cost intensive to have the knowledge have the people inside and so on.

So you can start maybe in three weeks if if you are fast enough or you can start with us uh tomorrow. Now if a company wants to start exactly the same as what we have now um and and do the same product as a product to resell and set that 49 that will be um very challenging because we we currently have over 10 million profile in our database that we understand how to communicate efficiently and every single messages that we having with those person make the system smarter. So each message make the the system and the agent more likely to succeed the next time you send a message.

So the data that we processing about people and profile in in the business environment make us a little bit smarter all the time and and I believe that what going to make make hard for people to catch up uh from from this interface perspective. Next will be to get integrated in people workflow like when you when you plug your CRM for something and whatever. Um like sending directly the work to the human where salesman say hey I have a good lead for you.

It need like it need a human intervention. He needs to get on a call or whatever like we aim to get irreplaceable by value as well. If if we make your business survive or even strive because we send you enough lead to fill your pipeline and grow your revenue, you you will never be able to leave us.

So for the audience, uh we we love featuring tools that actually work and it it looks to me like giving you value. You've done 3 million plus messages and the value proposition to a a client, Julian, and and feel free to correct me is, you know, for less than $100 a month, right? Uh you can with salesmind uh uh maybe spend a few hours training salesminds agent and they send out 150 200 uh almost 200 per day to prospects and your history shows that you're getting a 42% reply rate, right?

Then you can then converse and maybe take it off LinkedIn to whatever you're selling. It could be selling financial service, selling a software, whatever it is. And if you then divide that, you know, $100 a month by 150 a day times 30 days times 42%.

It's a extremely cheap, painfree way of getting prospects, would you say? Is that sort of the the value ad? It's getting that prospect right.

It's extremely inexpensive and quite automated. It's it's it's beautiful.

Amazing. I think you you you rephrase it and you you nail it very effectively, right? The cost of acquisition customer acquisition cost for sales mind is currently $9 per client.

$9. Okay.

So very sorry. So how did you get that number? What do you mean by uh cost of acquisition?

Because a reply is not acquisition. A reply is a reply. When you say acquisition, did that mean actually buying the service?

Yes. So customer acquisition cost for us is $9 because we eat our own dog food. We use sales mine to protect and generate.

Oh, this is for sales mind. For sales mine I see.

We cannot we cannot know that effectively for all the the users but for test mind we monitor that.

So you mentioned fundraising. Uh to the extent you're comfortable have you raised money? What round are you at?

And and and what's the plan once you've got the money? So we raised preed with a venture studio. So basically they are providing now the entire development team for years for a year.

So that has been significantly uh improving the quality of the product the fast uh the speed of the delivery that we were able to do as well. So we've we've passed that now we're raising uh feed and the way how we plan to spend the money uh from seed will actually to scale what actually works for us. So we need to increase the number of account that we are having the number of activities that we are doing for ourself.

So the the the the main goal for that will be actually for us to to scale. We have a little bit of product um that needs to be to get done like the email, the phone, WhatsApp will be the easiest way to open a larger town basically because you mentioned it yourself right out of the 10,000 people you need to reach maybe 8,000 of them are not on LinkedIn. So we need to reach them through other channel and and that what we will be able to do with the funding to improve the agent capabilities other channel and ensure that we are multi- channelannel and be able to communicate through the preferred channel of the lead.

I mean, you're one of many agents that we're seeing that are solving very specific business cases and there's a a number of different sales products that that we're we're invested in are also doing. How fast is this developing in your view and how do you see this uh enterprise agent space evolving over the next three five years?

Wow, that's a tough question.

It is. It is. But I think I think you probably know you have an opinion because you're playing the space, right?

Yeah.

Yeah. We see the challenge is is going extremely fast like from months I say like four months to six months we see new trends appearing um new challenges that we need to face uh and so on. So the the space is moving extremely fast especially like Fman is in a competitive landscape.

uh the sales industry is really big and and the time is just enormous right like 57 billion has been spent in sales and marketing tool last year but people still have an empty CRM so in the next three to five years where I see that the sales agent will have the most impact will be and at least the vision that we are having at salesman will actually make the funnel the entire yourself funnel entirely personalized. So it mean the first message is personalized, the reply is personalized, the seat that you're seeing is personalized and the interaction you have with the AI live avatar is personalized to you only and your specific need because you don't have time to spend 10 minutes on a website and on the person profile to try understanding what is the offer, what is the value, what is the benefit and so on. We just need to show you what matters to you.

how far would it be for me to prompt GPT plus connected to my LinkedIn and my Dropbox and my sales kit to say do what salesman does and is that a threat to you? It is not uh it is not a threat for us because you can eventually do it for one of your accounts but can you reliably do it for 100 accounts of LinkedIn or 100 email. Um so this is the use case that we we solve for agency uh and that actually a great a great moment for us to to switch if you if you notice that chip will actually do that.

