Is Autostep profitable?

No. Autostep is pre-revenue with no disclosed ARR, customer counts, or enterprise logos as of June 22, 2026. Its only traction claim is a vendor-stated example of one 10-person team uncovering $100K+ in operational waste, which is unverified.

How many people work at Autostep?

Autostep has 2 total employees as of a June 22, 2026 BuiltIn listing, founded in 2025 and based in San Francisco, with team categories limited to Founding and Go-To-Market roles.

Could AI or Microsoft replace Autostep?

Likely yes. Autostep's core function, pattern-detecting desktop activity and auto-generating agents, is LLM reasoning over logs. Microsoft already ships Viva Insights and Copilot analytics to the same buyer, and frontier labs are building computer-use agents covering the same discovery-to-automation chain.

Who is Autostep's biggest competitor?

YC peer Tasklet, pursuing an identical knowledge-work-agent thesis, reported over 1,200% growth since January 2026, $5M ARR, and a $20M raise at a $175M valuation from USV and Lightspeed, far ahead of Autostep's 2-person, pre-revenue stage as of June 2026.

Is joining Autostep as an employee risky?

Yes, structurally. Autostep has only 2 employees, no disclosed funding amount or investors, no confirmed enterprise logos, and sits between Microsoft's shipped Copilot analytics and frontier computer-use agents, both of which threaten its core discovery-layer product as of June 2026.

Has Autostep raised funding?

Autostep states it 'is raising' as of a June 22, 2026 Launch YC post, with an investor booking link tied to YC's P26 batch, but no round size, lead investor, or valuation has been publicly disclosed.

YC Spring 2026 · United States

Autostep

Desktop app that watches work, finds automatable tasks, builds the agents to run them.

Losing its protection

Assessed 27 July 2026 · Evidence as of 27 July 2026

The structural protection is eroding.

On the recordSince 27 July 2026 · Checkable from Q4 2026 · Every call on the record Checkpoint: Fundraise outcome (round size, lead, valuation) expected

The call

Autostep sits between Microsoft's shipped analytics and frontier labs' emerging desktop agents, with no disclosed customers and a strategy that has already shifted twice.

Autostep analyzes what you do on your computer and tries to turn those patterns into automated tasks using AI. The problem is that Microsoft already built similar tools into its Office software, and companies like Anthropic and OpenAI are creating AI agents that can do this kind of work directly from your screen without needing a separate service. The company would need a real customer base and proof it can do something these giants cannot to survive, but right now it has only a small team and no announced big-name clients.

Why

AI pressure: Anthropic, OpenAI, Google computer-use agents. Frontier labs are building agents that watch screens and act directly, skipping the separate discovery step Autostep sells. Could AI do the core job? Yes, likely.

Bigger players: Microsoft (Viva Insights, Copilot analytics). Microsoft already ships desktop activity analytics inside M365, bundled into per-seat pricing customers already pay. Could a giant copy it? Yes, easily.

What protects it. None visible yet: no disclosed customers, no pricing model, and the core function is a feature other tools already ship.

What would change this call

Autostep's Q4 2026 fundraise outcome (round size, lead investor, valuation) becomes public. Shows whether investors believe the data-flywheel moat can outrun Microsoft and Tasklet.
Microsoft adds task-discovery or ROI-measurement features to Copilot analytics within M365. Would show the incumbent absorbing Autostep's core feature as a bundled add-on.
Tasklet announces enterprise logos or ships computer-use automation ahead of Autostep. Would show the better-funded peer converting the same market first.

Key numbers

2 total employees, founded 2025 (2026-06-22)
Peer Tasklet raised $20M at $175M valuation (2026-06-22)
Two public repositionings in 4 months (Feb to June 2026) (2026-06-22)

Named risks

Core product is a thin LLM wrapper over activity logs, easily replicated by better-funded rivals or bundled by incumbents. Severity: high
No disclosed funding amount, investors, or valuation despite active fundraise as of June 2026. Severity: high
Zero disclosed enterprise logos or revenue at 2-person headcount undermines credibility of 'Compounding Context' data moat. Severity: high
Repositioned product framing twice in four months (Feb 2026 to Jun 2026), signaling unstable go-to-market strategy. Severity: medium
Microsoft and Salesforce can bundle adjacent ROI/analytics features into existing per-seat suites at no cannibalization cost. Severity: high

Rivals, as read in July 2026

Rival	Direction	Evidence
Tasklet	losing to	Tasklet reports $5M ARR and $175M valuation vs Autostep's 2-person, pre-revenue stage as of 2026-06-22
Microsoft (Viva Insights, Copilot analytics)	losing to	Microsoft already shipped Viva Insights and Copilot analytics to same buyer as of 2026-07-07
Frontier AI labs (OpenAI, Anthropic, Google computer-use agents)	losing to	Computer-use agents announced moving into desktop observation to automation chain, per 2026-05-21 analysis
Autotab	contesting	YC peer building 'digital robot' for repetitive desktop work, same conceptual territory as Autostep
ActivTrak and similar monitoring incumbents	contesting	Established players adding AI-scoring features per 2026-07-13 comparison roundup

Questions people ask

Is Autostep profitable?: No. Autostep is pre-revenue with no disclosed ARR, customer counts, or enterprise logos as of June 22, 2026. Its only traction claim is a vendor-stated example of one 10-person team uncovering $100K+ in operational waste, which is unverified.
How many people work at Autostep?: Autostep has 2 total employees as of a June 22, 2026 BuiltIn listing, founded in 2025 and based in San Francisco, with team categories limited to Founding and Go-To-Market roles.
Could AI or Microsoft replace Autostep?: Likely yes. Autostep's core function, pattern-detecting desktop activity and auto-generating agents, is LLM reasoning over logs. Microsoft already ships Viva Insights and Copilot analytics to the same buyer, and frontier labs are building computer-use agents covering the same discovery-to-automation chain.
Who is Autostep's biggest competitor?: YC peer Tasklet, pursuing an identical knowledge-work-agent thesis, reported over 1,200% growth since January 2026, $5M ARR, and a $20M raise at a $175M valuation from USV and Lightspeed, far ahead of Autostep's 2-person, pre-revenue stage as of June 2026.
Is joining Autostep as an employee risky?: Yes, structurally. Autostep has only 2 employees, no disclosed funding amount or investors, no confirmed enterprise logos, and sits between Microsoft's shipped Copilot analytics and frontier computer-use agents, both of which threaten its core discovery-layer product as of June 2026.
Has Autostep raised funding?: Autostep states it 'is raising' as of a June 22, 2026 Launch YC post, with an investor booking link tied to YC's P26 batch, but no round size, lead investor, or valuation has been publicly disclosed.

Timeline

2026-06-22Second Launch YC post repositions Autostep around 'Measure the ROI of Knowledge Work'
2026-06-22BuiltIn lists Autostep at 2 total employees, founded 2025, based in San Francisco
2026-06-22Autostep states it is actively raising via YC P26 batch link
2026-05-28Enterprise AI solutions guide shows adjacent platforms already selling agent ROI measurement
2026-02-27First Launch YC post positions Autostep as pre-automation task-discovery layer
2025-01Autostep founded, per BuiltIn company record

Revision history

27 July 2026

First assessed. Losing its protection. The text above is fixed as of this date and is not edited afterwards.

In the same band

Trellis Surtr Defense Systems HeyClicky