Latest / AI for Founders with Ryan Estes / Your Vibe Code Just Handed Hackers Your Database - Punit Bhatia, Founder of Fit4Privacy

Transcript

Your Vibe Code Just Handed Hackers Your Database - Punit Bhatia, Founder of Fit4Privacy · AI for Founders with Ryan Estes

• 0:00
  Your AI product just went live, Congratulations.
• 0:02
  You also just handed a hacker the keys to your customers data
• 0:05
  and didn't even know it. When your vibe coding at
• 0:07
  midnight, credit card burning, riding that dopamine hit of
• 0:10
  watching your idea come to life. Speed kills.
• 0:13
  But it's not killing your competition, it's killing you.
• 0:16
  What if I told you that testing your data is something 100% of
• 0:19
  founders claim to do and almost none of them do right?
• 0:23
  There's a massive difference between running a test and
• 0:25
  running the right test. One makes you feel safe, the
• 0:28
  other actually keeps you out of court.
• 0:30
  You need an AI agent, a virtual privacy advisor trained on your
• 0:34
  own compliance policies, quietly watching every other agent in
• 0:37
  your stack like a security guard that never clocks out, never
• 0:41
  takes a lunch break, and never sends your client database to
• 0:43
  the wrong server. That's buildable right now.
• 0:46
  And here's the gut check that will save your company in my
• 0:48
  compliant and is my customer data safe.
• 0:51
  If you hesitated, you have work to do.
• 0:53
  Welcome to AI for founders. I'm Ryan Estes.
• 0:56
  My guest today is Panit Bhatia, privacy AI and security
• 0:59
  strategist and host for the Fit for Privacy podcast.
• 1:03
  And the guy founders call when their house is already on fire.
• 1:06
  In this episode, you'll learn how to build responsible AI
• 1:09
  products that don't blow up your brand, what e-mail compliance
• 1:12
  mistakes are quietly costing founders right now, and how to
• 1:15
  create an internal agent framework that's actually
• 1:18
  protected. This one's for founders building
• 1:20
  AI powered products who want to move fast and stay standing.
• 1:24
  Want frameworks, playbooks, and insights like this delivered
• 1:27
  straight to your inbox every week?
• 1:29
  The smartest founders we know are already signed up at AI for
• 1:32
  founders.co, where we turn complex AI strategy into moves
• 1:36
  you can actually make Monday morning.
• 1:38
  Don't just listen, level up. Join us at AI for founders.co.
• 1:42
  And if this episode made you think or made you sweat a little
• 1:45
  or saved you from a very expensive mistake, drop us a
• 1:47
  five star review. Founders help founders, and that
• 1:50
  one tap goes further than you know.
• 1:58
  Welcome to the show. Thank you so much for joining
• 2:00
  me. It's a pleasure to be with you,
• 2:02
  Ryan. I'm very happy to be here and
• 2:05
  thank you so much for having me. Absolutely.
• 2:08
  It's my pleasure. And we're going to have a
• 2:10
  conversation about security and privacy largely and we'll get
• 2:14
  into some other kind of facets of that conversation.
• 2:17
  But I think this is kind of like at the tip of everyone's tongue
• 2:20
  right now, largely because we're just out here vibe coding and we
• 2:24
  have no idea what the ramifications are going to be.
• 2:27
  Plus we have, you know, word from Claude that they have some
• 2:31
  new model that's creating, creating a potential problems,
• 2:38
  put it that way. Zero day events on across the
• 2:42
  board, all over the place. So give us an introduction.
• 2:46
  Now I want to make sure that we mentioned that you've got 2
• 2:48
  brands. One is fit for privacy.com.
• 2:51
  That's fit #4 for privacy.com and you also have Grow skills
• 2:56
  dot store. So that's just sitting the table
• 2:58
  a little bit. Talk to us a little bit about
• 3:00
  your work. Sure.
• 3:02
  So I'm a privacy sourcing AI strategist.
• 3:05
  So what does it mean if a company is coming into Europe or
• 3:08
  CI and AI foundation, not AI company, innovation company says
• 3:14
  hey, we are doing the right thing in AI, we have a nice
• 3:16
  product, but how do we make this privacy compliance, security
• 3:21
  compliant? And then they come across
• 3:23
  buzzwords like responsible AI, ethical AI, human AI and you
• 3:28
  name it. And then also they hear if
• 3:30
  they're in EU, some laws and I will spare you of those
• 3:32
  thoughts, but some laws like the GDPR and EUAI act but and then
• 3:37
  they say, what do I have to do? Hey, I have a nice product which
• 3:40
  can do fancy things. I can utilize AI and do service
• 3:44
  to humanity. Now what are these people
• 3:45
  talking about? So I help them convert that
• 3:48
  fluff in their language to reality on what do they have to
• 3:53
  do and what all they have to take care of so that they're not
• 3:57
  fined and they're on the right side of the law.
• 4:00
  So we do it in two parts. One, we had advised them and
• 4:03
  consult them and that's what we do in fit for privacy.
• 4:05
  We focus on privacy, AI and security and sourcing aspect.
• 4:10
  And then typically companies say we want to do train our staff,
• 4:14
  train our management, train our key personnel on different
• 4:18
  matters. So be it quality management,
• 4:20
  security, AI or anything. And that's what we do through
• 4:24
  growth skills dot store. So it's aggregated the
• 4:27
  consulting and the training part of it separately.
• 4:30
  And we do all kind of training standard off the shelf
• 4:33
  certifications of bespoke training, suited, customized,
• 4:38
  adapted to the management talk, adapted to the staff and live
• 4:41
  online or whatever they need. That's what we do.
• 4:44
  And me as a person, as I won't like to serve people, I have a
• 4:48
  podcast and I have a YouTube channel named Fit for Privacy.
• 4:53
  And there we bring knowledge and experiences from experts and
• 4:57
  industry experts to people. And we also do webinars on a
• 5:00
  monthly basis. So that's how it is broadly
• 5:04
  speaking. Of course, we'll get into more
• 5:05
  details. Incredible.
• 5:08
  So I'm talking to the right person now.
• 5:09
  I'd like to kind of maybe start with a case study.
• 5:12
  So I talked to probably a dozen founders a week.
• 5:15
  I talked to a lot of founders and CE OS and what I'm seeing is
• 5:20
  they're working on their core offer a course in building their
• 5:23
  business. They're incorporating AI tools
• 5:26
  in different capacities within their business, within their org
• 5:29
  chart. And through that they're
• 5:31
  starting to to create products on the side.
• 5:34
  So either they're taking kind of the internal infrastructure
• 5:38
  tools that they're building for their team and being like, hey,
• 5:40
  I could spin this into a project product and, and starting to
• 5:43
  bring that to market, or they have their pet project of
• 5:47
  something new that they're kind of working on now.
• 5:51
  I I suppose there's like a, a double side of this kind of vibe
• 5:55
  coding coin, which is like there's this unbelievable
• 5:57
  elation that just wants you to token Max all of your credits
• 6:03
  until The thing is, is done. But there's also this sinking
• 6:07
  feeling in the in the background that's like, oh boy, like what
• 6:11
  are my what's my exposure and what are my vulnerabilities?
