AI-Powered Startup Workflows

Guy: [: 00:00:00

Today I'm interviewing, uh, David Hirschfield and he's with a company called Techies. And, uh, we've been talking a little bit and I'm incredibly impressed with some of the stuff that he's been doing. But before we get started, let me tell you a little bit about him. He is a 35 year software veteran. And UCLA physics alum.

artups, including one sold in: 2000

Welcome.

David: Uh, thank you very much. And actually now it's over. It's actually over 90 startups. I think I have a typo in somewhere in my bio. So it's over 90 startups during that period.

Guy: Well, that's a, that is just absolutely incredible. Well, what a, uh, what a backstory and, uh, so tell us a little bit more about what your backstory is and some of the things you've done in AI and marketing.

David: My backstory was in software development, uh, not necessarily in AI and marketing because I started a, my first software company in 92 was logistics, route distribution, inventory management, and that company I grew to 800 customers in 22 countries over the next eight years and sold out to publicly traded firm.

So I was successful then [: 00:02:00

It's been 18 years and we've worked with a lot of startups during that time. And the biggest problem that startups have is understanding their markets when they start. And that has led me into AI and developing a methodology for founders so that they can find product market fit. Quickly and start to generate revenue from their idea before even building the product as a way of proving product market fit.

an AI marketing perspective [: 00:03:00

But it turns out that it's not as easy to confirm as a lot of people think it is when they go off and start software companies. So using AI to really nail down this early adopter and product is just a critical success factor for founders now.

Guy: Yeah. And absolutely. Uh, how true. How true. Yeah. Thank you for that.

So, you know, one of the things that's, uh, really important is, uh, you know, how do you can automate, uh, the workflow within marketing or, or generally, but let's talk about marketing and certainly AI can simplify that whole process. So tell us a little bit more about how that works and how it can, you know, potentially even make, you know, these workflows very accessible for, uh, non-technical users.

s, uh, pick as an example of [: 00:04:00

And is that person really in a position to want to buy the thing that I'm trying to sell? Whether it's a product I. Or a service. And what signals are you able to find that indicate that they really are a potential person who would be interested or maybe a channel partner that has a lot of people that might be interested, that they could refer to?

ing all this information and [: 00:05:00

Information that they had, these signals that they pushed out there on the on the web and because what their position is and what their authority to buy is and things that we can assume based on their public personas. Then the next step to that is then crafting some kind of email drip campaign that speaks to them personally about.

continuing the conversation [: 00:06:00

So. Where the re end result of all this is, you end up possibly sending many, uh, fewer outreaches to much more highly targeted, much more highly valued potential target customers. And this is where I see AI going and people will appreciate getting those emails because it speaks to the things that they have said are important to them right now.

Guy: Yeah, absolutely. And, uh, you know, certainly getting the outreach right, and, um, and I agree with you, you know, a good drip campaign and just getting the, you know, that personalized touch really makes a big difference. Now, one thing you talk about as well is, uh, you know, getting and, and discovering new insights and using AI for that.

n even building them out and [: 00:07:00

David: Well, I, one of the most important things, I think for a marketer, uh, right now who might see AI as a competitor.

Um, you should see AI as like your buddy, your best friend, the person who you can tell all your darkest secrets to and get insights about all you know. That's really kind of the relationship you want to have with AI right now because ai. Is, is something that you can leverage to make what you do so much better.

Um, and so I'll give you some examples and I'm going to take this out of the context of marketing because I think it provides a really better understanding of what the value of ai. I'll give you actually two examples. 'cause these, and these are real world examples. They're both personal. So the first thing was.

about my wife wants to grow [: 00:08:00

And, and we start talking about it and then her sister calls as she gets on the phone with her sister. So I brought Chet GPT up on my phone and I said what my we're. Two people were in our early sixties and and gave it some information about where we live in Vista, California and considering the climate and our age, and we're not veg vegetarian, but we eat a lot of vegetables and my wife wants to grow them all, how many beds would we need and what questions should I even be asking anyway, so then it gave me a whole bunch of questions and I thought they were all perfect questions.

And so I had it answer the questions for me, and I liked the answers it gave me. Anyway, we had this conversation going, and by the end of it, I had, we had, I, I don't remember how many beds. It was like eight beds, or seven or eight beds that were four by four each, and it had exactly what should be planted.

lanted in those beds is the, [: 00:09:00

Right. And then you also need companion flowers because. Certain bugs, like certain vegetables, but there are flowers they actually like more. And you plant those flowers alongside the vegetables and now it draws all those bugs off. Right? And so it gave me a schedule for every season of the year, what should be in each bed.

