Builder.ai

» see builder.ai as it lived, on the wayback machine

~ the goodbye ~

builder.ai was in my linkedin feed on and off from 2020 onward. every time a paid-promoted 'we build your app with AI' ad ran, it was them or a clone. i never clicked. edge vision was the harder problem and it was already my job.

sachin dev duggal registered engineer.ai in london in 2016. the pitch was that a normal person could describe an app in plain english, a friendly assistant called natasha would help scope it, and the platform would assemble the codebase from prefabricated blocks with a light human review layer on top. it rebranded to builder.ai in 2019. deepak chopra was an early advisor. microsoft came in for a strategic investment in 2023 as part of a 250 million dollar round led by the qatar investment authority.

in august 2019 the wall street journal ran a story titled "AI startup boom raises questions of exaggerated tech savvy" and used engineer.ai as the lead case. sources described a workflow where customer briefs were handed off to human engineers in india. natasha was more of a project-management chatbot than a code generator. the piece did not kill the company. it just sat there as an unpaid receipt while the funding rounds kept coming.

by 2024 revenue was under review. an internal audit found that 2024 sales previously guided at 220 million were closer to 55 million. bloomberg reported round-tripping schemes with an indian content company called verse, invoices going back and forth to inflate revenue between 2021 and 2024. duggal stepped down as ceo in february 2025. manpreet ratia took the seat. by may the lender viola credit had seen the audit numbers and seized 37 million from the operating accounts, leaving five million in restricted funds. ratia announced the insolvency on 20 may 2025.

~ the chronicle ~

~ what the pitch actually was ~

the marketing was "you describe the app, natasha builds it." the reality inside builder.ai was closer to a low-code outsourced-dev shop with a friendly frontend. that is not itself a bad product. plenty of agencies have made money on managed offshore dev work. the sin was pretending the AI part was the load-bearing wall when it was mostly wallpaper. investors were not buying an agency. they were buying an AI story.

~ the audit ~

the round-tripping mechanic that surfaced in the 2024 audit was old-school revenue laundering. builder.ai and verse would each buy roughly equal-value services from the other, book both sides as revenue, and let the cash cycle back through. neither company got any real economic value from the trade. both got a bigger top line to show fundraisers. bloomberg's tally suggested the pattern ran for three years. by the time viola credit worked out the actual receivables it was already game over.

~ what it left behind ~

about 1000 people lost jobs across five countries when the platform went dark. duggal remained subject to a UK national crime agency inquiry into alleged money laundering, a separate matter unrelated to the round-tripping. some of the customer projects were mid-build when the servers stopped. the wider AI-app category kept selling. by mid-2025 there were half a dozen new no-code AI-builder startups pitching the same shape of thing to the same shape of investor, this time with actual llm generation under the hood.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »