Bicycle AI

Overview

Bicycle AI was a Bengaluru-based startup that offered customer support as a managed service, using a human-supervised AI system to handle Level 1 support tickets on behalf of client companies.Founded in mid-2016 by Abhishek Nayak, Arpit Mohan, and Arvind Singh, the company entered Y Combinator's Winter 2017 cohort with a thesis that machine learning could automate the repetitive, high-volume tier of customer support — reducing costs for clients without requiring them to grow their support headcount.The thesis was directionally correct: AI-powered customer support is now a multi-billion dollar category.

But the underlying NLP and ML technology available in 2017 was not capable enough to deliver reliable automation at scale.The resulting hybrid model — where humans remained deeply embedded in every interaction — was structurally too labor-intensive to generate viable unit economics.

After a failed pivot to a trivia quiz app, the founders returned remaining capital to investors and shut down.Both Nayak and Mohan went on to co-found Appsmith, which raised $51.5M and reached a $250M valuation, suggesting the failure was one of timing rather than founder capability.

Founding Story

Bicycle AI was founded in July 2016 by three co-founders: Abhishek Nayak (CEO), Arpit Mohan (CTO), and Arvind Singh, with headquarters in Bengaluru, India. ^[1] The legal entity was incorporated as Wise Mobile Technology Inc., a US holding structure typical of YC-backed Indian startups seeking access to American investors and customers. ^[2]

Nayak brought a strong serial-entrepreneur pedigree to the founding. He had previously co-founded Gharpay, a cash-on-delivery payment startup that was acquired by logistics company Delhivery, and ClinkNow, a mobile payments startup acquired by Ezetap — both backed by Sequoia Capital India. ^[3] His experience building and exiting consumer-facing startups in India gave him credibility with investors and a practical understanding of the operational challenges facing fast-growing companies — including the customer support burden that comes with scale.

Mohan joined as CTO in September 2016, bringing engineering depth from his time as a Senior Engineer at Exotel, a cloud telephony startup, as well as prior experience at Gharpay and ClinkNow alongside Nayak. ^[4] The two had worked together before, which likely accelerated the early product development cycle.

The third co-founder, Arvind Singh, headed the AI team during the YC batch. Beyond this structural detail — surfaced in a contemporaneous blog post by team member Matt Szaszko — Singh's background, specific role, and post-Bicycle AI trajectory are entirely undocumented in available public sources.

The founding insight appears to have been straightforward: customer support at scale is expensive, repetitive, and largely rule-based at the Level 1 tier. If machine learning could handle the majority of inbound tickets automatically, companies could serve more customers without proportionally growing their support headcount. The 2016 chatbot boom and rising enterprise interest in AI-powered automation made the timing feel right. The company entered Y Combinator's Winter 2017 cohort in January 2017, ^[5] operating with a split team between Bengaluru and San Francisco during the batch.

Timeline

• July 2016 — Abhishek Nayak, Arpit Mohan, and Arvind Singh found Bicycle AI in Bengaluru, India; Nayak begins as CEO ^[6]
• September 2016 — Arpit Mohan formally joins as CTO ^[7]
• January 2017 — Bicycle AI enters Y Combinator Winter 2017 (W17) cohort ^[5]
• March 6, 2017 — Team member Matt Szaszko publishes "3 Things I Learned Working at an AI Startup in Y Combinator" on HackerNoon — the most detailed contemporaneous public account of the product and team ^[8]
• March 21, 2017 — Bicycle AI raises $120K seed round led by Y Combinator, with Bossa Invest as co-investor ^[9]
• October 2017 — Arpit Mohan departs as CTO, likely coinciding with the point at which the AI's technical limitations became insurmountable ^[7]
• August 2018 — Abhishek Nayak's CEO role ends; founders return remaining capital to investors and wind down the company ^[6]
• September 2018 — Nayak joins Accel as Entrepreneur in Residence, beginning a deliberate period of reflection before his next venture ^[10]
• June 2019 — Nayak's EIR tenure at Accel ends ^[10]
• July 2019 — Nayak and Mohan co-found Appsmith, an open-source low-code internal tools platform ^[11]