I’m Kostas Hatalis, Founder of GoCharlie. 

We're a full stack generative AI company. We build small and micro language models, privately and constantly deployed for businesses in a variety of use cases. 

I’m a recent PhD graduate from Lehigh University in deep learning and first time founder.

I founded the company in the summer of 2021 to initially build a productivity AI app. From 2022 to early 2024 I raised about $2.5M.

We recently began to pivot the company from B2C to B2B as we’re now building personalized and private models for businesses.

A big part of the funding was to build out our proprietary models. Building AI models is a huge challenge, but we had the know-how for it.

We also planned to use the funding to build out our SaaS platform, team, and marketing initiatives.

Early on, we sought deep tech investors and we got some traction with them.

Our biggest checks came from deep tech VCs that saw the vision of needing to build one’s own vertical AI model.

The other capital  came from angels and VCs that were in the consumer app space along with investors who valued AI apps targeting marketing agencies and small businesses. 

This was also around the time ChatGPT had come out, so there was not yet a big explosion of ChatGPT taking everyone's lunch.

I spent all of 2023 doing outreach as a full-time job.It was mostly cold outreach and talking to people who knew other people. 

I basically spent the whole year doing nothing else other than back-to-back calls with VCs. It was exhausting.

I received about 300 rejections until I received the 30 or so VCs and angels that came into the round.

Of the 30 investors that came in, about 25 were angels or angel syndicates. Half of the angels were friends and family. The other angels and VCs that came in had specific theses for investing in AI and SaaS.

But I also pitched to hundreds of traditional VCs where things like MRR, EBITDA, and other traditional metrics (even at the pre-seed stage) is all they looked at and cared about.

I tended to focus more on getting introduced to VCs that wanted to build out a portfolio of AI companies, seeing the vision of what generative AI could become.

I strongly emphasized our vision, which is the need to build vertical models. I was hinting throughout the process at the need for businesses to own their AI.

This was primarily how I was able to raise. The vision was that we weren’t all going to be beholden to Google or OpenAI or Microsoft — there are going to be smaller players.

Even if these small players are not worth $100B, they will still capture parts of the pie. 

For both consumers and businesses, I believe it's essential for users to own their AI and have it tailored to their unique needs, rather than relying on generic services like ChatGPT.

When I initially raised, I was telling investors that I was raising $2M. That scared off a lot of early investors. 

After speaking with other founders, I heard about the strategy of reducing the round size. 

As a result, I then set out to raise $200k, and as soon as I closed that $200k, I then began raising another $200k. I incrementally raised more over time until I had the capital I needed. 

To drive urgency I always highlighted how fast generative AI was moving and that it was and still is pivotal to invest quickly so that we can move faster.

There was always a challenge in educating my audience to understand the vision. With a lot of VCs, I had to constantly educate them. 

Most of the time it would go over their head. The results would usually be a quick rejection from them.

Even today in 2024, there's still a huge lack of knowledge of generative AI — many just follow the hype. The VCs that I raised from tended to be very knowledgeable on AI.

If a VC could stand up a conversation and understand what a large language model was on a fundamental level, they were good targets for me as they could understand the big picture.

I pitched a number of big funds like Sequoia, IBM Ventures, Bloomberg, etc.

Maybe this was because I was a first-time or unknown founder, but they all told me that without substantial revenue or traction, I was too early. 

This makes sense as these VCs tend to invest in Series A companies and beyond.

But ironically enough, all these VCs later invested tens of millions into other pre-revenue and pre-product AI startups.

We've seen a bunch of these companies come up in the last few years. They raise hundreds of millions, sometimes billions, and then implode.

I’m scratching my head, saying, “What's the logic here? Why is there one standard for one founder but another for another?”

I'm still trying to figure this out. It's still happening too. Many AI startups are raising ludicrous amounts at insane valuations. I have not been able to understand why it’s happening.

Fundraise only when you need to, and fundraise early. Fundraising is all about getting people to buy in on you. 

I’ve discovered that there are only 2 things that help you raise money. 

First, revenue is king. You can close a check extremely quickly if you have the right amount of revenue. 

You can be selling AI powered toilet paper, but if you have high MRR and growth rates, VCs will throw money at you.

The other is the confidence that investors will have in you as an individual. Building that confidence is very difficult, especially those fresh out of college or grad school.

We can build bigger and better things over time if we fail at first.

If you are Pennsylvania-based, I highly recommend Ben Franklin Technology Partners. Their whole thesis is to invest in the latest technologies.

If you are more deep tech, a PhD founder, or are in SaaS or AI, I highly recommend the Stanford Research Institute. 

We worked very closely with them. Our scientists, myself included, worked with their scientists to build some models. Our whole experience was extremely positive. 

I praise them heavily because they're likely one of the only VCs out there that do research themselves. I've not seen any other VC firm actually do AI research.

If you are a consumer facing app I recommend Goodwater Capital. If you’re in retail tech I’d recommend XRC Ventures.

If you’re an immigrant founder, I’d definitely recommend Geek Ventures.

For founders building a B2B tech startup I also extremely recommend applying to the Alchemist Accelerator.

One of the best newsletters, which also happens to be very expensive, is The Information. I find that it is very up-to-date.

With books, the traditional ones like Zero to One and Blitzscaling are great.

For podcasts, I'm biased towards the ones done by the Alchemist Accelerator.

If you're looking to get advice on how to build a B2B enterprise-facing company, they have great resources. We did their accelerator as well. YC’s Startup School also has some great resources. 

I’m noticing a trend that startups are beginning to copy each other even more. I’m also noticing that this is what looks good to investors.

To me, for the last 2 years, almost all the companies in their batches look identical to each other. So maybe that’s a hint to copy others.

Especially in AI we are seeing a rise of companies mimicking each other in products and that seems to translate into success. 

For instance, Anthropic, Cohere, Mistral, and xAI are all mimicking OpenAI in some way. All these companies are now valued in the billions. 

Generative AI for writing is no different. You have companies like Writesonic, Anyword, Copy AI, and Jasper who are essentially clones of each other. But they’ve all raised quite a bit of funding.

I’ve seen a similar trend in the AR/VR and crypto/blockchain spaces. It’s all a bit odd to me because in entrepreneurship, we’ve all been taught to stand out and be different.

But from what I’m seeing, companies that copy each other tend to do quite well.