Vikram (00:00)
In a month, we went from 50 paying customers to 5,000 paying customers. It's intense. That's the thing they talk about. You can't figure out how many servers to add, how to scale the system? There's the classic signs of a product market fit where growth comes and you are just scrambling to deal with it.

Roland Siebelink — Host (00:51)
Welcome, welcome, welcome! Back at the Breakthrough AI Operators Podcast. I'm so excited that we already have another guest on the show this week. Today's guest holds a PhD in biomedical engineering. He built an enterprise software company from scratch to 350 employees across 66 countries, and then sold it. His company, Pictory, has turned 20,000 organizations away from traditional video production with a team of 60, total funding of 4.7 million in an industry where competitors are raising hundreds of millions. The question is, how? With that, everybody meet my guest, Vikram Chalana, the co-founder and CEO of Pictory. Welcome to Breakthrough AI Operators, Vikram.

Vikram (02:10)
Thank you, Roland. It's a pleasure to be here. Thank you for that really warm introduction.

Roland Siebelink — Host (02:14)
Many of your competitors have raised hundreds of millions of dollars. And you are very capital efficient. What is it that you see different about this industry from all the other companies in your field?

Vikram (02:45)
I think there's one big decision that we made early on. We were very early, generative models were just starting to come when we started the company in 2019. The Transformer paper was announced, and we saw what we could leverage out of the box. So we decided that we're gonna actually leverage everything that comes out of the box. We're not gonna build our own models. We were seeing the writing a little bit on the wall, even for generative image and generative videos, that if text is taking so much compute and so many resources to train, images and videos are just gonna be a lot. And we said, this is great. I'm sure there will be a lot of models that we'll have access to in the next couple of years if we just wait long enough. And I'm so glad that bet paid off.

Roland Siebelink — Host (04:39)
Was there also an experience dimension — like I've typically seen industries evolve like this, I can tell where the world is going?

Vikram (05:27)
Yeah, I wish things would translate in the AI age as good as you imagine, but they didn't. Predicting where markets are going, I think that's the hardest thing for anyone. I did not imagine the AI to grow so quickly and so well.

Roland Siebelink — Host (06:15)
You were already starting on this before ChatGPT came out and started off the generative AI revolution, right?

Vikram (06:24)
That's exactly right. We saw it slowly percolate and I was expecting that speed. And then suddenly ChatGPT came out, it just took the world by storm and that was completely unexpected. We rode the wave too a little bit.

Roland Siebelink — Host (07:00)
What's the current state of the different models out there? How much choice do you have? Is there a big risk of being too embedded with a certain provider?

Vikram (07:12)
There are three types of models that we use. Text models — Claude and OpenAI are just battling it out, leapfrogging each other every so often. That's great. I love that. We integrate with both for different use cases. On image and video, Google is amazing for both — their Veo models are probably best of the state of the art. We also integrate with ByteDance, Pixverse, and Amazon Nova models. But it feels like Google has a step up on all of this.

Roland Siebelink — Host (08:36)
You said the workflow you provide on the front end is actually very sticky for customers. Can you talk more about that?

Vikram (08:53)
Yes. At the end of the day, we are betting on that being our winning strategy. If you're able to integrate with the systems that customers use, if you're able to pull data in and understand the format in which they're used to and then output that in the formats that they're used to, then I become embedded in their workflow — whether it's for creating training videos or for marketing videos. That's where we really wanna play.

I'll tell you the story because it's very interesting. In a month, we went from 50 paying customers to 5,000 paying customers.

Roland Siebelink — Host (09:42)
Wow. That explains your grey hair, Vikram.

Vikram (10:04)
It's intense. That's the thing they talk about. You can't figure out how many servers to add, how to scale the system? There's the classic signs of a product market fit where growth comes and you are just scrambling to deal with it. Every system fails. The way we found it was also interesting.

Before that, we were trying to find the fit in larger companies, in product marketing teams in slightly larger organizations. And we had some success, but it was a struggle. For videos, almost all marketing teams had agency relationships and legacy agency relationships. They were very hard to break. And for customer-facing videos, the pixel-level accuracy was super important to larger companies.

We had this insight that maybe what we need to do is try a different market. Product-market fit is a combination of product and the market, right? Same product on a different market. And we found a distribution channel — AppSumo. We said, okay, let's do this campaign on AppSumo. They're a different market — SMBs, agencies. And that worked. The problems were the same, whether it was a large company or the small company. But they didn't have the budgets for agencies and they didn't have that quality bar. It just worked.

Roland Siebelink — Host (12:34)
It's almost like what Clayton Christensen called disruptive innovation — your product may not have been good enough to hit the top of the market, but at the bottom of the market there was actually a huge opportunity.

Vikram (12:48)
That's right. And now we're able to move upmarket because as the product has matured, as the technology has matured, we're able to start penetrating that market too.

Roland Siebelink — Host (13:00)
I wanted to ask you about your own personal journey. You started off as the CEO of a company that grew quite large, but then you consciously stepped back and became the CTO. And now you're in a CEO role again as founder. Which one is actually running Pictory right now?

Vikram (13:36)
My co-founder is the CTO. I'm running it as a CEO now.