And, and then you have succession planting when you take stuff out of the soil, what to put in that same soil, because if you grow an onion, what likes to grow after an onion? Right? And it gave me the whole plan. In five minutes this conversation we had. And then my wife got off the phone with her sister and I said, what, how about this?

better now than it was then. [: 00:10:00

The other example I. Uh, we just went to the Gem and Mineral Show in Tucson, which is the largest show like that in the world. And we had a goal because when you walk into our house in our entryway, we wanted some really beautiful p um, stone or mineral, either I. You know, something that we could put up on the wall right in the entryway or something that would stand up on the floor, like one of those big giant crystal amethyst, agates and I, and I understood.

You can find that spectacular stuff, really inexpensive. And this show is unbelievably large. It's the whole city. To give you an example, they have like 80 locations and the convention center, which is completely taken over by this show, is just one of those 80 locations. And not the biggest, we didn't know how to do this.

wasted your time. So I used [: 00:11:00

It did not only. Did it give us a schedule of what, where to go and in the right order so that we would be finding the kinds of things that we wanted to buy over a period of, over the three days that we were there. But also when to start to negotiate to get the best deals, how to negotiate with certain vendors, which vendors had the types of things that we wanted, where to eat in between, in the, you know, in the morning and then at lunchtime, and then where to go that was nearby.

ize what's possible when you [: 00:12:00

Guy: Yeah, yeah, absolutely. And um, uh, you know, and those applications like that are, uh, you know, so just so incredible. You know, when you're talking about. You know, that basically that workflow that you had in searching for that one particular piece to decorate your home, uh, you know what? That could easily be the workflow that you need to develop a new product or whatever it happens to be and launch it.

And so, uh, great example. I really appreciate that. Um, now AI can also be used, uh, on research and, uh, it can certainly accelerate that research. Uh. You know, from not only product research, but also, you know, general market research. And, uh, so tell us, you, you've developed a methodology there. Tell us a little bit about that.

e of your startup and and my [: 00:13:00

Now, if you're Maybelline and you're coming out with a new type of mascara, that's not the right fit for this because you already know who you're. Ideal customer profile is, and you know how to reach them and the language they care, right? So your marketing department's all dialed in there, but if, um, uh, and maybe you're just going after different age, demographic or whatever, but you, because of all the signals you've had.

So that's why you're coming out with this product. If you're an e-commerce company. This wouldn't be for that, but if you're selling a service or a software product or, or a technology pro or you know, a hard product and it's new in the market, it's, you don't know who the ideal customer profile is. And that's the problem of startups deal with and new product launches that are not in that ideal customer profile target to begin with.

Who is that ideal customer [: 00:14:00

And a root level problem statement is. Not the generic problem statements that you normally hear, but it's the one that starts with, I feel threatened because if I don't solve this problem, my future in some way is going to be much worse. Right? Just fill in the blanks along in there, and that's a root level problem statement.

r might be your. Restaurant, [: 00:15:00

You didn't know it until they leave a one star review or whatever the problem is, right? And both of those could have the same high level problem statement, which is, I don't have the information I need when I need it to make critical business decisions. So that's the difference between a high level. On a root level.

So we list all the root level problem statements, all the niches, and out of that then we score all the intersections between all the problem statements and the niches, and we score them on two levels. One is how much does that problem statement cost that stakeholder in that niche and a separate sheet, how much is the perceived impact that that problem statement has for that niche, uh, with that niche?

high perceived impact to the [: 00:16:00

Then you do deep dives on those to figure out which one. But this is a very long and difficult, tedious process. Requires collecting a lot of information from a lot of places. Easy to do this incorrectly, and so AI is just invaluable at helping us achieve this. In fact, we're building an AI model that will automate all, all that entire workflow to do all this scoring.

ell directly to those people [: 00:17:00

'cause that's the whole goal here.

,: 1000

And this, you know, can really, I go by that too.

,: 1000

Guy: Because you know, the, the problem too, when you're, you know, the founder and the, you know, an engineer or a software developer or whatever it is, you know, the first customer is probably a friend that you really know really well, and he's willing to kind of give you the benefit of the doubt, take on the risk.

th [: 00:18:00

And it sounds like what you're talking about with, uh, with Launch first and then supporting that with ai. First of all, speed that process and get, you know, those answers that you need so you can really accelerate, you know, and, and with less risk and more certainty, how to get that 10, you know, a hundred and then the thousand, uh, you know, customers as you're trying to grow

David: and less cost.

