# Founder10x - The New Rules of Building SaaS - Santosh Sharan, Co-founder & CEO, Zeer.ai

An episode of [VC10X - Investing, Venture Capital, Asset Management, Family Office, Wealth Management](https://invisiblenetwork.ai/podcast/1632806986), 23 Sep 2025, 53 min.

Audio: https://anchor.fm/s/a453ddd4/podcast/play/108699940/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2025-8-23%2F407996177-44100-2-04c7e6482a18f.mp3

Santosh Sharan is the founder of Zeer.ai and a go-to-market veteran with over 20 years of experience. Santosh believes we're at an inflection point where the traditional SaaS model is fundamentally broken. In our conversation today, we explore why the old playbooks don't work anymore, how AI is reshaping startup economics, and Santosh's vision for agent-to-agent interactions replacing traditional sales processes. ⭐ Sponsored by Podcast10x - Podcasting agency for VCs - https://podcast10x.com Zeer AI website - https://zeer.ai/ Santosh Sharan on LinkedIn - https://www.linkedin.com/in/ssharan 🎙️ Hosted by Prashant Choubey Follow Prashant on LinkedIn - https://www.linkedin.com/in/choubeysahab In this episode, we talk about - - Building AI startups versus traditional SaaS startups - The breakdown of traditional go-to-market strategies and rising customer acquisition costs (CAC) - Platform versus point solutions and the future of venture capital - Agent-to-agent interactions and buyer-led go-to-market approaches - Creating defensible moats through data flywheels and proprietary datasets & more Timestamps: (00:00) Introduction (03:05) Santosh Sharan's background and introduction to Zeer AI (05:34) Differences between building AI startups and traditional SaaS startups (07:51) The challenges of building AI products and the importance of going beyond being a wrapper (11:57) The rapid changes in the tech industry and the need for founders to be ready to pivot quickly (15:05) Impact of low-cost software development on business margins and valuations (20:17) Thoughts on venture funding and raising capital for AI startups (22:43) Platform solutions vs. point solutions in the current market (27:38) Creating a moat in the AI startup ecosystem (31:44) Changes in go-to-market strategies and traditional sales approaches (38:37) The future of buyer-seller interactions and agent-to-agent communication (41:49) Using LinkedIn for prospecting and generating interest (44:21) Cutting through the noise in the AI market and standing out (46:23) Predictions for the future of B2B interactions in five years (49:09) Overview of Zeer AI's buyer concierge concept For sponsorship or guest appearance requests, write to prashantchoubey3@gmail.com Subscribe to VC10X on Youtube, Spotify, Apple Podcasts.

## Transcript

[0:00] So all the playbooks that worked earlier, all the marketing
[0:03] channels suck right now because the cost of building software is
[0:09] going down. Everybody's going to build 1 and
[0:11] they all they want is like 3 or 4 customers to be to make it
[0:14] profitable because cost is so low.
[0:18] The go to market distribution narrative, what you're building,
[0:21] those things matter a lot more than building.
[0:25] The velocity and momentum itself is going to be the more.
[0:29] It's not velocity or development as much as velocity to which you
[0:34] react to market forces. Cost of building product is
[0:39] going to go down increasingly to 0.
[0:42] CAC is going to go increasingly to like Infinity.
[0:45] It will just keep on rising. I foresee that a lot of venture
[0:50] funds will fold up in the next 5 years.
[0:54] Customers are not just for revenue.
[0:56] Most founders make the mistake thinking customers paid revenue.
[0:59] Their role is done. Customers should help you Co
[1:02] create the business. The future of go to market I
[1:07] think is where buyers start the process and sellers help the
[1:12] buyer buy better. There's so much of noise and the
[1:17] noise is increasing that we are reaching a stage where almost
[1:22] being different is more important than being better.
[1:28] Hey everyone, this is Prashanth and I'll be host for the VC
[1:30] Tennex podcast and today we have Shantosh Sharan with us.
[1:34] Santosh is the founder of Z Rei and a go to market veteran with
[1:38] over 20 years of experience. Santosh believes we are at an
[1:42] inflection point where traditional SAS model is
[1:45] fundamentally broken. In our conversation today, we
[1:47] explore why the old playbooks don't work anymore, how AI is
[1:52] reshaping startup economics, and Santosh's vision for agent to
[1:56] agent interactions replacing traditional sales processes.
[1:59] So without wasting any time, let's dive straight in.
[2:02] This episode is sponsored by my own podcasting agency called
[2:05] Podcast 10X, where I help venture capitalists, asset
[2:08] managers, allocators and family offices and starting their own
[2:12] podcast and building a strong brand.
[2:14] Podcast NX is designed in a way to work well with the busy
[2:17] schedule of investors where we handle every aspect of podcast
[2:21] production so you can focus on investing.
[2:24] If you're looking to build a strong investor brand, starting
[2:27] a podcast should be right up your priority list.
[2:31] It's the only form of marketing where you can showcase your
[2:34] strong network. In other words, access to
[2:36] limited partners and founders and grow that network further.
[2:40] And contrary to what most people hear, it's still very early to
[2:44] start a podcast. There are at least 10X more
[2:46] newsletters and thousand X more blogs than there are podcasts in
[2:50] this world. You can learn more about our
[2:53] podcasting services at Podcast ten-x.com.
[2:56] I'll make sure to add the link in the description below.
[2:59] Enjoy this episode. Hey Santosh, good to have you on
[3:02] the BC 10X podcast. How are you doing?
[3:05] Very good, Tim, to be here. Yeah, it's a pleasure having you
[3:07] on. To start things off, can we have
[3:10] a bit of your background, your story and what you're working on
[3:13] right now? Sure.
[3:15] So for the last 1520 years, I've worked with various go to market
[3:21] and data companies. And during this process, I feel
[3:25] like V as in go to market teams got a bit greedy with data and
[3:33] automation and what worked earlier no longer works.