• 6:14
  So for folks that are kind of deploying these AI products now,
• 6:19
  what is kind of some words to the wise as far as the caution
• 6:22
  that should be taken right now? Well, you're asking me to
• 6:27
  synthesize all the consulting that I do in a few words, but
• 6:30
  let's try. So basically what we call is be
• 6:34
  responsible in the development of AI and what does it mean?
• 6:37
  Let's decipher it. Being responsible means you have
• 6:42
  the product and you have the service in mind, but make sure
• 6:47
  it's not infringing upon any of the laws, like you're testing
• 6:51
  for hallucination bias and things like that.
• 6:54
  All the things you are concerned that your product shouldn't
• 6:57
  have, it's not processing any personal data or it's keep you
• 7:01
  respecting the privacy of individuals.
• 7:02
  But what we do is we give them a set of principles or rules if
• 7:06
  you bought it or a policy, whatever you like to call a
• 7:09
  document saying this is what your developers, your team have
• 7:13
  to take care of. So we align with management and
• 7:15
  say, let's do you think you want to be ethical?
• 7:18
  Yes. Do you want to protect privacy?
• 7:20
  Yes. Do you want to be compliant with
• 7:22
  law? Yes.
• 7:23
  Do you want to be transparent? Sometimes yes, sometimes maybe
• 7:26
  that's how the answer is. So we give them these rules, we
• 7:30
  put them into policy and say, now let's train your stuff and
• 7:35
  you as manager, govern it. Govern it, meaning make sure
• 7:40
  that these are being followed. So we put practices and controls
• 7:43
  through which these are being kept in place.
• 7:45
  Like let's say we want to be responsible AI company.
• 7:48
  That means our data set should be tested.
• 7:51
  So are you doing testing? That's a simple control.
• 7:54
  And is your testing strong enough?
• 7:56
  Are you testing for bias? Are you testing for
• 7:59
  hallucination? Are you checking data quality?
• 8:01
  Now these are basic questions, but we formulate them and we
• 8:04
  align and harmonize with law. And that becomes your
• 8:07
  responsible AI tool kit or guidelines.
• 8:09
  And that's what you do. And then at the end of the year
• 8:12
  or two years or just before the product launch, say, hey, nice,
• 8:16
  you heard it, you put it in the document, you checked it.
• 8:19
  Now let's check if that is actually true.
• 8:22
  So we do what we call a self-control assessment or
• 8:25
  assessment by us, self-control assessment is they themselves
• 8:29
  check against some controls and by us is we check it and we say
• 8:33
  let's check if you're playing by the book, that is the book we
• 8:36
  created at the start. And if there are gaps, no
• 8:39
  offence, let's fix it before we go to the market.
• 8:41
  As simple as that. Amazing what?
• 8:44
  What percentage of businesses are actively testing their data
• 8:47
  right now do you think? Let's say how many say they are
• 8:51
  testing it 100% if nobody says we are not testing it, but the
• 8:58
  issue is are you testing right and are you testing enough?
• 9:03
  Yeah, and that's where the gap is, but everybody's testing it.
• 9:06
  I would be really worried if somebody's to, you know, putting
• 9:08
  it into the market without testing or says we are not
• 9:11
  testing. Yeah, you know, it, it rings
• 9:14
  especially true kind of in my world, you know, startup world,
• 9:18
  you know, startup founders want to move fast, do the fun stuff,
• 9:20
  let's talk go to market, let's talk talk ads and, and and sales
• 9:25
  and like all that fun stuff. But often times the biggest
• 9:28
  gains are just from fixing the leaky bucket, you know?
• 9:31
  Yeah, I think a lot of times, especially in startups, this is
• 9:34
  the area that they really need to focus on compliance because,
• 9:38
  you know, one, you're, you're facing legal repercussions if if
• 9:43
  you break the rules and also you're leaving yourself
• 9:45
  vulnerable for attack, you know, So maybe maybe twist that dagger
• 9:50
  a little bit. You know what are what are some
• 9:51
  of the some scary stories you've heard where people have come to
• 9:55
  you when it's too late? Well, one guy came and he had
• 10:00
  sent emails to an e-mail tool without thinking too much,
• 10:06
  without worrying about any of the privacy laws or any concern
• 10:09
  for individuals and started sending it.
• 10:12
  And they had configured the whole CRM system, you know.
• 10:16
  So everything was set, all the emails are configured and the
• 10:18
  whole database is set up. And somebody complained and now
• 10:22
  he comes to me and says got to fix it.
• 10:25
  And I say you have to do this. And he says, oh, but that means
• 10:29
  I will not have anybody to send e-mail to and we will have to
• 10:33
  start from scratch. I said, now that's not my
• 10:36
  problem because if you've come one or two years earlier, we
• 10:39
  would have set it up in the right way.
• 10:41
  Now you set it up is a mess and that's a problem.
• 10:43
  And second time it was there was a client of ours and we asked
• 10:48
  them, I was explaining you the rules and the principles and
• 10:52
  they said yes, yes, yes, yes, yes.
• 10:53
  And then when we went into testing, the second testing that
• 10:56
  we do or checks and we found a script that was reading all the
• 11:01
  personal data while the rules and the book and the guidance
• 11:05
  for the script said you shouldn't do it.
• 11:06
  And then I asked the developer, and the developer says, I know
• 11:09
  it was easier to copy all data. So we copy all data, but we
• 11:12
  don't read it. You don't read it, but the
• 11:18
  hacker can, and anybody with malicious intent can.
• 11:22
  So you better not copy it if you don't need it, copy what you
• 11:25
  need. That's the basic golden
• 11:26
  principle we work with. So use the data as necessary and
• 11:29
  copy the data or take the data as necessary.
• 11:33
  Quick side note for founders who are using AI to write content,
• 11:36
  you've probably noticed this. It's incredibly fast, but it
• 11:39
  doesn't really Dr. meaningful traffic.
• 11:42
  And honestly, that's because most AI generated content is
• 11:45
  just average. There's no real insight, no
• 11:47
  depth, no structure, missing things like internal linking,
• 11:51
  and Google's getting very good at filtering that out.
• 11:54
  So instead of trying to just prompt better, I actually came
• 11:57
  across something interesting called Neena.
• 11:59
  It's a system that combines human strategy and review with
• 12:02
  AI writing to create high quality SEO optimized blog
• 12:06
  content that's actually designed to rank and keep ranking.
• 12:10
  What I like about it is that it's done for me.
• 12:12
  Their team handles the topic strategy and the content review
• 12:15
  and data markups for Google and turns blog posts into actual
• 12:19
  long term traffic assets. So if you're a founder thinking
• 12:23
  seriously about organic growth, this is one of those things
• 12:25
  worth paying attention to. You can check it out at Try
• 12:28
  nina.co. That's TRYNIN a.cotrynina.co.
• 12:36
  Every week I sit down and ask myself one question.
• 12:39
  What do founders actually need to know about AI right now?
• 12:42
  Not the think pieces or the stuff that sounds impressive,
• 12:45
  the real useful, put it to work today stuff.
• 12:48
  Then I write it up, send it to you in your inbox, free.
• 12:50
  That newsletter is called AI for Founders, and I built it because
• 12:53
  I was tired of wading through hit pieces and people that are
• 12:56
  afraid of the future. Each issue I break down the
• 12:58
  tools worth your time, the strategies that are working, and
• 13:01
  the ways you can use AI to grow your company without burning out
• 13:04
  or getting left behind. Thousands of founders read it
• 13:07
  every week and I'd love for you to be one of them.
