Run a solo consulting practice on Claude Cowork agents with a three-pass quality check before anything reaches a client
A former strategy consultant describes a plugin that gives a solo consultant agents for proposals, kickoff packs, research and deliverables, with source checking and two independent audits before delivery.
Evidence: The author reports this. We have not checked it beyond reading the source.
The business problem
A solo consultant has to write proposals, plan engagements and produce deliverables alone, and cannot trust unchecked AI output in front of a client.
What was tried
After a one-time onboarding of two to three hours in which the consultant supplies methods, templates and examples, each engagement gets its own folder that the agent can see, which keeps client data separate. Agents draft a proposal from a request for proposals, a kickoff pack, a first deliverable and a contract from a template, with the consultant signing off. Quality control runs three passes: trace each claim to a source and rate it verified, probable or not established; have an isolated sub-agent check the document against a review grid and rebuild the model; then have two models from outside the system audit it. Findings are fed back and the audits repeat.
What was reported (mixed)
The founder says the plugin gives a 25 times average speedup measured across its users, with quality maintained, and that a draft moves from roughly 85 percent to near-final in about half an hour. In the demo, the source check rated 47 claims verified, 14 probable, 9 hypotheses, 13 contradicted and 6 unverifiable, and the external audit held the first draft until it was revised.
Limitations
A podcast interview with the product's founder, who promotes the plugin and a discount code. The demo consultant and client were invented, the timings are for one run and not a benchmark, and the speedup and 'client ready' figures are the guest's own estimates with no client numbers. The internal audit uses the same model family as the work, so it may miss shared errors. The guest says AI makes many mistakes and does not recommend full automation.
What you need
A Claude Cowork subscription and the paid plugin (about 400 to 500 euros a month at the quoted prices), plus your own templates and past work to teach it your style.
Sources
- Umbrex Unleashed podcast (guest sells the product discussed) ↗ Vendor case study, publication date unknown
Source published: unknown. Last reviewed here: October 12, 2026. Spot a mistake? Tell us.
Tools in this workflow
- Claude Cowork: Runs the agents inside a folder per clientTry Claude Cowork
Related workflows
Run a one-person business with a fleet of single-job Grok Bots with approval gates
A solopreneur describes a fleet of named Grok Bots, each with one job and explicit anti-jobs, handling briefs, inbox triage, research and social drafts, with approval before anything is published or sent.
Install Hermes Agent for scheduled jobs and a self-built skill library, reachable from Telegram, Slack or WhatsApp
The README for Nous Research's open-source Hermes Agent describes persistent memory, skills it creates after complex tasks, a natural-language scheduler and one gateway for chat apps.
Set up OpenClaw as an always-on agent for client onboarding, KPI snapshots and content drafts, with approval steps
A hosting company's tutorial lists 25 OpenClaw workflows, including scheduled KPI posts, client onboarding folders and emails, expense logging and content repurposing, with security advice for running it safely.
Run a four-agent founder team in one Telegram chat with OpenClaw
An AI phone vendor's guide describes a solo founder template with an orchestrator, a business agent, a marketing agent and a dev agent reached through one Telegram chat, with claimed time savings.