[3:38] So all the playbooks that worked earlier, all the marketing
[3:41] channels suck right now. Emails are not getting
[3:45] delivered, nobody responds to phone and so on.
[3:49] So this is where last quarter I was thinking about what's next
[3:53] and I decided to flip the funnel and do a violate go to market
[3:59] platform with Zier AI. So that's what we're doing with
[4:04] AI and agents. I think there's an opportunity
[4:08] to create a buyer LED go to market.
[4:11] So contrast that with what we have today, which is seller LED,
[4:16] where sellers are sending say 10,000 emails and only 10 buyers
[4:20] are in market to buy. So other 9000 and change looks
[4:24] like a spam, right? Buyer LED go to market is where
[4:27] buyers raise their hand and they start solving a problem where
[4:32] sellers participate and help the buyer.
[4:34] Now it turns out Gartner has been talking about Biolet for
[4:38] over 10 years, but it turns out nobody really built any
[4:42] solutions for the buyer. So there's no CRM for buyers,
[4:45] there's no tools for buyer to run their buyer journey.
[4:50] So we are building some of those tools for a Biolet agent to go
[4:55] to market platform. Yeah, that's quite interesting.
[4:59] And you're someone who has been in the space, you know, working
[5:02] on GTM strategies at different companies.
[5:04] You have some sort of experience in running all these things,
[5:08] especially GDM and such. So we'll talk along some of
[5:12] those topics as well, which I'd add to send to what you're
[5:15] building as well, right? So the first one on that front
[5:18] being you've said building AI startups is completely different
[5:22] from building a SAS startup, right?
[5:25] What's the most dangerous assumption you say that founders
[5:28] have about building AAI startup versus building a SAS startup?
[5:34] Yeah, yeah. So I didn't know before I
[5:39] started dipping my feet in SAS. If you think about it, there's a
[5:44] lot of front loaded work and then from there onwards it's
[5:49] like linear, linear or progressive input in order to
[5:54] get success. AI doesn't follow the same rules
[5:58] because the with the AI startup, you know, there are some very
[6:03] core AI which takes a while to develop and fundamental AI
[6:09] technology. But that's not what I'm talking
[6:10] about. Most of the AI apps like ours,
[6:13] right? Often times some part of it is a
[6:18] wrapper over GPTS. So you get an initial euphoria
[6:24] because in a matter of weeks you get to 50% of the app, right?
[6:29] And then the real work starts because then it takes another
[6:34] 3-4 months to finish to tweak to build all the apps.
[6:39] So that's something founders need to plan for.
[6:43] But also I think the real value is not in creating those
[6:48] wrappers, right? The real value, like in our
[6:51] case, we are putting together a network around it, almost like a
[6:55] social network or network of buyers and sellers that we're
[6:59] trying to solve a real problem, not just provide chargeable
[7:05] info, right? So those things take a while.
[7:08] One, one last point, I think building a narrative or the go
[7:14] to market for these AI startups is also not as simple as SAS.
[7:19] There's a lot more focus on value and outcomes than they
[7:23] used to be earlier with like subscription type services.
[7:28] Absolutely. That's quite interesting.
[7:29] And that's quite a shift that's happening, especially in the B2B
[7:33] space. So, so that's an important
[7:37] distinction. You made that OK, the first
[7:38] 5060% of the product founders can make maybe in a few weeks,
[7:43] right, with the tools that are available now.
[7:46] But the other 50% or 40% that's remaining, that's where the real
[7:50] world lies. And still you said that tech or
[7:54] AI itself cannot be your motor differentiator.
[7:57] It has to be some sort of network or something that's
[8:00] different about you and that's defensible about your business,
[8:03] right? So that's quite interesting.
[8:06] And how did you take that journey of building the initial
[8:09] 60% of the product at the the AI early on were the same tools
[8:16] that are available today were accessible to you or not?
[8:20] Or how do you see people playing that out right now, founders
[8:23] doing that, building the initial product early on and then, you
[8:26] know, tweaking it and refining it for the use case?
[8:30] Yeah, yeah. Our journey was a little bit
[8:34] different, but I'll, I'll talk In general the journey that most
[8:39] founders go through. I think there is some advantage
[8:44] of being able to build, put something together very quickly,
[8:48] which is what I'm seeing the good side of AI development or
[8:52] wipe coding if I can put a name to it.
[8:56] The good thing is you have something now you can take it to
[8:59] market and get feedback, right. So some of which we did as well,
[9:04] although we are building a compound startup, not an
[9:07] endpoint solution. So we are building like seven
[9:10] different apps in the and we're building a whole platform.
[9:14] So ours was a slightly different case.
[9:16] But let's say you're building an AI app and you can do 50% of the
[9:21] app in like a matter of weeks. Then I would recommend going to
[9:24] market right and getting some feedback and that helps with the
[9:29] product market, Fred, which earlier you couldn't because it
[9:32] costs so much. The front loaded effort was so
[9:37] much that this traditional development cycle of getting
[9:43] really knowing what you're building and being very sure and
[9:46] interviewing users and putting together requirements and then
[9:50] going building it and then taking it to the market.
[9:55] After like several months, that whole cycle has been disrupted.
[9:59] Now within two weeks of starting the company, you should have a
[10:02] potential customer using something.
[10:05] And as you're getting feedback, you are building it.
[10:08] So your product and engineering cycles have gotten compressed
[10:12] and now you're doing one week turnarounds rather than three
[10:15] months turn around. And that really helps companies
[10:21] attain product market for it much sooner than they would
[10:25] force us, Absolutely. Yeah, that's quite interesting.
[10:28] And another point you mentioned is that you cannot just be a
[10:31] rapper to charge GBT or other LLMS.