• 13:09
  Subscribe for free at AI for founders.co. 3 seconds to sign
• 13:13
  up and I promise it's worth it. So those are some of the simple,
• 13:16
  basic things people do wrong, and it's amazing that it still
• 13:19
  happens. Yeah.
• 13:21
  You know, email's a big one because a lot of the AI tools
• 13:24
  given to us now allow for a lot of different kind of
• 13:27
  sophisticated automations. You know, you, there's the whole
• 13:30
  instantly cold outbound part of the world.
• 13:33
  There is everything that's built in your inbox.
• 13:36
  There's great product called Warm Start.
• 13:38
  There's a lot of things that we can start to manipulate with
• 13:41
  with e-mail when you're talking about kind of like falling out
• 13:45
  of compliance. Is this largely for folks out of
• 13:48
  the United States? I kind of understand that, you
• 13:50
  know, in Europe they have different e-mail standards.
• 13:52
  There's different differentiation and standards
• 13:54
  between work and personal emails.
• 13:57
  Maybe. What are some areas in e-mail
• 13:58
  that that founder should be aware of?
• 14:01
  So when it comes to e-mail, the first thing is do you have the
• 14:06
  right to send those emails? So let's say I come to your
• 14:10
  website or you come to my website and you sign up for a
• 14:14
  webinar. Now I signed up for a webinar
• 14:18
  and you signed up for a webinar. That's fine.
• 14:20
  But afterwards you send me a product offer.
• 14:24
  Now that's a questionable thing. You can send it, but did you
• 14:28
  tell me you will send it? That's the first question.
• 14:31
  Second, did I agree that you can send me marketing information?
• 14:35
  That's the second question. Now we can always play with it
• 14:39
  saying I told you so or I ask separately, mind the US, the
• 14:44
  model is it preferred to tell you and the Europe the model is
• 14:49
  you ask explicitly. Now that's the way to do it.
• 14:53
  But most people, they don't do neither of that and they say no
• 14:58
  problem, the customer signed up, we will send him bunch of emails
• 15:01
  and that's where it goes wrong. They're not supposed to send
• 15:05
  because I didn't sign up for spam mail list.
• 15:08
  I signed up for a webinar or you signed up for a webinar.
• 15:11
  That's what people are for when you respect and of course it's
• 15:14
  nice you send them a message saying since you joined our
• 15:17
  webinar, we would be adding you to your our sort of newsletter
• 15:21
  or marketing emails. If it's not OK, please opt out
• 15:24
  or please opt in based on the region you are and the strategy
• 15:27
  you have and then that's OK. But don't just assume that's
• 15:32
  where it goes wrong. Boy, and does everybody does
• 15:35
  that? That's a major blind spot.
• 15:41
  Yeah. And I mean, especially now
• 15:43
  because, I mean, your inbox probably looks like mine.
• 15:45
  It's just like, you know, it's 200 emails a day, you know, I
• 15:49
  mean, and we're talking about major companies.
• 15:51
  You know, if, if there wasn't like, you know, Nordstrom at
• 15:55
  some point was sending me like 2 emails a day.
• 15:58
  You know, it's like this is unbelievable.
• 16:00
  But I mean, they do it because it works.
• 16:02
  So there's this kind of Gray area where they're like, hey, we
• 16:06
  didn't, we didn't explicitly say you want me to e-mail you 2
• 16:09
  times a day. But we know that it, the more we
• 16:11
  pump these emails out, the more shoes and socks we're selling at
• 16:14
  Nordstrom. So, you know, this is just a
• 16:16
  potential roll of the dice, I suppose that people are playing.
• 16:20
  What are some other really kind of common areas that people are
• 16:23
  kind of maybe exposing themselves unnecessarily?
• 16:27
  I would say first and foremost, people think it's a legal or
• 16:31
  compliance issue and it's a brand perception issue because
• 16:35
  once you're in the media that you did this, the brand
• 16:39
  repetition is gone. It's a company issue, it's a
• 16:42
  trust issue. It's more than a compliance
• 16:44
  issue. But what was typically wrong is
• 16:47
  management of emails, management of data, not paying attention.
• 16:51
  It's not not paying attention, not being aware of how to
• 16:55
  protect data, keeping data safe and misusing data.
• 17:01
  So you are like you were talking about I gave you or you gave me
• 17:05
  the data for webinar. It's OK to send me details of
• 17:08
  the next webinar or of course, once in a while and once in a
• 17:11
  week. But not 2 emails a day or three
• 17:13
  emails a day and bombarding me with every time 15% off, 10%
• 17:17
  off, 20% off. Give me some value, give me some
• 17:21
  value and one in five, one in six emails you can offer me.
• 17:24
  Of course, that's fine, but not every e-mail.
• 17:27
  So balance out value versus marketing is what I say.
• 17:31
  Rather than sales, build direct sales.
• 17:35
  And the other thing is paying attention to protecting your
• 17:37
  data. Yeah, across regions.
• 17:40
  So earlier I was talking about protecting in your database or
• 17:43
  in your systems, but also across regions.
• 17:45
  Where is your data stored? In EU, in China, in India, in
• 17:50
  Philippines, in US, where you got to be aware of it?
• 17:54
  Most business owners are not aware of it.
• 17:56
  And like, I don't know, we have AWS, we have Google Cloud, we
• 18:01
  have this cloud, yeah, but the cloud sits somewhere.
• 18:04
  It's not in the cloud. It's in a salt mine in Utah if
• 18:09
  you're lucky. But yeah, people like I got in
• 18:12
  the cloud, it's it's up there. Yeah.
• 18:15
  So I'm not saying where you have to store it, but I'm saying
• 18:18
  where you store it has some consequences for you.
• 18:21
  Yeah, Yeah. That's a good point.
• 18:23
  You know, and, and particularly now as people are, you know, I
• 18:27
  think a lot of folks, you know, six months ago gave open AI
• 18:32
  access to their whole staff. It was like, hey, figure out,
• 18:35
  use it, do it. Looks we have to be AI native,
• 18:37
  you know, and people just started uploading absolutely
• 18:40
  everything to Sam Altman and now they're starting to backpedal
• 18:44
  and say like, Oh my gosh, OK, what, what have we done?
• 18:47
  What you know, what IP have we, we given to training data to
• 18:51
  open AI. And you know, what personal
• 18:54
  information is the CEO Social Security number now training
• 18:58
  data and is showing up as some answer in somebody's query.
• 19:02
  You know, so a lot of folks are starting to think of, you know,
• 19:04
  let's get some Mac minis and let's start storing all of our
• 19:07
  data locally. Let's bring in a local LLM.
• 19:10
  So we're not pinging, you know, the, the large model servers.
• 19:15
  What? What's your kind of take on
• 19:16
  that? I think that's depends on what
• 19:20
  you're going to use your LLM for.
• 19:22
  If you're using LLM more as a Google more as a help for you or
• 19:27
  your staff, that's OK used use the standard data wherever it is
• 19:32
  doesn't matter because you are maybe making holiday plans.
• 19:35
  Maybe you're checking what your marketing strategy should be.
• 19:37
  Maybe you're checking how is your your competitor doing.
• 19:40
  So that's fine. But as soon as you get into
• 19:43
  managing your internal stuff, let's say I want to upload an
• 19:47
  internal policy and ask it to review for me or I want to
• 19:51
  upload a client policy and ask it to review for me.
