# Doe turns connected work context into supervised agent tasks and finished artifacts

Canonical: https://mudpie.ai/companies/doe/
Breadcrumb: [Home](https://mudpie.ai/) / [Companies](https://mudpie.ai/companies/) / [Doe turns connected work context into supervised agent tasks and finished artifacts](https://mudpie.ai/companies/doe/)
Author: Ali Abouelatta (https://mudpie.ai/authors/ali-abouelatta/)
Published: 2026-09-19
Updated: 2026-09-19
Research type: Company profile
Method: Company and accelerator sources checked 2026-09-19. Product claims are attributed to their sources; this is research, not a hands-on product trial.

Doe is an action-oriented AI work platform for teams that want agents to search their files, apps and conversations, then return a finished artifact or carry out an approved workflow. Its buyer is choosing an execution layer, not another chat window.

## What it does

Doe’s docs describe agents that can search connected tools and the web, analyze and edit files, build workbooks and dashboards, draft messages, update records, publish sites and turn recurring work into scheduled Loops. Users can inspect activity and sources, open artifacts and approve external actions before they happen. [Doe docs](https://docs.doe.so/)

The use cases are concrete: pipeline risk, CRM hygiene, market maps, pricing extraction, contract review, financial models, board metrics and recurring reports. Doe says it connects more than 75 apps and can work from web, Slack, email or text. The advantage is context continuity across files, prior decisions and tools; the tradeoff is that the platform can touch more consequential systems than a normal assistant.

## What to check

The approval boundary is the important product feature. A team should decide which actions are read-only, which create drafts, which update systems and which require a named human approval. The docs explicitly show this pattern for executive-assistant work, but the buyer should confirm permissions, audit logs, data retention, connector scope and failure behavior for its own apps.

The company’s launch describes SOC 2 Type II, HIPAA-ready handling, zero data retention with LLM providers, SSO, SCIM and audit logging. Those are company-reported controls; no account, connector or security review was used. The current public homepage was not substantive in the checked snapshot, so pricing and current packaging are best confirmed in a sales conversation.

The [YC profile](https://www.ycombinator.com/companies/doe) identifies Adrian Barbir as CEO and Richard Ou as CTO, with Ou’s prior Vapi, Wordware and second-time-founder background. That context supports the action-engine thesis, not a guarantee that every connected workflow is safe to delegate.

## My editorial take

I would try Doe first on a recurring, reviewable task such as a pipeline summary, market map or spreadsheet cleanup. Keep external messages, writes and publishing behind approval until the team has watched the full activity trail. If the work is a one-off question or the organization has no owner for connector permissions, a smaller tool will be easier to govern.

## Quick facts

| Field | Sourced detail |
| --- | --- |
| Buyer | Operations, revenue, finance, research and strategy teams |
| Product | Multi-tool agents that return artifacts and run recurring Loops |
| Integrations | 75+ apps are advertised, plus web, Slack, email and text |
| Pricing | Not published in the checked sources |
| Main fit question | Which recurring work can be delegated with visible approval? |

## Sources checked

Doe’s YC profile, public docs and listed homepage were checked on 2026-09-19.

## Cohort context

Doe is listed in Summer 2025. In our 2026-09-18 directory snapshot, 112 of 166 listed companies in that cohort have YC’s primary industry label B2B (67.5%). This is a current-directory comparison, not an original intake count or a performance ranking. [Nine-cohort dataset](https://mudpie.ai/research/yc-cohorts-2026-09-19.json).

## Public website snapshot

Observed 2026-09-19T16:18:02.176Z in raw homepage HTML. This records visible metadata and advertised links, not agent execution or product quality.

| Signal | Homepage observation |
| --- | --- |
| Product description metadata | Observed |
| Canonical link | Observed |
| H1 or H2 heading | Observed |
| Typed structured data | Observed |
| Docs/developer link | Observed |
| Pricing link | Not observed in this response |
| llms.txt link | Observed |
| Markdown alternate | Not observed in this response |

[Public observations](https://mudpie.ai/research/yc-homepage-links-2026-09-19.json) · [Collection method](https://mudpie.ai/research/yc-homepage-methods/README.md). Missing links here do not establish that a capability or file is absent elsewhere.


## Author disclosure

I cofound Lazyweb and publish Mudpie. This is an owner-written publication, not an independent testing organization. Research notes distinguish observations, sourced reporting and editorial judgment.