And if you can't get there, you fail for a fraction of what you would've failed for. If you go the what, the typical route is MVP pitch deck, try to raise money.

ow, for what you're building.[: 00:19:00

David: Okay. So for marketing, for project, uh, for developing a product, we use AI for all of that. So one of the things we now stepping outta marketing again, uh, just to talk about it from a planning perspective. One of the things that we think that we're particularly good at is how we do our estimates. Uh, project estimates for a software project.

Uh, they're very detailed and they're accurate, and we deliver on time based on the, uh, estimates that we give to clients, assuming they don't change scope in the middle of the project. And most software companies do not do that. They're not on time. They're usually off by, you know, factors. You know, X factors not just like percentages.

the entire year and all the [: 00:20:00

So that is something that we are really proud of to do that. It's a very time consuming, costly process to do these estimates. We are now, uh, just finished the first release of a tool that we built that's AI driven that will actually make us even better than we are now in building our estimates based on the same kind of methodology that we use.

So it mimics our estimation process, but uses AI to actually extract all that information. And to organize it into our plan. And our estimates are not just how long, how much it's gonna cost, how long it's gonna take, but also the team size when individuals on the team need to be scheduled into the project to deliver their part of the project.

ing about marketing, then we [: 00:21:00

What kind of campaigns, uh, what's the messaging? Creating all the drip campaigns on and on. And we use it for pretty much everything. And we just keep building our own internal workflows that pull AI in at different steps in the workflows to accomplish specific tasks that can automate all of our marketing.

Guy: Yeah. Yeah, absolutely. That makes a, a lot of sense. And, um, you know, that planning process and getting it right, especially in a software development, uh, arena where, you know, every, every day that you're off, you know, costs you money or costs your client money, it's never good and Right. And actually it's not only just the cost issue, it's also the timing.

'cause they, you know, that the time to actually, that you might lose in lost sales is, uh, very, very high.

and for us it's a matter of [: 00:22:00

Guy: Yeah, absolutely. So we've been talking about, uh, strategy and planning. Um, now let's talk about execution. One of the things that AI really can do is, uh, the personalization at scale. And, uh, so that when you do your outreach, I guess, whether it's email or maybe any other kind of personal outreach, whether it's maybe a text, uh, talk about that, that personalization component.

it. Picks one person on the [: 00:23:00

And develops like a social persona from a buying perspective, depending on the type of target it is. And by doing that, it can create a really personalized drip campaign that speaks directly to the issues that are current with that person. Uh, so that then when they see the campaign, it feels like somebody read.

Read about them as responding to something that they may have posted somewhere and is engaging them in a way that is organic and meaningful. I know when I get those emails, I don't care how the email was generated. I. I appreciate those emails. Engage me and I open them, and then I'll click a link or reply to the email and start to engage those people on the things that matter.

pen to care about right now. [: 00:24:00

So this is where AI is just so powerful, right? So if it sends five to 10 emails a day, which of course it can do a lot more than that, those are five to 10 incredibly well targeted emails that should most of them get engagement. As opposed to sending out hundreds every day. Mm. And having 1% of them give you engagement.

Guy: Yeah. Yeah. And that, and that math too is, uh, is easy to track and easy, easy to determine what is really gonna, you know, help you at, you know, at each step of the funnel. And, uh, and also hit the, the customer when they're, you know, with the right message at the right time. And, uh, mm-hmm. And I think, you know, that's what a lot of people are doing today is really using ai, uh, in marketing.

e it so much more effective. [: 00:25:00

Yeah. So tell us a little bit about what you found there.

David: Okay, well, uh, okay. There's different levels there. Number one, from the plan, going back to planning, what kind of content should I be creating that is going to one, make me an authority on something that my potential clients are gonna care about right now?

words in the right place and [: 00:26:00

So AI is just amazing at all of these things. If you are using some of the tools that are available to help with this, and also if you're driving it pro correctly, 'cause prompting is so critically important, writing really meaningful. Elaborate prompts. People think they can put in one, you know, short one sentence prompts without creating a lot of context and providing examples and things like that, and they just don't get out of AI what they're hoping to get, because AI is not prescient.

You know, it doesn't have any idea what you truly need to get out of it. So if you don't give it enough context. In the prompt that you write, then it, how could it possibly know what it needs to accomplish in the thing that it's creating? But it is getting really good at how it produces content.

Guy: Yeah, yeah.

years [: 00:27:00

Well now it's the same kind of thing, except now it's building out the prompt to structure that prompt so you can really get the answers that you're really looking for that make then the most difference towards whatever it is that you're trying to achieve. And, and that's, uh, that is absolutely the case.