[10:35] It needs to be something more. And I think one good example of
[10:39] for that would be a common example that most of you people
[10:42] would know is complexity, right? A lot of people say that it's
[10:45] just just a rapper, but there are two more things to it which
[10:49] meet the eye when you go deeper into it, is that they're
[10:52] aggregating different LLMS in a single platform.
[10:54] And the other thing is that they they have been specialized
[10:58] features and the UII would argue is better than other platforms.
[11:02] So they are actually still building for the user and giving
[11:05] them what they need while working with these different
[11:08] LLMS all in one platform. So that's the differentiator
[11:11] there and that should apply. How does that apply to your own
[11:14] startup? And that is something that
[11:16] founders need to think about, right?
[11:18] Absolutely, absolutely. That's it.
[11:21] Awesome. Great.
[11:22] So again, something that you've said is that founders these days
[11:29] need to be ready to pivot in just a matter of weeks, right?
[11:33] Because in changes and these new launches are happening every now
[11:38] and then. Almost everyday there is some
[11:39] new AI launches happening, right?
[11:41] And some launches can make what you're building obsolete or
[11:44] maybe you can do something better than what you were
[11:46] already doing. So give me an insight into that,
[11:48] how our founders should gear up to, you know, be ready to change
[11:52] at the speed of change that is happening in the industry right
[11:56] now. Right, right.
[11:58] So there's, there are multiple fundamental shifts that's
[12:01] happening right now, right? So it's important to understand
[12:04] that and the implications it has on venture building, right?
[12:08] So one of the fundamental shift is that price of building tools
[12:14] is going to 0, right. So earlier all the focus was on
[12:18] building. Right now you think about what
[12:22] you want to build and then the building part is easy, right?
[12:26] So the go to market distribution narrative, what you're building,
[12:30] those things matter a lot more than building because the
[12:33] nature. So the world we are going into
[12:36] is let's pick a category, say customer success software.
[12:41] So 10 years ago, say if there were 40 today, say there are 80.
[12:46] Next 12 to 18 months, there's going to be 800, right?
[12:50] So there's going to be a quantum jump in, in every single AI or
[12:54] SAS category because the cost of building software is going down.
[12:59] Everybody's going to build one, right?
[13:02] And they, all they want is like 3 or 4 customers to be, to make
[13:05] it profitable because cost is so low, right?
[13:09] So now this leads to your question about, you know,
[13:17] eventually the way you think, let's say you're building
[13:24] something, a customer success tool and you have a unique
[13:28] angle. And then you hear that somebody
[13:31] just launched something similar because the piece at which these
[13:35] tools are coming at us is furious, right?
[13:39] So the ability to pivot and come up with new ideas and how
[13:45] quickly you can kind of build upon some of your earlier builds
[13:50] is going to be the key. The velocity and momentum itself
[13:55] is going to be the Moat, right? And, and, and to, to double
[14:01] click on that. It's not velocity or development
[14:04] as much as velocity to which you react to market forces, right?
[14:10] Because development is going to be like we are saying lower cost
[14:15] and efficient and high velocity anyway.
[14:17] But then a lot of companies don't react to market very
[14:21] quickly. They don't like how competitive
[14:24] forces or, or market forces, right.
[14:27] The companies that are able to react like overnight almost or
[14:31] like in a matter of days rather than months, they're the ones
[14:37] that are going to win. Yeah.
[14:38] Absolutely. That's quite interesting.
[14:40] And the point you mentioned about that, there's going to be
[14:43] far more competition in the space, right, Because making
[14:46] these tools are SAS businesses, it's going to be so much easier
[14:50] that more people are going to be doing it, right, who would
[14:52] otherwise not build something like that.
[14:54] But now, since it's so easy, they can build it and put it out
[14:57] there in the market, right? So how does that impact the
[15:00] margins for these businesses and the valuations these are
[15:03] driving? How do you look at that, let's
[15:06] say, from an investor perspective, How should they
[15:08] look at this market now? Yeah.
[15:10] The venture investment has completely changed or will
[15:13] completely change, right. So there, there are a couple of
[15:17] numbers. So cost of building product is
[15:20] going to go down increasingly to 0, right.
[15:24] CAC is going to go increasingly to like Infinity.
[15:28] It'll just keep on rising like higher and higher to the point
[15:31] where my prediction is that it won't make any sense for
[15:35] Endpoint solutions to afford a CAC.
[15:38] Because if you're going to spend $500,000 to buy a customer, then
[15:44] why are you going to sell your subscription for $200, right?
[15:49] You'll run it. So everybody's going to try and
[15:53] become a platform or join a platform.
[15:55] So they could subsidized the CAC, right?
[15:59] Now back to venture investment. If you think about venture
[16:02] investment in the last 20 years, they banned on winner take all
[16:07] models. They hoped that, you know, two
[16:10] companies out of the 20 or 10 they invest in just have a like
[16:18] a crazy high returns and then they pay for all the losses,
[16:23] right? And the reason they would have
[16:25] high return is because the way the venture ecosystem work is a
[16:30] few companies ended up becoming premium vendors and they could
[16:34] drive high margins, right at the cost of everyone else.
[16:39] So look at every category like say CRM, the Salesforce,
[16:42] HubSpot, and then there are 400 of the CRM that fight for 15% of
[16:46] the market share, right? And that's the case with every
[16:50] market segment. There are like two or three
[16:52] companies that own 80% of the market share and then 100 plus
[16:57] to fight for 15%, right? So that model is broken now
[17:03] because when you have so many competitors and then there's
[17:11] going to be feature parity, right, everybody is field is
[17:14] going to begin to look the same in a world like that, there will
[17:18] be immense downward pricing pressure.
[17:20] So these premium vendors are going to see they'll have to.
