The response to our launch has been incredible — 30K views and 400+ love in 24 hours. A ton of folks booked our calendar for a demo and a ton more signed up for the waitlist; we couldn’t wait to deliver the product to everyone 🙂
Sending all the love to everyone who showed interest and took the initiative
And we are just getting started. Here’s our learnings from the time we got the idea around March, to validating and finally launching this week.
Learnings from the Launch:
- execution is the real moat but a little entropy is good for you
- talking to all kinds of folks that you can access, investors, customers, engineers, friends, fellow startups at similar stage etc,…
- the entropy from each of them would trigger so many thinking styles, and each would help test your thesis and improve
- hackathons, it’s insane how many awesome folks you could find at hackathons. in my book hackathons now rank much higher than niche meetups, last would be conferences 🙁
- frameworks are still too early to adopt for complex products like ours, they do provide no lock-in features, for ex not having to lock into single LLM provider is a huge win but you are essentially locking into the framework’s state of feature set, if you need a custom feature that’s really important for your agent, either you have to understand the framework in its entirety to add the feature or you have to start from scratch. after wasting 2 weeks juggling through 3 frameworks, came to the realisation that agentic frameworks are too early for complex production use cases like ours (imagine multimodal, memory, summarisation, persistence, HITL etc,..)
- Juggling through multiple things quickly kills your productivity in everything, when you have to manage the team, write code, talk to customers, increase reach, it’s easy to get into a mode of doing a lot of activities but not making any progress. time blocking is a blessing for me 😅. Pro-tip, use two-level time blocking, top layers sticking to a single theme like distribution or coding.
- Simple things like Post It Notes are pretty effective, have a physical aspect and you get to see the board all day from waking up to sleeping
- Build In Public and posting on LinkedIn, commenting is the best way to build the storytelling skill — most would think its hard to articulate and build a story on some technical task like adding logs to something, but storytelling is the most important skill an entrepreneur could have — how else would you be able to craft a story on every weekly update to your investors (metrics & projects aren’t going to move 2x every week 😛)
- If your O3 & Deep Research usage isn’t at least once in a day, you are ngmi
- A/B Testing the website copy with folks you trust would help a lot. Your target audience might internally care about something but would publicly appreciate only a certain type of copy because of the industry knowledge (ex: investors care about metrics but they appreciate problem-solution storytelling much higher)
- Execution is harder than most think, it’s easy to spend time in “Planning”, “Thinking” and “Researching” than being able to put your head down and keep executing through and through
- Sticking to a Schedule is the toughest thing, the pressure to meet the deadline while knowing full well that if you do that night out, you might mess up the schedule for the next few days. Following schedule is the best thing you could do to stay in the right mindset, you don’t want to spiral into “what am i doing, nothing is moving much”
- AI got much better but AI is also slow in many respects, and that destroys the flow mode like nothing else I have experienced — be extra conscious about your flow mode, and what you do when AI is taking 2-3 minutes to complete the task.
What I wish I could have done better
- I wish I got advice on the strategy from more folks, especially folks who openly and graciously offered to help in brainstorming but tolerating delays is mentally tough — “launch yesterday” is the only mantra
- I wish I spent more time talking to customers
- I wish I launched earlier 🚀