• 19:54
  In those things, you got to be making sure let it in your safe
• 19:58
  environment, licensed environment with the right
• 20:01
  settings at the right place and in a, as you said, local
• 20:05
  database or local setup of LLM. So you need to make that
• 20:09
  judgement around what you're doing and what's the sensitivity
• 20:13
  of it. And then of course, the
• 20:14
  judgement around what we were talking, you don't upload your
• 20:18
  client database and say these are the names of people can send
• 20:20
  the emails. Well, you have better ways of
• 20:23
  doing it, no? Yeah, absolutely.
• 20:29
  Yeah, interesting, you know, in, in around this idea of training,
• 20:34
  particularly with AI and kind of figuring out the right way to do
• 20:38
  it. You have several offers of
• 20:41
  training certifications, which is really interesting, you know,
• 20:46
  largely because it, we have certifications for artificial
• 20:50
  intelligence, professional AI manager, lead auditors.
• 20:54
  So, you know, I just thinking about creating a foundation of
• 20:58
  good practice. You know what what?
• 21:00
  What is the certification for AI professionals and AI managers?
• 21:05
  That's a very good question. See people when they talk about
• 21:09
  AI, did they think it's all about coding and development?
• 21:13
  Well, that is one part of it, but then it's about management
• 21:17
  of risk in AI. So that's where this certified
• 21:20
  AI risk manager kind of training comes in.
• 21:23
  Then it's about you are a manager, you manage a team of
• 21:26
  developers, you manage this risk manager, you manage the
• 21:29
  stakeholders, you manage the CEO, but you don't know how what
• 21:32
  they do. So you better do a certified AI
• 21:35
  manager or certified AI professional to kind of start to
• 21:38
  understand when they say I'm doing this, I have my code and
• 21:43
  this is doing this, they'll use some words and you will have no
• 21:46
  idea about it. So get familiarized with their
• 21:50
  vocabulary then because you are aligned with the thinking, they
• 21:54
  will be able to give you a better output.
• 21:56
  So that's the certified AI professional.
• 21:58
  But then as a manager or CEO, you're like, I have built a
• 22:02
  wonderful AI product and I want to tell the world it's safe AI,
• 22:06
  it's responsible AI, it's ethical AI.
• 22:08
  Now what do I do? Get someone trained in ISO
• 22:12
  42,000 and one, get it implemented as a practice in
• 22:16
  your company and then you tell your staff to do an audit and
• 22:21
  then you go to your client and say we are ISO 42,031 certified.
• 22:26
  Well then people will listen to you.
• 22:29
  They will be more comfortable buying your product.
• 22:31
  I'm not against AI. I'm rather for AI because for
• 22:34
  me, AI is like a knife. Whether you cut your finger on
• 22:38
  onion or bread, that's entirely your choice.
• 22:41
  But just because you cut your finger, it's not an AI fault or
• 22:45
  the knife fault. Use AI wisely, with protection
• 22:50
  and with and more importantly, be transparent.
• 22:53
  Yeah, remember these days when you buy a knife, it's written
• 22:56
  uses in some countries. Use it wisely.
• 22:59
  It can. It's very sharp.
• 23:00
  It can cut your finger. It can be dangerous to you.
• 23:03
  Same thing. AI is dangerous.
• 23:05
  So be transparent. Tell what it does, tell what it
• 23:07
  doesn't do and put it in a maybe say product notice or product
• 23:12
  sheet or end user licensing agreement wherever it fits or
• 23:15
  wherever you deem appropriate technology wise.
• 23:18
  What does it do for you? Yeah, yeah.
• 23:21
  It's like it doesn't matter who you blame.
• 23:22
  And you cut your own finger, you know, you're feeling the pain.
• 23:27
  You know, it's very tight. Yeah.
• 23:29
  Well, and plus everybody's spitting up agents now, you
• 23:31
  know, so basically we're all managing a fleet of agents.
• 23:34
  So it's almost, you need to have the basis of security and
• 23:37
  privacy built in, you know, to your for yourself.
• 23:41
  So there's a framework as you're starting to train out your staff
• 23:43
  of AI agents. That's a very tricky part and I
• 23:48
  don't know how we will manage it.
• 23:50
  Let's me be honest about it. I don't know how this agent
• 23:53
  story will go because I can create an agent for pretty much
• 23:57
  everything, basically anything, anything repetitive or see
• 24:00
  workflow based or process based that my staff used to do, I can
• 24:03
  do an agent and it's very good. And it has immense power for
• 24:08
  business, for humanity and for society in the long term.
• 24:11
  But how do we manage it and how do we control it?
• 24:13
  Because with the Tom or John, I would use to ask, hey Tom, how
• 24:18
  did you do this? Can you check?
• 24:21
  And I then will ask John to check it.
• 24:23
  But with the agents, OK, I'll ask one agent, which is Tom and
• 24:26
  another agent which is John to check Tom's work.
• 24:30
  And then where is the accountability?
• 24:31
  Because if Tom and John, I could replace them, I could fire them.
• 24:35
  Now with this AI Tom and AI John, I don't know who will be
• 24:38
  fired and who is accountable. Because the end of the day, me
• 24:41
  setting up the agents is the accountable party.
• 24:44
  But that's where there's a huge amount of risk.
• 24:46
  But I think we will find a way if we follow the principles of
• 24:49
  Responsibility Act, if we follow the law and if we do the right
• 24:53
  thing, we will find a way. But the challenge become when
• 24:56
  people, as you earlier mentioned, cut corners and in
• 24:59
  essence of speed and profits go a little bit too fast.
• 25:02
  That's where the challenge is. Challenges, not with their
• 25:04
  challenges, would be us trying to be greedy.
• 25:07
  Yeah, well, it's interesting because what I want now is I
• 25:12
  want an AI agent of Punit and I I want him to check on all my
• 25:18
  other agents to make sure that there's security and privacy.
• 25:21
  You know, ultimately, you know, there's a marketplace of agents
• 25:24
  that people are building their own internal agents like this
• 25:27
  pretty good. You know, here you can come pick
• 25:29
  up these skills or pull it off my git or whatever.
• 25:32
  It would be kind of amazing, you know, to have that's like
• 25:35
  another agent, which is like security and privacy.
• 25:37
  So before things are deployed or after things, or you know, if
• 25:41
  you're checking data, then you have that agent that's
• 25:43
  continuously checking with all the agents, making sure they're
• 25:46
  compliant. I agree with you and that's the
• 25:50
  way I see future. You can have a virtual security
• 25:54
  advisor or privacy advisor set up and use it to guide you.
• 26:00
  But what does that virtual privacy Advisor or security
• 26:03
  advisor feed on? The policy, the guidelines, the
• 26:08
  rules, me and the C level or the executive level will create.
• 26:14
  So we create those and feed them into your AI or LLM internally
• 26:18
  and then your LLM can be supporting the implementation of
• 26:21
  that. So that's how I see it.
• 26:23
  So you're very much on track that we need these privacy and
• 26:26
  security and the AI advisors going forward.
• 26:29
  Yeah, and. And I think it's going to become
• 26:31
  more and more of an issue. I remember who had that big
• 26:35
  photo database back in the day. They're like stock photos.
• 26:41
  I want to remember their name because they are kind of doing
• 26:43
  some nefarious things. It'll come accents.
