Introduction

Hi, everyone. My name is Tata Maitisyan. I am an AI enthusiast and a marketer.

Career Journey and Evolution

I would say that my entire career has been shaped quite differently. A lot of people ask me, why do you keep changing the types of things which you do? It's because I follow my passions.

I started a corporate career, spent 11 years in the corporate world working for organizations like Deloitte and Nike. And then I've decided to move to the startup world and I joined a startup which later became a unicorn and I was heading there international.

Early Interaction with AI

expansion team that was the time in 2019 when i was first touching the ai because all of a sudden i was managing artificial intelligence lab back then generative ai was not so big so we primarily worked with machine learning computer vision etc and fast forward now

Nike marketer turned to a startup executive and then quit all of that to help AI startups grow their user base and business. So that's what I do nowadays.

Using AI in Day-to-Day Operations

And today I'm here to show you something which I hope a lot of you do. But before I jump into some of the demos and explanations, I'd like to ask you a question.

How many of you do market research, competitor research, lead generation in your job? Can you raise your hand?

Okay, about a half. Do you do that manually? Raise your hand if you do that manually. No one. Okay.

Do you use AI for that? Okay. Most of you do. Awesome. Great.

I hope today I'm going to show you some of the examples of the types of ways I use that day-to-day and how we use that with our clients. Hopefully, that is going to bring something new to you.

If it doesn't, please share the comments. I'd like to learn from you as well.

This is a very nice small space so we could actually exchange opinions.

Challenges in Selling and Research

What is the topic of today is something which you do a lot as an entrepreneur, as a founder, you try to figure out how to get insights, understand what companies do, you try to find the competitors, you try to analyze the market. And then ultimately, what Reggie was saying, you try to get to people and actually make them answer.

And what I have figured in the last years working with the startups is that one of the hardest things is actually selling.

Finding people, and it doesn't matter if you're an investor and you're trying to sell to a startup, hey, here is our money, please take it if they're good. Or if you are an entrepreneur trying to find a customer, or if you are a consultant like I am myself now, trying to find startups to work with me.

At the end of the day, the world turns around this thing of finding the right people.

And what triggered this whole research thing was when I understood how much time we actually spend on research. It was really manual.

Why we need this? So I figured that there are few key use cases.

Key Use Cases for Research

Most times we do that when we want to understand the product positioning. Let's say you develop a startup, a solution which is what you think works for a specific segment, but you need to understand what words to use, how to actually sell it, which segment specifically to target.

Another reason could be for product development. You need to understand which kind of tools or features to use, of course for sales, of course for lead generation.

And there are multiple ways you could do that.

Automating Research

There is a way which is pretty straightforward, and I'll show you how to automate this because it gets really hard. Question is how to automate this. So I'll today show you a couple of tools and ways you could automate this whole process of finding the specific topics online.

A lot of things are already on LinkedIn, on Twitter or X, I should say X, I guess, and Reddit. A lot of people every day discuss specific topics which might be in your industry.

There are a bunch of different social media listening platforms, but what I was struggling with that still a lot of this work was quite I would say manual.

Second thing is understanding market dynamics and understanding what's happening in a specific market. If you are in finance and investing, if you're considering to partner with specific industries, you've got to do this kind of fast.

And of course, you need to research the current customers if you have any, and you need to create lead generation. Today, I'm going to show you the number one and number four because we don't have time for everything.

However, if you guys are interested of how to do the market research, I don't know if this QR code is going to work. If not, we're going to share with everyone. But there is an article with all the prompts, all the guides, all the tools which I use to do the market research. So you could just go ahead, take it, copy it, and try to use it in your workshop.

Social Media Listening

Let's focus on the social media listening and understand which kinds of tools you could use. I've actually, with the help of ChatGPT, which Reggie really likes, have put together a list of companies and analyzed which are being used for social media listening, a.k.a. go ahead and scrape whatever is out there on Reddit, on Twitter, on LinkedIn and try to find the right people.

If we look at the way these platforms work, a lot of times they listen, but they are not able to do any sort of an action. So let's say I see some comment, but then I'm not able directly to reply to that. So I need to open the platform, create an account, and then go ahead and reply. And that gets kind of manual.

Up until some of the new tools emerged quite recently, which automate the whole process for you. And let me show you how that works.

