Tasklet AI’s rapid scale signal
Tasklet AI reported eye-catching growth—1,200% in Q1—and $5M ARR, and it’s compensating top hires extremely aggressively, pointing to a model of lean teams paid well while leveraging AI for scale. (x.com) That hiring and pay pattern suggests some startups are trading headcount for high-impact roles plus automation, a model that changes how early-stage companies think about talent and operating leverage. (x.com)
A 9-person startup in San Francisco is saying it hit $5 million in annual recurring revenue on April 8 and grew more than 1,200% this year, then closed a $20 million Series A round a week earlier on April 1. (arr.club, wellfound.com) Tasklet sells a simple pitch: tell an artificial intelligence agent in plain English what work to do, connect it to your tools, and let it run in the cloud 24 hours a day. Its own product pages say it can use application programming interfaces, webhook triggers, and even a browser in the cloud to handle recurring office tasks. (tasklet.ai, shortwave.com) That product matters because it goes after “knowledge work,” which is the catchall for jobs done across inboxes, spreadsheets, chat apps, customer databases, and internal tools. Tasklet says the old pattern of clicking through dozens of apps is being replaced by one agent that can generate its own interface, write code, and take actions across systems. (wellfound.com, tasklet.ai) The founders are not unknown operators trying this from scratch. Tasklet says the company is led by Andrew Lee, who co-founded Firebase before Google acquired it, and Jonny Dimond, and the team includes people from Google, Amazon, and OpenAI. (arr.club, wellfound.com, ycombinator.com) The company is also unusually explicit about how small it wants to stay. Its Wellfound page says the team has nine people, and its hiring copy says “nearly every line of code here is written with AI” and that it reaches for AI first when solving problems. (wellfound.com) That is the operating model investors keep looking for in 2026: software revenue growing like a normal startup, but headcount growing more like a boutique law firm. Carta wrote in August 2025 that startup hiring had slowed in most sectors while salaries rose for artificial intelligence roles, which fits a world where companies pay up for a few key builders instead of hiring broad teams. (carta.com) Tasklet’s public job listings show that pattern in miniature. The only visible role on Wellfound is a customer support job paying $115,000 to $155,000 plus 0.1% to 0.2% equity, while outside compensation data puts many founding-engineer roles closer to the mid-$100,000s in salary before equity, which makes any premium above that stand out. (wellfound.com, glassdoor.com, pave.com) The bet is straightforward: if one top engineer using strong models can do the work that used to require three or four people, then a startup can afford to pay that engineer far more cash and still come out ahead. A company with $5 million in annual recurring revenue and nine people would be at roughly $555,000 of annual recurring revenue per employee, which is an unusually high productivity figure for a young software company. (arr.club, wellfound.com) That does not mean the model is proven. Tasklet launched publicly in October 2025, says it is still in beta, and is selling a category where reliability, security, and cost control decide whether “agents” become real business software or just a flashy demo. (shortwave.com, tasklet.ai, cognitiverevolution.ai) But the signal is already clear enough to change founder math. Instead of asking how fast they can add 20 employees after a fundraise, more startups are asking whether 5 elite hires, a stack of models, and a lot of automation can get them to the same revenue line faster. (wellfound.com, carta.com, arr.club)