• 26:47
  Are Shutter stock? No, both of those are fantastic.
• 26:50
  It was, you know, it was maybe 15 years ago.
• 26:54
  Anyway, what what they did is, is they flooded their entire
• 26:58
  catalog onto Google Images to basically when people were
• 27:02
  looking for pictures, they, they, they put them up.
• 27:05
  And then invariably, you know, web developers would take those
• 27:08
  pictures and use them as stock photography on the websites.
• 27:11
  And then they would do reverse image lookups and then basically
• 27:15
  send, you know, angry letters say like, hey, you stole our,
• 27:19
  our, our intellectual property. You owe us $4000.
• 27:23
  And they, they did that notoriously forever.
• 27:26
  And I, I don't think it had any legal footwork and I don't think
• 27:31
  they even wanted to press it, but enough people were scared
• 27:33
  that they paid them like the release money.
• 27:35
  They're like, OK, OK, sorry, I, I made a mistake.
• 27:37
  Here's the money. And that was basically their
• 27:39
  revenue model. And what I could see happening
• 27:43
  is, is something very similar because we have the power to,
• 27:46
  you know, basically, you know, publish what whatever it is,
• 27:51
  some data or a pixel or something here and there that
• 27:53
  people can trace, say it's mine, and then reverse engineer and go
• 27:56
  after everybody who used it and say like, hey, you owe us
• 28:00
  licensing rate of XYZ. They'll figure out the number
• 28:03
  and then basically, you know, extort them, so to speak.
• 28:08
  That's a big challenge in the AI world because now you're lifting
• 28:11
  the company. But what if AI does it?
• 28:14
  If you're AI agent or M&M or somebody does it, who is to be
• 28:18
  blamed? You, the creator of agent or the
• 28:22
  configuration manager of agent or the agent?
• 28:25
  Well, if you're blaming the agent then you're in for the
• 28:28
  wrong place because that doesn't exist physically and can't be
• 28:31
  penalized. Yeah, 100%, yeah.
• 28:34
  And that's another thing that we're kind of wandering into is
• 28:36
  like, yeah, who's liable for the actions of the agents?
• 28:40
  You know, they're they're getting more and more autonomous
• 28:42
  and ultimately they're kind of tracking back to one person, but
• 28:45
  a lot now you have agents directing agents.
• 28:48
  So it and it becomes like really relevant right now.
• 28:51
  I mean, we haven't seen anything, I don't think super
• 28:54
  catastrophic yet, knock on wood. But you know what, what Claude's
• 28:58
  talking about with that Mythos model sounds like pretty crazy,
• 29:02
  you know, like the capabilities. Now, a part of that also is this
• 29:06
  seems to be the new hype cycle is every time there's a new
• 29:09
  model, they come out and, you know, try and scare us so that
• 29:12
  they could build regulatory moats around and what what
• 29:16
  they're doing, what they're doing individually, you know,
• 29:21
  and then they leak it out six months later and you're like,
• 29:23
  oh, it's a little bit faster, it's a little bit better.
• 29:26
  But this one seems a little bit different now.
• 29:27
  What's your take on Mythos and kind of what that's telling us
• 29:31
  about security and privacy? It tells us that security and
• 29:36
  privacy is in a very difficult place.
• 29:38
  Let's be honest. And typically what tends to
• 29:42
  happen with laws or privacy or security, these evolve when the
• 29:46
  threats come because you have a problem and then you create a
• 29:52
  law and same thing from privacy and security perspective, people
• 29:56
  don't even know that's a threat. That's a challenge.
• 29:58
  And so that's where the responsibility lies with the
• 30:02
  creator or the business owner to do it right and to be
• 30:06
  responsible and to be educated on what is expected.
• 30:09
  And then if you take a fundamental principle based
• 30:12
  approach, I will be, I'll give you an example.
• 30:15
  I was working with the company and we said you should be
• 30:17
  transparent. They said one of their values is
• 30:20
  transparency. So we created a policy for them
• 30:22
  and said, let's put it on your website.
• 30:25
  You said, hey, that's not possible.
• 30:28
  And I said what? You are transparent.
• 30:30
  You said transparency is one of our values.
• 30:32
  So in the policy we created, what are the rules you will
• 30:35
  comply with? No, but our competitors can see.
• 30:38
  But there's nothing that the competitors can see.
• 30:41
  It's no proprietary model, no proprietary code, nothing.
• 30:45
  And I had a huge challenge and they didn't agree.
• 30:48
  And I had another one who said transparency the value.
• 30:50
  And they were very, very willing to.
• 30:52
  So that sometimes these privacy, security or laws also emphasize
• 30:56
  or amplify how true you are to the values you are projecting to
• 31:00
  the outside world. You're saying transparent, but
• 31:02
  you're not willing to put in a small policy of five pages.
• 31:05
  It just says, what are the rules you abide by and you create.
• 31:08
  I think customers will create stress, but you are a
• 31:10
  transparent company. Now that's, you know, an ethical
• 31:13
  dilemma most companies have. So whether meters or whether
• 31:17
  Claude or anything or even the DeepSeek or all these that that
• 31:20
  are coming up, they are a huge, huge challenge.
• 31:24
  Privacy and security and laws are not yet 100% adapted to
• 31:29
  that. So if you look at the principles
• 31:31
  of the law, the guiding principles, the ethos of the
• 31:34
  law, then you're safe if you apply that way.
• 31:37
  But if you're looking for where is the line item that doesn't
• 31:40
  allow me to do it, you won't find it.
• 31:42
  And then the law allows you. But if you look at you've got to
• 31:46
  be responsible and ethical and taking care of privacy of
• 31:50
  individuals. Am I doing it right?
• 31:53
  And typically I tell my clients, if you had a 16 year daughter or
• 31:58
  a son, would you let them use this product?
• 32:01
  Would you find it safe if they use it?
• 32:02
  Or maybe if you have 12 year old daughter and son or son?
• 32:06
  And sometimes they're like, but it's not made for kids.
• 32:09
  Did you say it in your product that it's not made for kids?
• 32:12
  These are the kind of checks you have to put in place.
• 32:15
  So if it's good for you, if it's good for your kid or a daughter
• 32:18
  or a wife or a mother or a father, it's a good product.
• 32:22
  So don't look for hiding behind the law.
• 32:25
  I tried to find the law and it didn't say you can't do it.
• 32:29
  Did your heart or did your head agree in unison that you can do
• 32:33
  it? That's where I say I said it's a
• 32:35
  brand and a trust issue. It's not a compliance issue.
• 32:38
  It's a trust issue. If you trusted for your family,
• 32:42
  if you're trusted for your near and dear ones, it's a good to
• 32:45
  like they have these financial aggregators.
• 32:47
  Give me access to all your bank accounts and I would create a
• 32:50
  nice graph and tell you what your projected savings are going
• 32:53
  to be in 10 years. Yes, why not?
• 32:56
  Would you use it? Would you attach your bank
• 33:00
  accounts? Probably not.
• 33:03
  And you expect your customers to do it.
• 33:06
  So don't create a facade of things which you won't use it.
• 33:09
  And if you will use it, if you're tech savvy, then go to
• 33:11
  somebody who's a little bit less, you know, your grandpa or
• 33:16
  your mom or your father and ask that question and then factor
• 33:19
  their concerns in. That's the way to do it.
