Our startup had its 6th anniversary last week during a very exciting time for us. @TaskletAI is on an absolute tear, growing 92% MoM right now riding the hype around @openclaw. We have the right product at the right time and we feel incredibly fortunate.
But it wasn’t always like this! The road to get here has been incredibly bumpy — and on this anniversary I’ve been reflecting on how far we’ve come to get here.
Founded as a "Slack killer"
We started as Subspace Communications Inc in early 2020 with a team of 7 — all ex-Firebasers (aka ex-Googlers). Our plan was to build the killer workplace communication tool that would replace both Gmail & Slack in one fell swoop.
We were arrogant and foolish after a long string of successes at Google with developer products, so when our initial product idea failed to land with customers, we kept pushing, believing we just had to get the details right.
Doubling down on email
Another 6 months later though it became clear this wouldn’t work, so we pivoted — this time focusing on bringing back many of the best feature of Google Inbox and aiming at prosumers.
This went better with alpha users and internally we loved using our product, so we kept building. We put a ton of work into the brand & UI polish and in building scalable and complex backend systems believing this would be a rocket ship when it launched.
And initially - it was! We launched @Shortwave to huge fanfare. Lots of people missed Google Inbox, and we got a ton of incredible press. Tens of thousands of people signed up. Our servers strained under the load.
But then came the churn — week after week we kept bleeding users faster than we found new ones. Our numbers went down month after month.
So we pivoted again — this time focused on the one area we knew there to be serious demand: high-end business users like the ones using Superhuman. We streamlined the product UX, tweaked the brand & added power user features, but it didn’t make much difference. Users had no reason to choose us over market leader Superhuman.
ChatGPT changed everything
But then ChatGPT launched, and the practical utility of AI suddenly became clear. In a memorable moment at an offsite our advisor @JamesTamplin gave an epic demo of what AI could do if combined with email data.
So we pivoted again — this time going all-in on AI. We were going to be the AI email client.
Hard choices
This pivot was harder. We had over-hired, and for the wrong roles, so had to do layoffs. We had a lot to learn technically, and many product sacred cows had to be slaughtered. But we moved quickly to adapt and launch our initial AI features — email summaries, drafting assistance, action item extraction, and so on.
The Shortwave AI Assistant
The real killer feature though took a little longer to build: our conversational AI Assistant. At its core was a search stack that combined semantic & metadata searches to enable really good AI-powered Q&A. It was a hit, and revenue started to grow.
Growth was still not as high as we needed though, and one of our 3 cofounders departed. We continued to look for ways to accelerate the business, including adding team collaboration features to our email product for customer support & sales use cases, but nothing worked. We kept the team lean to conserve cash.
Our first AI Agent
In fall 2024 my brother @binarybits shared a podcast episode where @labenz interviewed the founders of @boltdotnew where they explained how their wildly popular agent worked. Inspired by what I heard, we started experimenting with a fully agentic approach in Shortwave, and… it worked! Really really well!
When we shipped our new agent, growth re-accelerated. Users were no longer limited to just looking up info and doing a few simple tasks like writing emails. The agent was smart enough we could wire it up to every feature in Shortwave, and beyond.
All-in on AI agents
This was the shot in the arm we needed to keep going, but we were again out of position to seize the opportunity fully. We’d run out the clock on our funding, and our team again did not have the right set of people to capture this opportunity. We did another round of layoffs, filled some key team gaps, and raised a bridge.
We started adding integrations with other tools — CRMs, project management software, messaging apps. Then HTTP MCP came out, and the ability to connect any tool quickly become clear. We doubled down and were among the first clients to support HTTP MCP.
Pretty soon we had users using Shortwave who had no interest in using our email client. They just wanted our AI agent & integrations, but wanted to stick with Gmail for their UX. How odd! We started getting a lot of messages from customers like this:
Tasklet
It was starting to look like the use cases for an AI agent with really good integrations went way beyond email, and that the email client form factor was getting in the way for non-email use cases. So we made a bold bet: build a brand new product focused on AI agent automation across integrations.
We took everything we’d learned about building agents & integrations and started work on @TaskletAI. We moved as quickly as we could to get it into the hands of customers, with our first real users using it in prod in less than 6 weeks.
A few months later we launched publicly and it was a massive hit. The feedback was incredible and revenue growth took off. People wanted more integrations, more triggers, more computer capabilities, smarter agents, etc.
We worked furiously to keep up, and each new launched accelerated growth further. Our v2 agent architecture, sandbox, and our redesign were especially impactful. AI coding tools let us move incredibly quickly despite our small team size (just 8 people).
To the moon 🚀🚀🚀
In January, Tasklet alone added more recurring revenue than we’d added in the first 4 years of Shortwave, and Shortwave was growing too. We finally feel like we’re on the rocketship we set out to build.
With the pace of change in AI and the leverage that AI coding gives us, I expect the next 6 months to be even crazier than the last 6 years combined. We have launches lined up already for Feb & March that I think dwarf everything we’ve launched today. It’s going to be fun.
Never give up!
So, if you’re another startup founder — don’t give up! There’s always a move to make. Just be ruthlessly honest with yourself about what you’re hearing from customers and what is needed for success, and move as fast as you possibly can.