Tools for Automation

Okay, here is an example of a platform called Growlix. They are doing this AI search marketing optimization.

1How the platform works is that you introduce here your business. You have to include as much contextual information as you could about your business. The second thing you do is ideally you need to include information about your tone of voice.

I am dropping the tone of voice guides very soon as well on my SAP stack. So if you guys want to check it out, that might be also helpful how to train to write in your tone of voice.

One of the hardest questions when it comes to working with LLMs. And well, another thing you could do is also add specific tags. Let's say B2B SaaS I'm interested in or I'm interested in B2C marketing or whatever the topics are where the keywords are. which could replicate your business interests.

You save, and then you go to mentions. It starts searching, today's 24th of April, yeah, and then you could see on Twitter, for example, you would find specific posts, and there would be kind of a lot, and you would see already drafted answers. And the best thing is that what you could do, you could rewrite if you don't like it.

Most times, to be honest, I work a lot with a ton of voice. I'm not very happy with the results, so I would rewrite the answer. And you could publish right away. And the answer goes directly to the platform from your company.

Why this is very cool? If someone is already searching for advice, if someone is already looking for a specific solution, they are asking already questions, this is a very fast way to get to these customers. Whether you're doing a customer research where you're trying to sell something, if you're offering any advice in a specific area, And this social selling is heavily underrated.

I work with quite a bit of companies. In the last, I would say, 18 months, I've worked with 12 startups. And out of 12 startups, I think only two did social selling. So I'd say that's heavily underutilized if we look at the scene. and very, very helpful.

The same would go on LinkedIn and the same would go on Reddit. And those are, if we speak about B2B software space in particular, one of the three main platforms, right? That's where people discuss stuff.

This is something which is simplifying my workflow quite a bit. These tools get to become quite expensive.

You would see that some of the tools which I have put in the spreadsheet, I've updated the pricing here in the free trials as well. Some of them cost thousands of dollars and they primarily target large corporations, Nestle's of the world, and like, I don't know, Nike's of the world, et cetera, someone who has that budget, but they are also cheaper tools.

I like to find underdogs. I like to find smaller startups who are able to either create a custom solution for you or be able to update some of the features or have a faster support. Why? Because they also tend to get more affordable as well rather than these giants.

Also, the giants, they focus a lot on this B2C space. They don't do so much B2B. So B2C meaning they monitor Instagram, Facebook, and not necessarily so much they work on LinkedIn or Reddit. And I think that's that's a misstep you would see here tagged as well like what platforms they are in and do they track them and you would see that some of the biggest tools they actually don't.

Lead Generation Workflows

So that's one of the ways we could work with the tools.

The other way is the famous lead generation workflows, which you could see happening a lot on LinkedIn. If you are on LinkedIn, there are these people saying, I have built this clay automation flow. You could go or whatever flow like on some platform.

You could go into this platform and you are going to be just copying this prompt, adding these things that's going to generate the leads and the contacts. And then you are going to send those emails and you are going to get responses. And I have generated $300,000 of revenue just in this month with that.

I felt I'm human, enthusiast and a human. So I felt it was like, okay, I'm going to try this out. I'm going to build this flow. I'm going to build this automation and it's going to work. It's going to build some, bring some real business to the customers.

And I was just on my phone while Roger was talking. I was checking the latest reports from one of these automation workflows we have built. To give you an idea, it costs us about $1,000, $1,500 a month for one of my clients.

And out of 3,000, 2,000, 3,000 emails sent every week, we get from 8 to 11 replies. And out of them, there are zero or one positive reply. Most times, one.

And I was like, what's wrong? What am I doing wrong?

I went around all of my network, all of the growth people in the Silicon Valley, in London, where I lived before Switzerland. In Zurich, I have a couple of people who do this as well. And I was like, guys, something is off. This thing is not working. What am I doing wrong? I had a bunch of calls with the supports of these companies. Something was off.

When something is off, what I normally do, that's how my technical brain works, I break down the problem to pieces. What is breaking down the problem to pieces is understanding how would I do this sort of manually.

Manual Process vs Automation

I decided to run this manual process on one of my clients and see what's going to happen. Let me show you what's the process and what do we get. The process is kind of simple.