• 33:22
  It's a good gut check, you know, would you have your kids use
• 33:25
  this? And it, it makes me a little bit
• 33:27
  skeptical. You know, like Chamoth is kind
• 33:29
  of famous for saying like he wouldn't let his kids, you know,
• 33:32
  use social media. And it's like he came, he was
• 33:34
  like OG Facebook guy, you know, And it's like, well, man, don't
• 33:38
  you? That's kind of an ethical
• 33:40
  conundrum and maybe just speaks a little bit about your
• 33:43
  character. If like you're like, oh, this is
• 33:44
  great. Yeah.
• 33:45
  Check it out. So it's printing cash.
• 33:47
  Of course I'm not letting my kids use it.
• 33:52
  So different rules, one for you and one for others.
• 33:54
  Yeah, exactly. You, you spoke about your
• 33:56
  clients a little bit, Who are your favorite clients and we are
• 33:59
  these enterprise, are they SMBs? My typical clients are SMBs who
• 34:07
  are looking to expand into EU or looking to comply with the laws
• 34:11
  and do writing, do trying to do the right things so they can be
• 34:14
  in healthcare, they can be innovative AI companies and
• 34:19
  they're in the launch phase. That's where they are.
• 34:22
  Yeah, interesting. And that that brand was Getty
• 34:26
  Images. Shout out to those guys.
• 34:28
  Copyright. Trolls Getty.
• 34:31
  Yeah, Getty. It's the copyright trolls.
• 34:33
  I, I, I don't know what, how that ended up.
• 34:35
  Maybe I'll do some research and double back on it.
• 34:37
  But I remember it was a, a big, you know, demand letters and
• 34:41
  legal extortion letters. Anyway, we're, we've reached the
• 34:47
  time in the podcast where we're going to enter the quantum
• 34:50
  round. Are you ready for the quantum
• 34:52
  round? Let's go for it, all right?
• 34:55
  We're going to do the the founder challenge to kick things
• 34:58
  off and the founder challenge is brought to you by kitcaster.com.
• 35:02
  You can get booked on top podcast with kitcaster.com.
• 35:05
  So that here's here's a good one.
• 35:07
  What's a revenue stream you killed even though it made
• 35:10
  money? I had some courses on privacy
• 35:15
  certification, they were making money but I killed it because I
• 35:19
  didn't want to update the courses again and it was a lot
• 35:22
  of overhead to maintain them. Yeah.
• 35:26
  You know, a lot of folks try and make courses and it seems like
• 35:29
  you found a lot of success with the courses.
• 35:33
  Talk a little bit about that, how you kind of choose which
• 35:35
  direction to go in and how do you kind of like, you know, put
• 35:39
  a stamp of approval where people can really use that as kind of a
• 35:42
  leverage in their career? So I put it in terms of what do
• 35:47
  people want? And when I say people, there are
• 35:49
  two categories of people. One, you as an individual and
• 35:52
  second, you working for a company.
• 35:54
  So either you want to certify yourself and grow your career
• 35:57
  and in that case the certification courses come in or
• 36:00
  else you are in the organization and you want to entrain a team
• 36:04
  of people, get them certified or you want the entire staff to be
• 36:07
  trained. So those are the three kinds of
• 36:09
  training we provide. If you're an individual, we help
• 36:11
  you with standard ISO based, EU regulation based certification
• 36:14
  training. If you want to get your teams
• 36:16
  trained, you can do the standard training or we can do a custom
• 36:21
  training. And if for all staff that's a
• 36:22
  short capsule of 30 minutes or 60 minutes.
• 36:25
  And most importantly, when you say I'm bored or executive team
• 36:30
  don't listen or don't understand, we help them in
• 36:33
  their language to understand what the law is without talking
• 36:35
  the law. Yeah, you know, a big part of
• 36:38
  your, your, your work is training.
• 36:40
  And one thing I know about people is change management is
• 36:44
  hard, man. People don't want to do things
• 36:46
  differently. So kind of how do you tackle
• 36:49
  that? Do you kind of begin with a
• 36:51
  certification, say like, OK, here's a five day certification.
• 36:54
  Once you complete that, everyone's going to have kind of
• 36:56
  a framework for understanding, and then you start to implement
• 36:59
  kind of one-on-one or small group kind of teachings.
• 37:04
  Very similar to what you said. So what we do is we usually do
• 37:07
  what we call a discovery workshop.
• 37:10
  So we get all the key decision makers around the table and say
• 37:14
  we are going to have a discovery around what is responsible AI or
• 37:18
  privacy protection combined, call it digital trust, which has
• 37:23
  one hour of training so that we align on the vocabulary and key
• 37:26
  terms. And then we ask them set of
• 37:28
  questions. Are you doing this?
• 37:29
  Are you doing this? Are you doing that?
• 37:31
  In a, in a span of two to four hours, we have a very clear idea
• 37:35
  on what they're doing right and where the gaps are.
• 37:38
  After that we give them a report in one week saying these are
• 37:40
  your gaps. And that includes saying get a
• 37:44
  couple of people on five days certification course, get all
• 37:47
  your staff trained on it. Thank you.
• 37:50
  You have been already trained but we can do a second round of
• 37:53
  it explaining you now the policy we will create for you so that
• 37:56
  you can talk to your staff in the right way.
• 37:58
  That's the approach we do to discovery workshop which has
• 38:02
  inbuilt the training for key stuff and then follow it up with
• 38:06
  the actions which include training for key personal, all
• 38:10
  staff and management and combined with the creation of
• 38:13
  policy and identification of gaps.
• 38:15
  And we are very flexible. If you want to close the gaps
• 38:18
  yourself and you have the skill then you want to obtain people.
• 38:21
  Good luck. Do it yourself no problem.
• 38:23
  But if you want us to help you hand hold you and get to a jump
• 38:27
  start, then we are also happy to do it and they're both types of
• 38:31
  clients we have. Love it.
• 38:33
  So discovery call certification implementation that that's a
• 38:40
  great framework. Where, where do people, where's
• 38:43
  the challenges for folks, you know, because I know it's hard.
• 38:46
  I know it's hard getting them on board where where they kind of
• 38:49
  fall through the cracks. You mentioned the change
• 38:52
  management. I think the first challenge is
• 38:54
  not knowing what to do or well, if you should do it.
• 38:58
  That's where the Discovery workshop helps.
• 39:01
  But the second challenge is being convinced about it, that
• 39:04
  you want to do it. And many of our clients, they
• 39:06
  come in for the Discovery workshop and they're like, we
• 39:08
  will do it next year because it's not so urgent.
• 39:12
  And then sometimes they come back after one or two years when
• 39:15
  it becomes urgent and it's already late, but it's usually
• 39:19
  now last minute change. So that's where it is.
• 39:21
  But the most important challenge, as you, I would say,
• 39:24
  is understanding that's important and having a sense of
• 39:28
  urgency around it. Yeah, I love that.
• 39:32
  If you were entering college as a freshman this year, what would
• 39:36
  you study? What would I study?
• 39:40
  I would study prompt engineering because I know people talk about
• 39:44
  AI, but I would like to learn how to manage AI, how to utilize
• 39:49
  AI for my, my own benefit. Because in that context, we have
• 39:54
  new recent going to launch very soon, a prompt engineering
• 39:57
  course, which will be free initially for people.