[17:25] So the margins will drop in general because if you have like
[17:31] say 2X3X number of vendors, the time is not growing 2X3X in the
[17:35] same time. So then everybody has to make
[17:37] room for these new vendors and by definition these new vendors
[17:41] are going to be lower cost. So the highest pricing vendors
[17:46] are going to have the and then outcome based pricing is making
[17:51] its way from SAS pricing. So that will cause another
[17:55] downward pricing pressure. So generally these new pricing
[17:59] models are first incorporated by new entrants rather than
[18:04] existing companies. So I foresee every SAS company
[18:11] is not agents company and new vendors in the SAS companies are
[18:16] going to be agents that have outcome based pricing.
[18:19] So you only pay when you have positive outcome.
[18:21] Otherwise it's free or usage based costs like you only pay
[18:25] for tokens or something. And the biggest target for them
[18:30] would be the the winner take all models, which means VC model is
[18:36] like at a lot of stress. And I foresee that a lot of
[18:42] venture funds will fold up in the next five years and we will
[18:47] probably see outcomes. We'll begin to see outcomes
[18:51] where they'll be super, super efficient companies like look at
[18:59] cursor lovable type companies, right?
[19:02] With like 50 people, they're doing hundreds of million
[19:05] revenue, which wasn't possible before, right?
[19:07] So we will only get more and more efficient.
[19:11] So VCs are not going to shut down.
[19:14] It's just that they were hoping for say two or three unicorns to
[19:19] pay for all the 20 mistakes they make, right?
[19:22] But now they will get one bigger cause like extremely profitable
[19:26] companies to pay for all the 20 mistakes.
[19:31] So they look for the one cursors and the lovables.
[19:36] So they'll still be, it's just there's going to be more
[19:38] polarity, there's going to be more failures, and the successes
[19:42] are going to be bigger and far fewer.
[19:45] Absolutely. That's very interesting taken as
[19:47] a founder yourself. Have you decided to take on
[19:52] venture funding? If yes, how are you thinking
[19:54] about how much should I raise? At what valuation should I raise
[19:57] in order to not, you know, put additional pressure on myself,
[20:01] you know, that, you know, becomes a more difficult
[20:04] situation for me to operate in or making promises that are hard
[20:07] to fulfill in this industry right now with the rapidly
[20:10] changing environment. So how how are you thinking
[20:12] about it as a founder of GDI? Yeah, we haven't take it venture
[20:23] funding. We do intend to maybe accept
[20:27] there has been some interest and inbound inquiries.
[20:29] So in the next few months we might close around I those days
[20:35] are gone when you needed large rounds to scale a company.
[20:39] Now you can do it with less. And that just leaves, you know,
[20:46] it's just a healthier way of growing companies and more
[20:49] efficient way. So I don't see, at least in the
[20:55] beginning, needing a lot of capital.
[20:59] But again, having said that, in the long run, who knows?
[21:02] You know, the companies have their own life cycle and it'd be
[21:06] silly to not pour fuel when you know things are growing.
[21:12] So I've learnt in life never to say no to anything.
[21:16] But having said that, I, I don't think you need hundreds of
[21:22] millions of dollar to get to like a billion dollar valuation
[21:26] like you used to before. I do like the efficient way of
[21:30] growing companies, but a lot of these efficient companies,
[21:33] including Cursor and lovable, if you look at them, they're all
[21:36] raising hundreds of million, even though they have like a
[21:39] small team, right? So the money, it's not that the
[21:45] money is going into go to market and not product, right?
[21:49] Go to market is getting more and more expensive at the same rate
[21:52] that product is getting cheaper. And this is also my, I guess
[22:01] interest and skill. I do think companies that can
[22:06] master, go to market or find ways to lower CAC are going to
[22:12] have a have the real Moat. Absolutely, Yeah, that's a very
[22:17] profound impact. That product is going to get
[22:19] cheaper and GDM is going to get more and more expensive with the
[22:22] rising competition. Quite interesting.
[22:25] And you mentioned about the platform versus point solutions
[22:28] thing, right? How every startup would need to
[22:32] have some sort of platform in order to, you know, sustain the
[22:34] business that they're doing and grow it further, right?
[22:37] Can you dive deeper into that with some examples on what are
[22:40] point solutions and what are platform solutions and why they
[22:44] will win? Sure.
[22:46] So let's say you're selling into a sales team and let's say you
[22:51] have a point solution, say an outreach AI solution where you
[22:55] can send emails through AI agent.
[22:59] It personalizes the e-mail based on the recipient's profile,
[23:04] right? And let's say you sell that
[23:06] solution for, I don't know, $99 or 299 a month right now.
[23:12] It all worked fine when you were using your own solution to do
[23:18] customer acquisition, but now it's becoming more and more
[23:22] difficult. Your cost of customer
[23:24] acquisition keeps rising. So there was this research by
[23:27] Winning by Design in Europe that I ran into last month.
[23:32] It they looked at top 50 SAS companies and they found that
[23:36] the customer acquisition cost for public SAS companies have
[23:41] gone up by 60% in the last three years.
[23:44] And it's, it's going to keep going up, right?
[23:47] So now if CAC is very high, let's say it's $1000, it's
[23:51] costing to you to get one customer.
[23:53] Now $99 doesn't pay for it, right?
[23:57] If it is like 5 dollars, $10 even $50 you would pay for like
[24:03] it it, it is still good business, right?
[24:06] But if it gets into hundreds of thousands of dollars, then it's
[24:09] not a good business anymore, right?
[24:11] So then this is an endpoint solution.
[24:14] So now what this company could do is one of two things, right?
[24:18] Either it can say, look, I can add because cost of building
[24:23] tool is cheap. So what I'll do is I'm, I
[24:27] already have 100 customers, right?
[24:30] So instead of charging $99, can I charge $1000 from these
[24:33] customers? Then I can pay for the CAC,
[24:37] right? And how do I charge $1000?