I'm not entirely sure they will allow you to connect with with LinkedIn itself directly to do what we do. Um but eventually that we don't see that as a threat because one person that is trying to automate his own profile is not entirely the the audience that we are looking for. So you you you can do it by yourself.

You can play if you're a builder and you likes it and and I know you likes it, right? You're in the same space. you have a sparkling mind and and a beauty and and and great hand to to build a product.

just yeah I wasn't I wasn't picking on sales mind necessarily Julian because a GPT or Perplexely are obviously they're all launching or have launch agents right I I was playing with Perplexity Labs this week and and it's not as good as Menace or GSpark so my question is less on sales but really saw these LLM tools that are evolving into agents and making easier uh you know there's always that danger of them disrupting especially a successful company like yours right so I'm just curious how you think about that and how you defend against that

the the way how we think about that and we defend about that is by the number and the scaling uh capabilities that we have I think if tomorrow open AI is launching a system for replying your email. Yes, it will be able to do it for yourself that works. Uh but for the entire um set of client that I didn't see have or making it easy for the entire company to actually integrate that I think it will be a little bit more challenging over the the the coming years.

Uh but that's only the very first top of the funnel, right? Where you build a little spark link. So where we see that cesman will have a very big advantage will be to actually have um adaptive micro site where again the messages is only a door for you to go to the next step in the sales funnel and having a way more personalized uh experience from that.

So we're not just sending messages on LinkedIn. We're just like moving the lead in the funnel and and moving it down and qualifying it. So this open will be able to send the messages but will it be able to make it move in the funnel?

Good question. Uh I hope no. Uh but if they do it something different in that case.

Yeah, I think that's probably as good an answer as we can give at this point. It's moving so fast. Um a geopolitical question.

Uh um and I I don't know enough about your tech stack, but China obviously is leading with quite a it's primarily open source, right?

How do you see the use case of China versus US technologies?

Um great question. Um we we see that the models have different strength uh each of them for different use case. So we leverage I think five or six different ones currently for all the process that we have the qualification identification matching and so on.

Um, now the competition between both I have no idea. I have no idea. I love China so much.

So I have bias in in the balance.

It's also a lot cheaper, right?

Yes. Yes. And extremely good, right?

The reasoning model are amazing. So I would say it is really surprising that I've been able to release the model that fast eventually. But we know China is really leading the the innovation.

So I'm really confident that in the in the next few years it's going to be everywhere. I was in China a few few weeks ago and actually I see that now in BU map they're integrated deepseek. So when you go on a business profile like Google Google business right you can find the the time when they open the address and so on and you can like text deepseek to actually answer the question from there.

So I think what are the the big advantage China have in the in the race is they have one one pass only. So if if they decide to put deepseek in all BU it will be instant. If they want to put that in which one billion and half people will have access to AI right away right now.

So the adoption rate will be incredible for them.

I have a question about enterprise adoption. I I put it in my initial Q& list Q&A list for you. Um uh uh you know we're we're spending some time with enterprises.

Um and it's different than what you and I have built at the startup world. It seems larger companies they want a lot of them see the the value of enterprise agents right but the issue of governance and ethics privacy and security those are major issues to be solved right are you talk are you talking to enterprises that are thinking about let's say deploying salesmind or similar agents and how are they resolving these major issues that frankly is very new, right? Like you know if if you want to deploy a sales mind and plug it into a a enterprises database, what if the agent goes haywire, hallucinates or agent decides to steal your data or whatever, right?

H have you have you come across this issue? I'm just curious. It's because it's top of mind.

Yeah. Yeah, we do have enterprises client actually and and the way how they manage that that was funny because I got in touch with them but the sales process is just so much longer. I think we started with one pilot, one person and and one team, 10 people and and now like 80 or 90 people in in the company for the sales.

Uh but what changed is they replaced the people that were like manually sending the outreach because that was the job of 15 people actually to go on LinkedIn. Try to understand what the person is doing. Try to qualify the lead based on the information they have.