• 40:01
  Because I want to educate people that it's how you use AI, not AI
• 40:06
  is useless. Yeah.
• 40:08
  What, how do you, what's your approach to kind of prompting
• 40:12
  now? Are you very specific?
• 40:13
  Do you kind of like do research on the prompt and then deploy
• 40:16
  it? Are you speaking, I mean
• 40:18
  everyone says speak to it casually.
• 40:20
  You know what? What's kind of your take?
• 40:23
  I have what we call a reactor framework.
• 40:26
  That's what I use. So reactor for me means you
• 40:30
  first tell it what's the role. If I'm using say, ChatGPT or
• 40:35
  Copilot or Perplexity, doesn't matter what it is, I would say,
• 40:38
  hey, you are my McKinsey consultant and you're going to
• 40:43
  give me this sorry role. Then I'm going to give an
• 40:47
  example of what does good look like.
• 40:49
  Then I'm going to tell what am I aiming to achieve?
• 40:52
  Why? Why am I doing this?
• 40:55
  Then I'll give it some context. I'm in this situation.
• 40:58
  This is the company, this is my need, this is my desire.
• 41:01
  And then I'll get use this material and finally I'll ask
• 41:05
  outputs of role example in context text and output and we
• 41:09
  call it the reactor framework. And that's what we baseline our
• 41:12
  prompt engineering on. And then once you have given it
• 41:15
  all these elements, they're bound to get very high quality
• 41:18
  input. Yeah.
• 41:21
  Output from AI input for you. Yeah, it's interesting because I
• 41:25
  have built up so much context in ChatGPT for personal stuff and
• 41:29
  clawed for professional stuff that often times I'll be like,
• 41:32
  hey, this is the project I want to do.
• 41:34
  Give me a a prompt that sounds like I would write it and
• 41:37
  they'll give it to me or, you know, and then like fine tune
• 41:41
  from there. And then another thing I like to
• 41:43
  experiment because they keep saying casual language, casual
• 41:46
  language. I'm like, all right, so I'll
• 41:48
  open up a project and say do the thing I want you to do the way
• 41:51
  that I like it. Just see what.
• 41:53
  Happens. No, that's I think what I tell
• 41:58
  people is in past 10 years ago, if you would hire an intern or a
• 42:02
  fresher or a person in your team, you would give them a week
• 42:06
  or 10 days or two weeks to train them, to guide them.
• 42:09
  What is the way of working in this company?
• 42:12
  Do that with AI. Why do you expect instant
• 42:14
  results? Train your AI, guide your AI,
• 42:17
  give it information. And even with your intern or
• 42:21
  senior manager you hired, you used to explain them, hey, I
• 42:25
  want you to do this. This is the problem, this is the
• 42:27
  challenge. Then don't ask, hey, what should
• 42:30
  I do for my audit? Which audit where, who is doing
• 42:34
  it, which context and how urgent it is?
• 42:37
  Do you want me a system for an audit which is one year later or
• 42:41
  do you want an audit next week? So.
• 42:44
  So give it full context and it'll help you.
• 42:47
  Yeah, I love that. Aside from compliance, what AI
• 42:50
  experiments are you you playing around with or a side projects
• 42:54
  are you doing with AI right now? It's the site project because as
• 42:58
  a founder and as an entrepreneur myself, I have the same
• 43:01
  challenges that anyone else has. So I'm looking at utilizing
• 43:05
  different elements to, you know, ease the workload.
• 43:09
  For example, marketing, what kind of social media strategy we
• 43:14
  need to have, How do we optimize our SEO?
• 43:16
  How do we get to our customers? What do they like to read?
• 43:20
  How do we create landing pages that convert?
• 43:23
  How do we write emails that convert?
• 43:25
  So I use AI locked for validation on these things and I
• 43:29
  also use it for ideation and things.
• 43:33
  I'm still not there with AI agent enough.
• 43:36
  We do have a couple of agents which check some band visibility
• 43:41
  and things like that, but those are the thing.
• 43:43
  And then, like you in our podcast, what I use AI for is
• 43:48
  finding good titles, transcription and writing
• 43:51
  descriptions. And of course we have the human
• 43:54
  element which checks and corrects it, but we use it for
• 43:56
  the first draft. Yeah, I love that.
• 43:59
  What are some of your favorite marketing tools that you're
• 44:01
  using for AI, whether it's landing pages or or generative
• 44:04
  artwork, things like that. So in terms of LMMS, we are
• 44:08
  using Perplexity Copilot Gemini and there's something called
• 44:13
  Centra dot AII, don't know if you heard that.
• 44:16
  So that's a set of AI agents, your virtual assistant, your SEO
• 44:20
  and so on and that those are the ones we are experimenting with.
• 44:24
  And then in terms of pure marketing, I mean the typical
• 44:29
  CRM solution along with the, the for me the Copilot works the
• 44:34
  best because we are user, we are a Microsoft shop, Microsoft
• 44:37
  environment and Copilot integrates well and that's what
• 44:41
  is serving us the most. Love it, love it.
• 44:44
  You've got a lot of successes and you got your hands in a lot
• 44:46
  of pots. I love it.
• 44:48
  Any spectacular failures you'd like to share?
• 44:52
  Many. Well.
• 44:58
  There are always failures in life.
• 45:00
  The question is not failures. The question is what we learn
• 45:03
  from those failures and how do we grow.
• 45:06
  So even my podcast arrived out of a failure.
• 45:11
  So I was into this business of setting up my company and
• 45:16
  somebody said you should record some interviews, conversation
• 45:20
  with people. So I recorded four or five and
• 45:24
  then I had no plan, no idea I was recording.
• 45:28
  And then one fine day somebody introduced me to a very high
• 45:31
  profile person and he said when is it going live?
• 45:36
  I said next week because I didn't have a good answer.
• 45:42
  And then I went to my coach and I said, what do I do?
• 45:44
  He said, just put it on a share, what do you say membership site
• 45:47
  and make it visible and I'll say, hey, I'm not going to do
• 45:50
  that. I'm not going to.
• 45:52
  I felt like a big failure because I have 7 or 10
• 45:56
  interviews done and I've duped them.
• 45:58
  But to the last one, I committed next week it'll be I sat
• 46:02
  together, edited the podcast, launched it on YouTube and I
• 46:05
  found Anchor dot FM in those days.
• 46:07
  Now Spotify published it and then man, I released it and I
• 46:12
  learnt the week after, oh, there can be a trailer.
• 46:15
  So I said no problem, let's cut it and put the trailer next
• 46:18
  week. So for one year the trailer was
• 46:20
  coming 11 week after the episode.
• 46:23
  And then we said hey, no, no, no, no, learn along the way.
• 46:29
  I'm sure you would be looking at the more dramatic failure, but
• 46:33
  that's the kind of failures I go through.
• 46:35
  So I quickly turn it into how to make it learn because I'm an
• 46:39
  entrepreneurs, we have to make sure the things work.
• 46:42
  Like even day before somebody said I gave feedback on the
• 46:45
  ebook and the person said it's not done one week later and I
• 46:51
  say why? Well, I created in a different
• 46:54
  tool. I can't create clickable table
• 46:56
  of contents. So no problems.
• 46:58
  Put it on the put it to the audience for version 2.
• 47:00
  You can do it because the ebook is no good in your SharePoint or
• 47:05
  in your folder. The ebook is good on a landing
• 47:07
  page wherein people can consume it.