[24:39] I talk to the customer and find out what are the other tools
[24:43] they're using and can I build those tools because they like me
[24:47] and cost of building is anyway going down to 0.
[24:51] So they send emails now maybe they need data, maybe they do
[24:57] need to do enrichment or less and they're working with the
[25:00] data company, right? Can I build that myself and
[25:04] charge another $99, right? Can I maybe after sending an
[25:10] e-mail, they want to do a dialer or a call, Can I integrate A
[25:13] dialer and charge them another $99, right?
[25:17] So you'll see more come and and that starts becoming a platform
[25:22] going from a single point solution.
[25:24] So that is one way to go. But some founder could say,
[25:27] look, I cannot build all these tools myself.
[25:31] There are already those tools exist.
[25:34] So why don't we just partner because they have to acquire
[25:37] customers costing them $1000. So there's a dialer company or
[25:42] there's a data company, right? So this founder could go to
[25:46] dialer company and data company and say, look, you guys are
[25:48] probably seeing $1000 acquisition cost as well and you
[25:52] can't afford just like I can't afford.
[25:54] So let's ten of us get together and and spin up a platform and
[26:02] then we'll just subsidized our CAC.
[26:04] The same customer comes in, we'll sell the customer $1000
[26:09] goods and then individually we'll subsidized CAC, but split
[26:14] the revenue amongst us, right. So I foresee that there's going
[26:19] to be rise of numerous such platforms through partnership.
[26:24] Right now only large companies like HubSpot or Salesforce or
[26:28] Zoom Info can build these platforms in house because they
[26:33] have thousands of engineers and product managers to kind of
[26:36] invest small endpoint solutions. Obviously they can't build like
[26:41] end to end work flows, right? But they can partner with other
[26:44] companies that already have. So I do see rise of like on
[26:51] demand platforms using you know, tools.
[26:57] There are some integration tools now.
[27:00] So using those tools you can like tray or you know, tools
[27:03] like that you can put together or integrate multiple of these
[27:10] platforms together. So yeah, that's the that's kind
[27:15] of something that I see happen as the cost of CAC keeps on
[27:20] rising. Absolutely.
[27:22] It's quite interesting. I'm talking about Moat.
[27:25] You know, we have touched upon it previously in the
[27:27] conversation, but how do founders, especially working in
[27:31] the AI space, think about a Moat that's tangible enough,
[27:35] defensible enough in this environment?
[27:38] That's, that's a really important question for founders,
[27:44] right? Ultimately, you don't start a
[27:47] business if you don't have a mode or you don't have a, at
[27:51] least a theory towards a mode, right?
[27:53] And sometimes the assumptions you make could be wrong.
[27:57] But I, I believe that every business should continue to work
[28:01] towards a strong mode, right? So I'll talk about the mode that
[28:05] I have seen in AI. But another framework to think
[28:09] about the business is that I find useful is every month I
[28:14] question myself that whatever effort I made to make to
[28:20] generate say $1,000,000 in revenue, next month I should be
[28:24] putting in less effort or with the same effort, I should be
[28:28] generating more revenue, right. So the whole goal of business is
[28:32] to build systems and automation. So it keeps getting better every
[28:36] month. It keeps you keep subsidizing
[28:39] your effort for next month, right?
[28:41] And the way you subsidized the effort is you build systems and
[28:45] automations and insights and best practices that keeps making
[28:50] you better and better, right, In a very systemized way.
[28:54] The problem is most founders, they think the goal of running
[28:58] business is so they make revenue, they sell something to
[29:01] customer and make revenue. And if they make the same
[29:04] revenue with the same effort in month, one month to monthly,
[29:08] they still celebrate. I look at it and said now this
[29:11] is a really bad business because as the business grows, it'll get
[29:14] complex. You'll have to end up spending
[29:17] 3X5X more effort, right? The goal is as the business
[29:20] grows, you have to spend half X quarter X for the same revenue,
[29:25] right? And that's where you create
[29:26] growth flywheels, right? Now back to AI, how do you
[29:31] create that? So first of all, every business
[29:35] should think about how their customers are subsidizing their
[29:40] future business. Customers are not just for
[29:42] revenue, right? So most founders make the
[29:46] mistake thinking customers paid revenue.
[29:48] Their role is done. Customer should help you Co
[29:51] create the business, right, Co create as in say marketing help.
[29:56] If customers really believe in you, they could go on LinkedIn
[29:59] and you know, maybe promote you. So founders should figure out
[30:05] how to get the customers to promote right.
[30:07] Maybe the customers could provide data in addition to
[30:10] revenue that makes your business, you know, end up
[30:14] having some build some data assets.
[30:17] Now AI ventures, ultimately the performance of AI really depends
[30:23] on proprietary data and systems that you build around it.
[30:29] We are in the early stages. So all AI seems to be the same.
[30:33] They all seem to perform as good as chargeably, but there will be
[30:38] difference in performance and some will do.
[30:42] And we'll begin to see in coming days, months.
[30:46] And I think all AI ventures should strive on putting
[30:51] together some kind of data flywheel.
[30:54] So data contribution networks where users that are
[30:58] contributing data, nonpublic data get some incentive in
[31:04] return. And I think that's training AI
[31:09] around these vertical data sets is going to be the next
[31:11] battlefield. Absolutely.
[31:13] Yeah, that's again a very solid insight there.
[31:16] That data can actually become your mood even in the AI world
[31:21] as well, because that proprietary data is not
[31:23] available to every other company.
[31:25] So if that's available to you, then that's how it would, right.
[31:28] Quite interesting. So, so you mentioned about how
[31:31] the traditional GTM or the traditional approach to GTM is
[31:35] completely not working anymore and there's a new approach that
[31:38] needs to be taken. Do tell me about what's your
[31:41] take on that and what's working now?
