Podcast Host at Relevant Software

Total recall with Fireflies.ai – an AI platform that transcribes and analyzes your meetings:How to learn by trial by fire


Relevant Founders

Listen to our podcast in which tech founders reflect on their journey of building a successful startup and reveal their secrets to success.

Youtube Logo
Apple Podcasts Logo
Spotify Logo
Google Podcasts Logo

With the world going digital and remote work being the future, the number of meetings has increased, resulting in Zoom overload. Non-stop meetings are merging into one, and crucial information is being overlooked. So, what if everyone had their own AI assistant? That was the inspiration for Fireflies.

We at Relevant Founders recently had a fascinating conversation with Krish Ramineni, CEO and co-founder at startup Fireflies.ai, about making meetings more productive in terms of preparation and execution while maintaining their value. 

YouTube player

Fireflies concept and purpose

Many tech leaders are now working on chatbots, conversations, and natural language processing systems. But they are mainly intended for customer-facing organizations such as sales, contact centers, and support. Krish Ramineni and his co-founder Sam Udotong had a much broader, whether it’s one-on-one, HR, recruiting, or sales meetings. 

It all started with their AI notetaker named Fred that automatically joins meetings and starts recording, transcribing, analyzing, and summarizing. After the meeting, the AI assistant sends a full transcript to whomever the organizer chooses, and can also surface action items, tasks, metrics, pricing, and other topics of interest.

Now, customers have begun using this platform for cases that the Fireflies founders hadn’t even considered, such as sending Fred to meetings they can’t attend and reviewing the notes later.

Building architecture “on the fly”

“Different companies take different approaches to architecture,” says Krish. “Most startups that don’t have a lot of funding, like us, bootstrap; we wrote the code, glued it together into one giant monolith, and threw it up on the wall. It made no difference whether the code was neat or ugly. The most important thing is that the product works.

Other companies, such as Apple, have a lot of funding and resources. They are perfecting everything from the code level, the hardware level, the chips, and the actual end product.

So there had to come a point when we realized we couldn’t settle for a band-aid solution. As a result, there was a transition period in 2019 during which we ripped everything out and rebuilt the architecture from scratch.

That was essential because, with the pandemic, volume shot through the roof. After all, everyone went remote. We would have been in a complicated situation if we hadn’t done that before January 2020; even in 2020, we hadn’t completed the necessary migration, microservices, and architecture changes. As a result, we had to do it on the fly.” 

Hard work behind the scene

“It’s been quite a journey over the last few years. We introduced GPT and open AI to improve the summarization techniques so that it looks like a human summary, and we are happy about it. But that’s always a work in progress. A lot of infrastructure goes in, so the team is built 24/7.

But at the root of it, we’re also building a massive voice-over IP company, meaning we have to have the infrastructure, real-time streaming, the architecture of YouTube, Twitch, or Skype and having Fireflies join all these hundreds of calls. 

I always say I can be late for a meeting, but fireflies can’t. Fireflies always have to show up and work 24/7 daily, every hour of the day, on weekends and holidays.”

True perfection is never achieved

“We’ve had to take a hybrid solution where you use out-of-the-box services like AWS, servers, or Kubernetes. It helps you deploy faster. Then there are other parts where you have to build things independently.

Sometimes, as much as I hate it, you have to put something out there that’s imperfect, then shape it, improve it, and iterate on it. And then your customer’s stress tests it, and breaks things along the way. So your true feedback is when it’s in the hands of real people using it because they find bugs, edge cases, and things you would never identify. And so, I think we had to learn by trial by fire.

True perfection is never achieved; you must chase it forever. Every time you reach one checkpoint, you will see new problems, scales, and volumes that the system cannot cope with. It’s our ongoing obsession with perfection that drives us.

Learn to listen better in order to succeed

Regarding leadership, you must be a good listener and be able to process things from multiple people’s perspectives before making a decision. Sometimes there is no right decision, you might have to pick a choice; go with your gut!

Teach your team to think a certain way. Sometimes I felt like, if there’s a problem, I have to go and jump into that problem and solve it myself. But instead, if I define a problem, I now bring it up for discussion. Let people work on the problem to find a good solution. Sometimes they’ll come up with a better solution than you could have ever come up with.

Global team for global product

Initially, we had no funding, so we needed to be very resourceful. In 2021, we had to hit what I would call team blitz scaling. So, we had to build teams across the globe. And we realized global could be a strength for us because we are running a 24/7 product, and Fireflies is being used across 1000s worldwide. So, we’ve had to build a platform where it works 24/7.