• 47:10
  And just because I gave feedback, you don't have to
• 47:12
  block. So I was like, hey, I'm a bloody
• 47:15
  failure. I can't even explain them that
• 47:17
  it's better off with people. So every day you feel like some
• 47:20
  little or some little other thing.
• 47:22
  How did I end up with this? And same way we did Google Ads
• 47:25
  and we failed drastically. I mean, we put 10s of thousands
• 47:30
  of EUR and I think converted. And I was trusting the marketing
• 47:33
  team. And end of the day, at certain
• 47:36
  point I said, Nick, stop, let's do an audit, let's look at it,
• 47:40
  what's not working, why it's not working.
• 47:42
  So yeah, every day it happens. How many you want?
• 47:46
  I love that. I love that because you know,
• 47:48
  ultimately there's going to be people that are listening that
• 47:51
  are still sitting on the bench and they're not getting in the
• 47:54
  game 'cause they're trying to craft something perfect.
• 47:56
  They got their head in the sand and they want something perfect.
• 47:59
  But when you kind of release yourself to say, I'm going to
• 48:01
  build and cut in public in a capacity that I'm comfortable
• 48:05
  with and I'm just going to concede that it's going to be a
• 48:07
  little bit clumsy at times and I'm going to make mistakes.
• 48:10
  And that becomes a part of it, you know, and I, I think it, it
• 48:13
  really releases a lot of stress and pressure if it's like, hey
• 48:16
  man, it doesn't have to be perfect.
• 48:18
  You know, like 90% is good enough, especially if you're
• 48:21
  talking about some of these founder LED initiatives, you
• 48:23
  know, being a big advocate for podcasting.
• 48:25
  Obviously I love that story because it's like, man,
• 48:28
  sometimes you just got to get it out there, you know, and then
• 48:31
  after 100 episodes, take an audit, be like, OK, what have I
• 48:34
  done? Well, what have I done?
• 48:36
  What, where can I improve? And then you get better over
• 48:38
  time. I think that's brilliant.
• 48:39
  You nailed it. Done is better than perfect.
• 48:42
  Yeah, yeah, I needed to hear that myself.
• 48:48
  I've got my precious little projects I don't want to release
• 48:51
  yet, you know, get it to the world.
• 48:54
  Let it breathe, opponent. This has been absolutely
• 48:58
  amazing. Give us a shout out for your
• 49:00
  podcast so I can put it in the show notes.
• 49:02
  It's called fit for privacy. Is it on Spotify and Apple and
• 49:05
  everywhere else? Yeah.
• 49:06
  So the Fit for Privacy podcast is available on YouTube and all
• 49:10
  podcast platforms. You name it and it's there.
• 49:13
  You just say the Fit for Privacy podcast and you will find it and
• 49:17
  it has an episode every week. I mean episode every two weeks
• 49:20
  and a tailor and an episode alternate weeks as I said, but
• 49:23
  now in the right order first and then the episode.
• 49:27
  Not like the way I started. Love it, love it.
• 49:30
  And for companies that are looking to build their trust in
• 49:32
  the digital age, I'm really glad that you pointed that out too,
• 49:35
  that like, sure, there's there's legal ramifications, there's
• 49:38
  security threats, but ultimately, boy, one of these
• 49:40
  these things pop. It could be it could destroy
• 49:43
  your reputation forever because people are going to LO Ms.
• 49:46
  they're asking about your company.
• 49:47
  If that's the first news story that comes up, good luck finding
• 49:50
  customers. So you you need to go check out
• 49:53
  fit for privacy.com and of course, if you're looking for
• 49:57
  some of these courses that you can go train yourself, train
• 50:00
  your staff, get certified, it's grow skills dot store and Punit.
• 50:05
  Where's the best place for people to connect with you
• 50:07
  personally online? Best is LinkedIn.
• 50:10
  Find me Punit Bhatia and connect with me.
• 50:13
  And I'm here to help you, here to serve you, here to guide you.
• 50:17
  Tremendous. And wherever you're listening to
• 50:18
  this, you will see that link to LinkedIn in the show notes.
• 50:23
  Punit, thank you so much for joining me.
• 50:25
  Thank you so much, Ryan, for having me.
• 50:27
  It's an absolute pleasure to have this conversation with you.
• 50:30
  Appreciate it. While you were recording your
• 50:32
  last episode, Casey Taylor was out there pitching your podcast
• 50:36
  to brands and securing sponsorship deals.
• 50:39
  Well, she's not yet, but she could be.
• 50:41
  Casey is Howdy's AI partnership agent and she does something
• 50:44
  pretty wild. She actually secures
• 50:46
  sponsorships for your show. Not a match list that you have
• 50:49
  to work through yourself, not a template you have to send.
• 50:52
  She handles the outreach to follow up the whole thing and
• 50:55
  brings back partners. You started your podcast because
• 50:58
  you had something worth saying. Casey makes sure that you get
• 51:01
  paid for sayingitgomeetcasey@gethowdy.com.
• 51:05
  That's GETHOW di.com. 2 minutes to get started, then Casey takes
• 51:10
  it from there. That's howdy.
• 51:11
  partnershipsgethowdy.com. In three years, AI fluency will
• 51:17
  be a baseline expectation for every college grad, not a
• 51:20
  differentiator. It'll be a requirement.
• 51:23
  The students getting ahead are the ones building that skill
• 51:25
  right now in high school and college before it becomes table
• 51:28
  stakes. ainativestudent.com is the program doing that weekly
• 51:33
  one-on-one coaching, live cohort sessions, on demand courses and
• 51:38
  the fastest growing AI chat for students all tailored to your
• 51:41
  students specific career path. Pre Med they learn how AI is
• 51:45
  transforming diagnostics. Pre law they learn legal AI
• 51:49
  tools business they learn to think like an AI power operator.
• 51:52
  1 flat membership less than a private tutor.
• 51:55
  Cancel anytime. And if you're a founder
• 51:57
  listening to the show, you already know the cost of falling
• 51:59
  behind. Don't let your kid pay that
• 52:02
  price. Go to ainativestudent.com, apply
• 52:05
  in a couple minutes, and get your student into the cohort
• 52:07
  that changes their trajectory. AI native student.com and send
• 52:11
  them to the chat. It's totally free.
• 52:15
  If Instagram changed their algorithm tomorrow, how many of
• 52:18
  your followers on Instagram could you actually reach?
• 52:20
  Now ask yourself how many people are on your e-mail list.
• 52:23
  The gap right there is the gap between a personal brand you're
• 52:26
  borrowing and one that you actually own.
• 52:29
  The best founders in the world, The ones getting the deals, the
• 52:32
  stages and the inbound. They have newsletters.
• 52:34
  They show up in inboxes every week.
• 52:36
  They build real relationships at scale.
• 52:39
  Inbox Alchemy makes that happen for you.
• 52:42
  They write your weekly founder newsletter, build you a stunning
• 52:45
  newsletter website, and grow your list to 25,000 engaged
• 52:48
  subscribers in six months. All managed end to end.
• 52:52
  So it's generating pipeline while you focus on building your
• 52:55
  company. Algorithms are someone else's
• 52:57
  decision. Your e-mail list is yours
• 52:59
  forever. Head to inboxalchemyco and start
• 53:03
  building the audience. No one can take away from you.