[31:44] Yeah. So certainly some parts are
[31:46] working. So for instance, you know, if
[31:50] you're a small company and you don't care about your reputation
[31:55] because nobody knows you, so you don't mind sending out, say,
[32:00] 100,000 or million emails to your entire market because you
[32:04] just want some revenue that just destroys the channel for
[32:07] everybody else. That's just bad manners, right?
[32:10] Because now buyers are going to like shut off and never respond
[32:15] to anyone else, not just you, right?
[32:16] So, and that's we, we see that happening now.
[32:19] I talked to founders and they're like, OK, I just sent 10,000
[32:22] e-mail. How does it make a difference in
[32:25] the world is so big? There are millions of people
[32:28] like how is my 10,000 destroying?
[32:30] But turns out the cost of data and cost of automation, it has
[32:35] gone down so tremendously. So 20 years ago only DNB was
[32:40] selling data at like 300, three, $150,000.
[32:44] That was the enterprise to even use data, right?
[32:48] Automation they were like that was another 100 and 5200
[32:52] thousand. So even just to start sending
[32:54] those emails it would cost half a million.
[32:56] So IBM was doing it right and some large company, but not
[33:00] every startups today a company gets funded say $50 million.
[33:06] The CEO gets 100 emails because now at $50, every startup is
[33:10] sending some triggered e-mail and they're thinking this is
[33:14] signal based and this is healthy.
[33:16] And but I've had the CEO say like, what do I do with this?
[33:20] Hundreds of they find a way to send it all to junk, right?
[33:24] I feel like what was good, what worked, and perfectly rational.
[33:29] In fact, there's a good playbooks people are sharing
[33:32] amongst themselves, but if everybody starts using it, it
[33:36] becomes a terrible playbook. So I have the vantage point to
[33:39] look at what was good. In fact, I helped create this
[33:42] space, right? And I feel responsible.
[33:46] I didn't think far enough. It seemed to help everyone and
[33:49] now if everyone uses it, it's like totally destroyed the
[33:53] market right now. There are some pockets of hope
[33:57] that's working. So for instance, Clay can enable
[34:02] you to run like say micro campaign.
[34:06] Instead of sending 10,000 emails, you could send 10 at a
[34:08] time with lots of signals and lots of triggers and AI and
[34:13] those are healthy. They still perform well and I'm
[34:16] a big proponent of those emails. So there are many such pockets
[34:23] and workflows that still work. I do a lot of LinkedIn
[34:26] prospecting that works for me. Happy to double click on that.
[34:32] But what I am doing is more Biolet.
[34:37] So I feel like and I don't foresee that the market will
[34:40] entirely shift from sales to biolet.
[34:42] I'm just trying to have say 5% of these transactions to move to
[34:47] biolet, so buy a lead and agent to agent.
[34:52] So let me explain. There are two points of
[34:56] inefficiency I see in current go to market.
[34:59] One is 50% of go to market cost goes into lead acquisition or
[35:04] demand Gen. right? And then there there's a lot of
[35:09] inefficiency because you send 10,000 e-mail, only 10 people to
[35:13] find 10 people that are in market to buy.
[35:15] You do 100 calls across like 4 weeks only to find four or five
[35:20] that are interested, right? So, and everybody stays busy
[35:24] doing this call SDR teams and AES and so on.
[35:28] But then their, their cost structures are built around to
[35:33] have only like 1 percent, 2% conversion.
[35:36] And like there's a lot of wastage in the system.
[35:39] So one part of our solution is everybody that logs into our
[35:45] platform, we give them an agent, a seller and a buyer and so on,
[35:51] right. And then when sellers or buyers,
[35:55] because buyers have the same problem, number of vendors are
[35:57] growing earlier they could easily buy from like 20 vendor.
[36:01] Now when there are 200, how they going to buy, right?
[36:04] So the buyer cost goes up. So we are having agents talk to
[36:09] each other. So seller agent and buyer agent.
[36:12] So agents go to seller agent and then they have the seller agent
[36:15] pitched to them. And these agents talk to each
[36:18] other and when there's a match, then they set up calls between
[36:22] humans. So now instead of the seller
[36:28] doing 40 calls to find those one or two that are really
[36:32] interested, now those one or two calls are showing up on the
[36:35] calendar and skipping the other 39, right?
[36:39] And, and these are like already start with 30% likelihood of
[36:43] closing because buyers are looking at only like these
[36:46] matches are happening with only the top three, right?
[36:49] So, so that is one part of our solution.
[36:53] The other part is so that is the end deliverable, right?
[36:58] Since love us because of the high engaged needs, right?
[37:02] But then how do you get to that? So that's the other part.
[37:07] I think buyers, they need answers to some of their
[37:11] question that's burning for them and they need a personalized say
[37:15] grade, not a Gartner grade or AG two grade that is the same CRM
[37:20] grade for the whole industry. But every SMBs behave
[37:23] differently than mid market manufacturing behaves different
[37:26] than IT and services. And there's so much complexity
[37:29] in the world, right? But then there are all these
[37:31] one-size-fits-all grades. So what we do is we let the
[37:35] buyers decide, ask certain qualifying questions and we look
[37:41] at the data out there, send out an agent that's continually
[37:44] looking for solutions to problems.
[37:48] And then during that buyer journey, we allow vendors and
[37:52] agencies and consultants to participate and solve the
[37:56] problem for the buyer. And while they are solving, the
[37:59] buyer also hires them sometimes, right.
[38:01] So this becomes again an intend, high intend legion method.
[38:07] So the future of go to market I think is where buyers start the
[38:13] process and sellers help the buyer buy better, right.
[38:18] And then in the process they they end up selling their
[38:20] product. And that is just the healthier
[38:23] way. No 10,000 emails would be sent
[38:27] and still they will all get the same 3 or 4, you know, high
[38:31] intent leads in their calendar. Yeah, let me take a pause.