David: So here's a simple workflow for most marketers that they could follow, like regarding the content. The first thing is ask AI to do and use the deep research tool on what type of content should I plan to produce content so that I could, whatever you're trying to accomplish. Build authority in my particular vertical profession.

ut a plan together for this? [: 00:28:00

To help you even think through, what should I be thinking about? What should I be asking? How should I approach this before I actually do anything? And it'll give you all that context and then you can ask it. Okay, here, those questions are good. Look, go ahead please give me an answer to each one of those to start with.

And then it'll do that and you then you can say, expand on on your first answer. 'cause I. Put a plan together to execute on that. So let's say I want to become an authority as a software developer. So what are the topic and who are my verticals? I, I want it to be in healthcare vertical. I want it to be for small to medium product related healthcare companies that are selling into healthcare.

developers? And if it says, [: 00:29:00

It'll usually, it'll give you that information in advance. Then say, okay, now I want to build content strategy for becoming an authority for building software that would engage the buyers of software development services in that market. What questions should I be asking and then do it again? I. It'll give you the questions and then, and just keep going through this process until at the end you've got an entire execution Plan, a list of topics in the right order, um, the length of the articles where you should be pushing art, different articles of different length or videos or podcast interviews or whatever it is.

I'm going to generate a lot [: 00:30:00

Guy: Absolutely. That makes so much sense. Um, man, I wish I could, uh, we could keep on going. You've got some great points here and certainly a lot of help for, uh, you know, the audience. Uh, but, um, let me ask you one last question and that is, um, so what advice would you give to an up and coming new marketer as.

They are now almost competing against AI for that job and for that, uh, that value that they're, they're trying to provide as they get out into the, uh, into the working world.

David: Well, if you're competing against ai, you're gonna lose so, and not because I. Well for, for obviously reasons, right? AI costs almost nothing and it, the results that it gives are pretty spectacular in a lot of cases.

I to create better marketing [: 00:31:00

Or whatever, 'cause they, that's what they want to see. They wanna see people that understand how to make AI really fly in their organizations. And if you can show them how you can do that, you're going to be at the top of the list in terms of being hired or being contracted or whatever role you're playing in marketing to the company that you're talking to.

Guy: Yeah, absolutely. I love that. And, um, and I think you're right, you know, and you said it early on too, which is, uh, you know, AI has to be your best friend. And, and this, you know, that, that, that answer is, uh, clearly the, the way that you have to do that. So, um, uh,

n you're talking to ai, I, I [: 00:32:00

Um, and I wanted to confirm that I got the right answer. So I'd say thank you. That's perfect. Um, because that would help inform it that, um, how to answer the questions in a way that I was right. And there have been studies done since then that if you're really a nice ai you get better responses from it.

Nobody really knows why. Plus what happens if AI takes over the world? I wanted to remember that. I was really nice to it.

Guy: Well, it is gonna take over the world. That is the challenge. So I like what you're talking about. Yeah. Yeah. Well I guess the please and the thank you, or at least the thank you anyway, is, uh, a confirmation that the answer is better than it might otherwise have been.

So therefore you're kind of giving it some feedback that may then be used to, uh, you know, be incorporated moving forward.

here's, like I said, there's [: 00:33:00

Guy: Yeah. Yeah. Fantastic. Well, I hate to say it, I'm gonna start doing that. So I thank you for the, uh, thank you for the advice. That's pretty good. So, uh, where can, uh, folks read, uh, reach you and, uh, and, and learn more information about you and your company? I.

David: Yeah, go to techies.com and techies is spelled T-E-K-Y-Z.

Or you can email me. The email won't be in the link, but if you made it to the end of the show, I feel like I can share my email with you. It's david@techies.com. And I'm easy to find on LinkedIn, David Hirschfeld. Um, there'll be a, there should probably be a link in the show information below, but, um, I'd love to hear from you.

e. Yeah, I can give you some [: 00:34:00

Guy: Fantastic.

Yeah, thank you for that. And, uh, so it's, uh, techies.com, TEKY z.com. Uh, David, thank you so much. This was really fantastic. And, uh, for our audience, uh, definitely stay tuned for many other se uh, videos in this. Series on the backstory of marketing and ai. And if you'd like to please download, uh, other excerpts, uh, for my book and other information, uh, at marketing machine.pro relevant.com.

David, thank you so much. This was fantastic.

David: Thank you guy. I really appreciate the opportunity to be on your show. Yes.