Company profile · 2 min read
Meridian: cited agent work for management consulting teams
Meridian describes an agentic operating system that turns a firm’s precedent into cited client-ready work in a single-tenant cloud.
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
Published · Updated
Meridian connects a consulting firm’s prior engagements to research, drafting and delivery workflows with citations and permission controls.
What it does
The company homepage says Meridian unifies engagements, decks and models. Its product page shows a workflow that searches prior work, drafts an assessment and delivers a deck, model and summary with citations. Security says each tenant is isolated and agents mirror existing permissions.
Buyer and task
Consulting firms that want to reuse prior engagements while keeping permissions and source citations attached to deliverables.
Workflow boundaries
The page example is illustrative product UI, not an audited engagement result. Pricing and implementation timeline are not documented.
What I would ask
Ask what counts as comparable precedent, how citations are checked, and which deliverables remain human-owned before client delivery.
Why it fits
Verra’s public product page describes Meridian as a way for consulting firms to search prior engagements, use that context in a new assessment and deliver a deck, model and summary with citations. The security page says tenants are isolated and agents mirror existing permissions. The useful idea is not “AI writes a deck”; it is that a firm’s prior work becomes a governed source for the next piece of client work.
That can fit a consultancy with repeated industry questions, reusable models and a painful search problem. A team could begin with one engagement where the relevant precedent is known, compare the generated answer with the original source and keep the final deliverable under a consultant’s control. The same workflow can fail if “similar” work is not comparable, if old assumptions are mistaken for current facts or if a citation is technically present but not useful to the reviewer.
The public UI is an illustrative product surface, not an audited engagement outcome. Pricing and implementation timing are not documented. The decision gate is therefore provenance: can a consultant tell why a piece of prior work was selected, correct it, preserve client permissions and remove confidential material before delivery? This profile uses the company’s current Meridian naming and the official Speedrun identity as listed in the official Speedrun profile.
Quick facts
| Field | Sourced detail |
|---|---|
| Buyer | Consulting firms with reusable engagements and deliverables |
| Inputs | Prior engagements, decks and models |
| Outputs | Cited assessment, deck, model and summary |
| Controls named | Tenant isolation and permission-aware agents |
Sources checked
official Speedrun profile · company homepage · source page · source page
Sources checked — September 20, 2026.
About the author
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
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.
First1000 ↗ · X ↗