[38:35] That's a lot to absorb. Yeah.
[38:37] No, that's quite interesting. How agents are engaging with
[38:41] agents and discussing whether they should buy or not and then
[38:45] setting up calls of humans that OK, I find like these ones and
[38:48] you can look at them now quite interesting where we are
[38:51] heading, right? Great.
[38:52] So something that you mentioned on your LinkedIn ones is that
[38:56] you have to, you had to switch voice renders 3 times already,
[39:00] right? How should companies structure
[39:03] their tech stack for this constant change?
[39:08] Yeah. So AI is undergoing so much of
[39:11] change and churn. These vendors themselves are
[39:16] building new models and changing business policies, right?
[39:21] And they're some are raising more money than others.
[39:24] And so the entire market is for lack of better word, if I can
[39:30] say, unstable. So we did, we did run to a large
[39:39] vendor that over time we wanted to do wise cloning, they wanted
[39:45] enterprise license, they wanted us to, they changed their policy
[39:49] where they wanted us to pay hundreds of thousands of dollars
[39:51] for that feature which we thought would be included with
[39:54] low end pricing. So it then we switched, then we
[39:58] got a we went to a smaller vendor that were very open to
[40:02] our to working with us, but then the smaller vendor had
[40:05] performance issues. Because there was a tremendous
[40:09] lack from time to time. So there were infrastructure
[40:11] issues. So yeah, we are.
[40:15] And then I moved to like another vendor and had some some more
[40:20] issues. So my learning is that things
[40:26] are changing so fast. It's really important to stay on
[40:31] top because we could have wasted each time we switched.
[40:35] It took us a week too weak to change API and integrate to new
[40:39] vendors. But if we didn't peel off
[40:42] quickly, it would be even worse. We would have wasted even more
[40:48] time. And now we have built the code
[40:49] in a way that it will be really easy for us to switch to new
[40:53] vendor in the first time we thought we are never going to
[40:55] make this switch right? Second, third time now we are
[40:57] totally prepared. We are even anticipating say
[41:02] vendor switch if need be, right. So I think all founders should
[41:05] think about that switching LLMS or switching other vendors and
[41:10] providers is and once you find a good one, it's, it's good to
[41:15] just stick by it. But there is going to be some
[41:19] uncertainty and performance issues in AI in the coming until
[41:26] everything's stabilizes, right? Surely you mentioned how you use
[41:32] LinkedIn for prospecting, right? And you see, I see you're very
[41:36] active on the platform, right. So can you give us an insight
[41:39] into how you use the platform and use it to generate interest
[41:43] into the product that you're building and also gain traction
[41:47] from there? Yeah, yeah.
[41:50] I'm kind of early stage in what I'm building.
[41:53] So I'm not exactly selling on LinkedIn, but I've done that in
[41:56] the past say with RB to B, but I was in more in a sales mode.
[42:01] But the way I, so my playbook is I have chosen a certain ICP and
[42:08] content pillars that work for my ICP and my audience, which is
[42:13] around sales, marketing, founder, entrepreneurship and so
[42:16] on. And then I write content for
[42:20] those ICP, right and the business I'm working on.
[42:26] And then when people engage with that content, I have some kind
[42:29] of automation in the bucket where I scrub the people that
[42:34] are engaged with the people I want to talk to.
[42:39] And then when I want meetings, I just have that configure that
[42:43] automation to drop my calendar link with A1 line message.
[42:49] And I find I have like 50% conversion and these meetings
[42:53] come inbound. I don't even have to send
[42:55] emails, right? So because the key is I have to
[43:00] do this within a few hours of my post because people when they
[43:07] engage with the post, they've already gotten some value.
[43:10] And if I have an ask right after they got value, they generally
[43:15] want to reciprocate versus if it's a cold e-mail.
[43:19] So I'm using this this hello effect or this micro celebrity
[43:24] fact that lasts for a few hours with around content, right?
[43:29] And then it goes away and then people forget you.
[43:31] So I'm trying to leverage that into having an ask for
[43:36] prospecting. Absolutely.
[43:39] Yeah. That's a very interesting
[43:40] approach. It seems you have a high heavily
[43:43] bent on, you know, optimizing the high signal buyer intent or
[43:47] something who's already engaged with you in some way.
[43:49] Even the platform is built around that, that the Biden
[43:53] trust is there. You just have to mass them with
[43:55] the right product that is right for their needs, right?
[43:58] So that's amazing. Got it.
[44:01] So let's talk about the attention economy.
[44:04] Attention seems to be the scarceest resource and B to B,
[44:08] what's actually working to cut through the AI noise?
[44:12] You being AI product yourself, how do you make sure you stand
[44:15] out and try to make your product stand out and be heard in the
[44:19] noise that's all around you? Yeah, yeah, there's, there's so
[44:24] much of noise and the noise is increasing that we're reaching a
[44:29] stage where almost being different is more important than
[44:33] being better. So if you're making a building a
[44:36] new company, I would say try to do everything different if you
[44:39] can, right? And it's going to get
[44:41] increasingly more and more difficult to be different
[44:44] because you build a new product or add a new feature within a
[44:48] week or two, all your competitors will copy it because
[44:51] it's so easy to like build those interfaces, right.
[44:55] So I think go to market part leadership community, these
[45:02] things offer opportunities that are still untapped or being
[45:06] different product will increasingly not be the
[45:09] opportunity to be different. Now, some products that are hard
[45:12] to replicate and but they're, they're also expensive and far
[45:17] in between, like you can't just copy open AI, right?
[45:20] So, but a lot of the product that are more like others,
[45:27] you'll find that they'll still succeed if the founders are
[45:34] creating a community, creating trust, credibility to talk
[45:39] leadership or content like this. I think if people want to buy
[45:44] from others, they can trust when the market gets noisy.