Campfire program: only the best remain

If we’re not improving ourselves or not getting better by learning from our mistakes, then, what’s the point? Our whole reason for existence is trying to perfect and craft our skills.  This is what we are striving for.

A growth mindset is a culture of always trying to learn, be inquisitive, and do things you don’t know. It is important, especially in a remote context. It’s easy to do the bare minimum because your manager is not monitoring your effort. 

In Fireflies, you have to be proactive; it takes a certain type of DNA to work for our company – and it’s intense. So, we created a program called The Trial by Fire or campfire program. If people cannot excel in the first one-month or two-month checkpoint, then they can’t keep up with the pace here. It is way too intense for them. 

Many folks in Fireflies work 40 hours a week and those will be the hardest 40 hours you will ever work compared to working for a corporate company. At a startup, unfortunately, there is no other way than to work hard. 

Fail-fast philosophy with a small correction

“The fail-fast concept is great. In the early days, you’re experimenting with lots of features. But you don’t want to fail too fast when looking at your core team and infrastructure. 

Product market fit is not something that happens once; when you’re building something, Product Market Fit occurs everywhere. It happens in the people you hire, the teams you make, the features you roll out. So we must also keep trying to keep pace with the market.”

From PM of Microsoft to CEO

“I was a product manager at Microsoft. Being a PM means you have influence but no authority. You can make changes, but you don’t manage a team where people report to you directly. They don’t owe you anything. 

From my time at Microsoft, I realized that great products require a lot of dedication and attention to detail. Even if you were shipping a small feature at Microsoft, you had a bunch of people look at it and review it before deploying. So with this engineering culture of shipping, great products came.

The next crucial thing I learned was that respect from engineers comes if you know what you’re talking about. But there’s a common thing in the industry where PMS gives wishy-washy high-level feedback. And the engineers do all the hard work alone. 

But I am different and don’t want to make my engineers’ lives harder. So, I will do everything to remove all the roadblocks and ensure they don’t have to spend extra energy doing things they don’t have to write code for. Therefore, I have to be very clear in my thought process and say no to too many things. I think one of the most important life lessons is learning to say no.

And I think when I came into Fireflies and started building a team, it was that on steroids because there is no right or wrong answer to what you’re doing. In reality, you’re responsible for the ideas, the data collection, the building, the shipping, and the ultimate reaction to what you ship. 

I believe you’ve gotta be in the weeds, roll up your sleeves, and get dirty. That’s how you earn the respect of your engineers, not by writing a high-level spec or being wishy-washy in your answers or vague.”

What the future holds

“Even when I look at it today, I think voice has become ubiquitous. You see voice assistants in the household, the B2B, and the enterprise space. I’m amazed at how fast transcription accuracy and stuff have improved drastically in the last three years. Technology, such as voice recognition, AI, transcription, and NLP, has become democratized and is used regularly both in the cloud and on mobile devices. 

Also, I’m really excited about AI understanding intent and coming back with answers. Imagine doing that with all of your team’s knowledge. You can go back and say, “What was our architectural decision six months ago? And why was it done, and who decided to do it? Theoretically, Fireflies could pull that up. Like, once we build this sort of technology, that is incredible. I mean, that would make life 100 times easier as well.”

CEO’s advice: just do not give up

Finally, we asked Krish Ramineni what advice he would give other founders to gain such recognition and get on the cover of Forbes 30 and Entrepreneur.

“Honestly, the Entrepreneur magazine cover or Forbes is a by-product of focusing on building products that customers love and making them better. And all the other stuff results from creating something that satisfies the needs of customers. And if they like it enough, they’ll talk about it. 

The greatest form of marketing is when someone speaks positively of you and recommends you to another person rather than you promoting yourself. 

I take a responsibility to improve as much as possible. Yes, I’m going to make mistakes along the way. But I don’t give up. Half the battle is not to give up.”

Written by
Podcast Host at Relevant Software
I am tasked with engaging with tech experts from all across the globe. I aim to provide an open, comfortable and informative space for our guests to open up and share their trials and tribulations of creating their product. We at Relevant provide our audience with the knowledge they need to better develop their product.

Do you want a price estimate for your project?


Do you know that we helped 200+ companies build web/mobile apps and scale dev teams?

Let's talk about your engineering needs.

Write to us