Roland Siebelink — Host (13:46)
Where do you draw the line between the two roles?

Vikram (13:55)
It's how much you are spending on the tech side. If I'm spending my majority of my time on the tech side, then I really should be the CTO. But if I'm spending time on the market, with customers, with investors, that's more of a CEO. The hard thing is giving up control. That's the hardest thing because you built the baby, now somebody else is making decisions. Being okay with that, being comfortable with that was the hardest thing.

Roland Siebelink — Host (16:26)
What would you say to people who are newer to the founder-CEO role that are struggling with delegating?

Vikram (16:50)
This is gonna be key to scaling. You have to work on the business, not in the business. That is absolutely key. If I'm involved in every decision and I am also coding and I'm running campaigns, I'm working in the business. Versus, on the business, I'm building the team, I'm empowering them, I'm making sure that the processes are defined. You can get to a certain size easily working in the business and micromanaging everything. And then at a stage you have to delegate; otherwise, you're just not gonna have any balance in life and the organization suffers.

Roland Siebelink — Host (17:51)
Was the product-market fit something you were consciously hunting for, or did it just happen?

Vikram (18:29)
No, no, it was consciously looking. It was a lot of iterations before that to even get to this point. We have a rolling set of experiments — either in terms of market or product changes — and we're just running through that. Sometimes you just abandon something; we try it and it didn't work. Other times, double down on something. Data is actually one of our core values.

Roland Siebelink — Host (20:18)
I should have looked them up and then I could have asked you all these challenging questions about how much you live your own core values.

Vikram (20:30)
I can talk through it because we live it ourselves every day. We call it CREDO. The first one is Curiosity. R is Respect. E is for Expeditiousness — moving fast. D is Data. And O is Openness or transparency. This is how we operate. It defines me, it's my personal values, my co-founder's values — we all align to that completely.

Roland Siebelink — Host (21:18)
Have you had difficult moments where you had a clash with core values and you had to decide to take action?

Vikram (21:39)
Yeah, there are times we've had to fire people. They were hiding things. It clashed against our core values of being open about things. There are times when we had to lay people off and it's really hard. But trying to exit somebody with respect is a big deal. And to me, that was one of the most important things.

Roland Siebelink — Host (22:46)
Somebody who was a fit in the first one million may not be the fit for the one to ten.

Vikram (22:56)
Yes. And somebody who brought you to ten million may not get you to the next milestone, the hundred million.

Roland Siebelink — Host (24:43)
How do you keep that zeal alive of always wanting to do better, not becoming a boring old corporation too quickly?

Vikram (24:43)
Sometimes I feel like if a product works well and it's doing the right thing for the customer, you probably want to leave it alone. I think a lot of products die because the founder got bored and added too much complexity into it. I certainly was tempted towards that in my previous company — Winshuttle. It just worked. It was a beautiful product that connected SAP with Excel and SharePoint. And then we started adding more complexities around workflows and collaborative Excel and stuff like that. We made it too hard for the end user.

My coping mechanism: if the product works and your go-to-market engine is also working around that, keep that going. Use the go-to-market engine to add new things so your NRR can potentially improve, as opposed to re-architecting the whole product every couple of years. Try to resist fixing something that is not broken.

Roland Siebelink — Host (27:06)
What was Vikram like as a child, as a teenager? Was there anything predicting that you would be a successful founder-CEO one day?

Vikram (27:06)
I don't think so. I know I was very curious. I was very driven as a child. I grew up in India — a very competitive society. I did not see myself as an entrepreneur when I was a kid. Although I was surrounded by entrepreneurs in my family. My grandfather was probably the most inspiring entrepreneur in our family. He was in the trucking industry and he was running his business until he was beyond 90 years old. I saw him and the passion and the zeal. Explicitly, it didn't appeal to me at that time, but I think that left a mark for the future. Because I was pursuing to be an academic. I did a PhD. My first job was a research scientist.

Roland Siebelink — Host (28:47)
What drew you back to entrepreneurism in the end?

Vikram (29:00)
Even as my first job as a research scientist, the thing that I got very excited about was product management — trying to build a new product and then showing that product to customers. Bringing that from zero to existence and having their eyes light up, that part of the journey just fascinates me.

Roland Siebelink — Host (29:27)
Last question, Vikram. What's your one piece of advice for founders just starting out?

Vikram (29:47)
Sometimes people don't take enough chances. They're too risk averse. And I think even when things are going well, you still have to continue to take the risks. Especially if you're starting out, you have to jump in. Just do it, just jump in.

Roland Siebelink — Host (29:55)
Well, Vikram, if people want to hear more about Pictory or about you, where do they go?

Vikram (30:24)
pictory.ai is our website. There is a free trial. Just try it out there. Turn your blogs to video, turn your idea to video.

Roland Siebelink — Host (30:41)
Thank you again, Vikram. This has been an amazing conversation. I'm sure the listeners are gonna find this super useful. Thank you for being on the show.

Vikram (31:06)
Thank you, Roland. It's been a pleasure talking to you. It's amazing.

Roland Siebelink — Host (31:10)
Likewise. And for the audience, we will have another amazing founder on the show next week. So keep listening, stay tuned.