[45:49] Absolutely. That's hitting the nail on the
[45:51] head. I think building trust,
[45:53] credibility and even a brand if you can in that noisy market, so
[45:58] many people competing for their attention.
[46:00] I think if you manage to build a brand on all the product, then
[46:04] that can really help in the buyer decision and get more
[46:08] traction for you, right. Let's talk, talk about the
[46:11] future of B to B, right. Five years out, how do you think
[46:16] the buyer seller relationship is going to look like?
[46:18] We are already seeing agent to agent interaction.
[46:21] So what's it going to look like in five?
[46:23] Yeah. I think about the future a lot
[46:26] given what I'm building. I think the future of B to B is
[46:31] B to C and increasingly we'll see more.
[46:35] So cost of products will keep going down, they'll they'll be
[46:41] more agent driven, more outcome driven if you think about.
[46:48] So there's going to be less sales driven because when you're
[46:52] selling for $1,000,000 a $100,000, you have sales led and
[46:56] it's still pays for. But if you're selling for more
[46:59] products for hundreds of dollar, then it's more marketing LED or
[47:02] more PLG, more community driven or alternative go to market
[47:06] models, right? I do foresee sales teams
[47:10] becoming much shorter than they are today, replaced by agents or
[47:16] marketing. I think.
[47:17] I foresee marketing becoming bigger, more investments going
[47:21] into brand building and into other marketing initiatives.
[47:29] I do, I do. I think AI and agents will play
[47:35] a big role, but in in a much different way than they do
[47:38] today. Because today what people are
[47:42] trying is either just providing more insights incrementally or
[47:46] they're trying to replace like a salesperson, but then it is
[47:51] replacing a salesperson but doing the same inefficient
[47:54] things they were doing, the salesperson was doing.
[47:57] And that's why the AISDR companies are not working as
[48:01] well. I think once the dust settles,
[48:07] AI is going to be used very differently.
[48:09] It'll be used like any other productivity tool.
[48:13] You're just going to have less number of reps their quotas will
[48:16] double and they'll need AI to like wipes do wipe selling or
[48:21] just to make themselves superhuman, right.
[48:24] So rather than have AI to replace them, so they'll know a
[48:30] lot more about the buyers, a lot more about the context, their
[48:34] own company and know how to, what to say when, how to and how
[48:37] to accelerate deals. And a lot of the inefficiencies
[48:41] that we have in go to market today, which is where half the
[48:45] employees are spending time, those inefficiencies will just
[48:48] go away. And I hope some of that goes
[48:50] away because of bio LED go to market that we are building.
[48:54] Yeah, absolutely. And moving to my last question,
[48:57] before we try to wrap this up, walk us through Zai Digital
[49:02] clone platform. What BDB use case use cases are
[49:06] you seeing the most traction with?
[49:09] Yeah. So this, this agent platform
[49:14] that we built, that's a small piece of the overall picture.
[49:18] It's going to take us another few weeks or a couple of months
[49:21] to build the entire platform. So that's why we have released
[49:25] just one part of the platform. But look at it this way, we
[49:30] don't talk about the bigger platforms.
[49:33] So a lot of the what I've shared is not exactly what I'm building
[49:37] because I feel like it will confuse people.
[49:39] So it's difficult for me to just talk about 1 component of it.
[49:44] But look at it this way, right? So there are a lot of chat bots
[49:47] on the websites you can go and it can and answer to your
[49:52] questions. But the role of those chat bots
[49:56] are really to drive you towards a meeting with a SDR or
[49:59] salesperson, right? Or if you're in a support side,
[50:03] then to resolve any tickets. There are no chat bots that are
[50:10] built for buyers that will that's designed to help the
[50:17] buyers, but is sitting on the vendor website because that's
[50:19] where buyers do their research, right?
[50:21] But is designed to help the buyer through the journey.
[50:26] And even if it means that the buyer ends up not buying with
[50:30] that vendor, but buying with a competitor because that's a bit
[50:33] of it, right? No chatbot and no AI is
[50:36] designed. All AI is designed to push the
[50:39] vendor, vendor solution to the buyer, whether they whether it's
[50:44] a best fit or not, right? So we are building this is kind
[50:48] of the key differentiation, we call it buyer concierge.
[50:52] So some vendors will take our solution, put it on the website
[50:56] and really help the buyer. And in the process they generate
[51:00] a lot of trust and credibility with the buyer, right.
[51:03] And we, our research is showing that buyers appreciate vendors
[51:09] that help them with the buying process.
[51:12] They have the first right of refusal to the revenue.
[51:16] Otherwise, the buyers don't even want to involve the sellers
[51:20] unless the sellers are cooperative to the buy process,
[51:23] right? So we are building tool and this
[51:25] agent platform is one of the first tools that will sit on
[51:29] vendor website. But it's been designed to help
[51:32] the buyer and it will allow the buyers to do like a micro RFP or
[51:38] ask some qualifying questions with the vendors.
[51:40] And if the vendor cannot answer, like if the agent cannot answer,
[51:44] it will escalate it and have humans find answers.
[51:48] And so there are small tools that are built into this
[51:53] platform to help the buyers buy better.
[51:57] Absolutely. Yeah, that's quite interesting
[51:59] what you're building there. All right, so with that, we come
[52:02] to the end of this conversation. Thank you for making time for
[52:05] this, Santosh. It's been an incredible
[52:07] conversation. Loved all the insights.
[52:09] They were unique, very profound bunch of those.
[52:13] Thanks for coming on and I wish you a happy building.
[52:15] Thanks. Prashanth and I enjoyed this
[52:17] call. Likewise, Thank you.

Shows are listed from Apple Podcasts' charts and directory, and Podcast Index where it is used. Every show and episode belongs to its publisher and plays from the publisher's own feed; this site hosts none of them.