It doesn't include a lot of LLMs or whatever else. You use prompt to very detailed one where you explain what is the market you're researching, the specific segments, specific jobs to be done, the types of companies you're looking for, and any details you want to generate. Then you verify the contacts because you need to find the email addresses and no GPT, no child GPT, no, I don't know, whoever you use, Claude or Gemini are going to give you those email addresses.

So you need to go to one of his platforms. There are plenty of those. I primarily use Apollo or Hunter to verify the emails.

And then after that, you go back to GPT and you generate personalized messages and then you start sending those emails. So yes, I use a lot of GPT in the process, but there's a lot of human thinking in the process of who I should target.

So what we did is... Let me show you the result first. We generate, I don't know if you could see, let me make this a little bigger.

This generates a company list with specific information about the company, like the headquarters, how much round they've raised with the recent news from the media, and a personalized email to the founder. And this kind of output allows you to just go copy paste and send the emails or set up some sort of automation and give this to your assistant to track manually. Why I like to do this manually, manually I mean sending them manually and then checking the results instead of using automation system is because I could understand what I'm doing right or wrong.

And how that works, in fact, is that there is a prompt. It was in the QR code which I shared. You could just go and test it for yourself.

I share the prompt with a GPT asking to come up with a list of companies. It generates the list of companies. Most times what would happen is I understand that my request was way too broad.

and i didn't realize this up until i started doing this manually because when i started looking at the list of 50 companies 50 companies is something i can comprehend is that sometimes we as founders or consultants look at the market and we say i want all the marketing software service solution companies as a target or i want all the e-commerce companies as a target. And then it comes up with a really random list of companies.

And one of my friends says, if you have a shitty brief, you have a shitty output. And this is exactly what happens. And I guess why the whole automated process was not working.

Insights from Narrow Targeting

So I spent a lot of time going back and forth with the LLM asking in. please create another list, now focus on this specific segment. Why that works better is because you as a human in the process try to understand how to better figure out the segment, how to better find a solution within that segment which is going to work.

And another thing which I also understood is when you start mixing up companies, for example, from different rounds, from different types of segments, the messaging kind of is also plain because you think very broad targeting. And this process taught me something is that we need to go super narrow in the way we target and build the messaging flows.

It also taught me that If you just give a task of like generate this whole list for B2B SaaS, which is a very broad segment, you're going to get a very diverse lead list.

Of course, it sounds very generic now when you think about it, but if you go to the clay automations, that's actually how that whole system works to be able to send 8,000, 3,000, 10,000 emails a week or a month. it has to go to all these types of companies and send kind of a generic message.

So to cut the long story short, the experiment is live for already three weeks. We have sent, in the first batch, 17, in the second batch, 10 emails. And so far, we didn't just get positive responses.

We had four companies starting the trial. And this was something which I didn't expect because we were like, okay, how is this going to work?

We're already around this automation flows. We still have zero trials from the automation flow. We have few positive replies in the last three months, but these started proving right.

So the founder was like, hey, let's try just scale these and work with more segments. And now what we're doing is we're building this custom list.

And the next step, which we already started with Automation Flow company, who is helping us to build these flows, is, hey, these are the types of narrow segments we want to target, these are the types of messages we're going to send, and this is what is working for us and where we get responses. And their conversion rates are also improving, at least we're getting some sort of replies from companies. Before it wasn't even replies, like any positive replies, this was close to zero, now it's close to one, a little bit of an improvement, you know.

Scaling Successful Methods

This has generated quite a few really good leads, few partnerships for us, and I'm hoping that now when we start scaling this, we would be able to maintain the quality.

There is a prompt. If you guys want to copy that, as I said, there is a QR code. You could just go ahead and try to copy it.

The thing is with copying, though, you would see here there's a lot of details included in here, right? That all should come from you, and that's where the human thinking should be.

How much AR you got, how specific companies you want to target, which specific segments, which jobs to be done. And that's where those details really matter and the context really matters. So I usually spend a lot of time on the prompt and then a lot of time negotiating with a GPT or Claude to make this thing work.

Closing Thoughts

If you guys want to connect, this is my LinkedIn.

I didn't want to make this too long. I know there is pizza already getting cool.

But if you have any questions, I'm here to answer.

I hope this was short and sweet. So let me know if this was helpful, if you want to learn anything more.

I'm here for the rest of the evening.