Start a message put the lead in the in the in the in the Google sheet and then report the Google sheet at the end of the day. So the guy was lasting not even 30 connection today. uh so they had like basically a bunch of people doing that.

Now those people have been that that team have been reduced. So I guess those people now have other attribution uh and higher responsibility. Now what happened is they have validator.

So each time they come up with a new agent, with a new campaign, with a new mission and so on, they have someone that is validating the quality of the AI understanding and fine-tune if needed because we give you access to everything. We let you handle the modification, if the AI hallucinate a little bit or if is not entirely correct or the way you will promote it yourself. So they have someone that is controlling that, but one person is enough for the entire company.

Gotcha. Um so Gary Tan uh who's the Y Combinator I think a few months ago said in the latest batch of YC companies um he surveyed the CEOs um and they said that he's I think he mentioned like 90% of the YC latest batch are using agents to build to build software right so instead of having 10 devs they've got maybe uh like like two or three and using agents um really uh uh you know if you if you think about that it really is um what I mean YC is selectively finding entrepreneurs that are using agents to build software right and so this is really the the whole idea of a lean startup and we we had a news newsletter that talks about that um how do you see that and and are you guys AI native as well?

Yes. Yes, definitely. Uh this is one of the first thing that we we do when we onboard people.

We give them a chat GPT plus account and a curl account so that we ensure that they they have at least the bathing tool to to do what they need to do. Um we actually migrate one part of the application that could have take weeks uh before with codeex this weeks. uh so we leveraging significantly the AI we have training internally about one uh one champion that have identified the right tool and the right capability of the tool and the right use case.

So he's leading the adoption within the company and within the other developer. So we have like someone that is uh finding the way and and showing the way for the other collaborator to actually leverage AI more in the code. But I've been extremely surprised that Codex now is able to send PR basically for you.

You're just giving a task and and it send PR on on GitHub.

Do you guys have a goal? Are you guys going to go to two million per employee?

My team was joking about we can we can replace our people with Codex. What I what I believe is we're going to have to keep the people and just make them 10 times more productive.

help themsel with AI to get the 10 times more productive. I think we're not at the 10 10x yet but we're working on that. So if if gamma is 1 million per employee that that may be the next our goal.

What my takeaway with this conversation is it doesn't really matter from a client standpoint whether you're using deepseek or using this text. What matters is that you're specifically solving a problem and the problem is uh you you've trained an AI to essentially learn the product or the service and you you've now very smartly integrated into LinkedIn which is as you said is a billion people. So, you'll find a lot of clients there.

And you you're now uh at least in salesminds, each your own dog food, you're you're acquiring new customers at $9 per per client. Um I would argue um is cheaper than that. um in the sense that if we didn't have salesmind and you're having people call right the amount of time that you spend training or retraining sales people turn over uh at management time uh is that usually isn't really boxed in.

So if I were to to compare using salesmind to do a sales lead versus the traditional way it's a it's a massive factor of cost and time. Would you say that's and I haven't done that but it just seems to me it's a lot more intuitive because once you train it it it scales up so much quicker right and it gets smarter especially with the memory piece. Yeah.

So so to me the $9 is a fantastic number but compared to do traditional way I'm sure if you box all that in it's actually I think a much more higher multiple of productivity. I'm not I didn't say that very clearly. Hopefully that makes sense to you and and maybe you can comment on that.

Yeah, that make a lot of sense. What we see as an average in the B2B industry is a lead cost 200 to $250 as a cost of acquisition, right?

Wow.

With salesmind because we are using our own solution, we we've done like $9. So it's significantly lower like we we could not actually achieve that with any other uh way I believe.

Great. Great. Okay.

Hey Julian, Saturday morning. I'm sure you've got lots of things to do. Do you do you want to tell us how the audience can reach you?

Our sales mind.

Yes. Uh fairly simple on LinkedIn. We are we are super visible.

We're creating more content about AI in sales and we we're really evangelist about how AI can make the the businesses 10 times more productive as you I guess. So LinkedIn will be the the best way. Julian Gada on self mind or self mind.

on LinkedIn, sorry, or sales man on LinkedIn and it would be very easy to find.

Fantastic. Well, hey, thank you so much. Uh maybe we can get you back on uh in a few months to to see how you you guys are growing and and uh see the ver Yeah.

Yeah. Fantastic. Thank you so much, Julian.