Meet the Manager: how Vezra decides which agents to hire
Every Vezra workspace starts with one agent — the Manager. It runs the intake, proposes the team, and handles the hiring. Here's how it thinks.
The first conversation you have on Vezra isn't with a support chatbot. It's with the Manager — the agent that spins up with your workspace and runs your intake. The Manager's job is to understand what you're trying to do, propose the team to do it, and then hire that team on your behalf.
We get asked a lot about why we designed it this way. The answer goes to the heart of what makes agentic products feel good or frustrating.
The problem with template galleries
The default pattern for an agentic product is a template gallery. You land on the app, you see a grid of roles — "sales agent, marketing agent, analyst agent" — and you pick the ones you want. It's fine. It's also the reason a lot of these products feel like a Pokemon binder: lots of types, unclear which you actually need.
Template galleries treat the user as the expert. You have to know which agents you need, which integrations they need, and how they should fit together. For the 10% of users who have a clear picture, that's great. For the other 90%, it's decision fatigue on day one.
We added a template gallery to Vezra — it's there if you want it, behind the "hire a template directly" option on the welcome page. But we led with the Manager because the Manager solves the other 90%.
What the Manager actually does
When you first open Vezra, the Manager greets you and asks one question: what are you trying to get done?
You can answer in any level of detail. Some users type one line ("I need help running my Shopify store"). Others write a paragraph. Others pick a starter chip and refine from there. The Manager listens, asks a small number of sharp follow-up questions, and then proposes a team.
A proposal looks like: "Based on what you've told me, I'd start with Sarah as your support agent — she'll handle your customer email and Shopify tickets. And I'd add James as a growth agent for your email campaigns. Does that sound right?"
The Manager uses real first names, not role names. We think "Sarah" is a better teammate label than "Customer Support Agent v2." Names come from a rotating pool so your team feels like a team, not a spreadsheet.
The Manager is scoped tightly
One of the harder design decisions was what the Manager is not. It's not a general-purpose assistant. It won't answer unrelated questions during onboarding, help you write a press release, or triage an inbox. Its job is team setup, and it stays in that lane.
The reason is quality. An agent that tries to do everything does all of it poorly. The Manager we ship is narrow, opinionated, and fast. If you ask it to do something off-topic, it gently redirects: "that sounds like a great job for a Content agent — let's get your team set up first, then we can add one."
How it picks integrations
When the Manager proposes a team, it also proposes the integrations that team needs. These are pulled live from our connector catalog, not hardcoded — which means the moment we ship a new integration, the Manager can recommend it.
This is deliberate. We've watched competing products where the intake agent has a stale understanding of the catalog and recommends integrations that no longer exist, or misses ones that were added last week. We wanted the Manager to always be current, so we wired it to read the catalog every session.
When you want to override
You can always skip the Manager. The welcome page has a visible "browse templates" escape hatch that takes you straight to the gallery. Some users know exactly what they want; we don't want them fighting our UX to get there.
But if you talk to the Manager for three minutes, you'll usually leave with a better team than you would have picked from the grid. That's been true in every user research session we've run.
Why we put so much effort into the first agent
First impressions in agentic products are disproportionate. If the first agent you interact with is vague, or slow, or misreads your situation, you infer — correctly — that every other agent will be the same. If the first agent is sharp, you trust the rest.
The Manager is the agent that earns your trust before the rest of the team walks in. That's why it exists, and that's why we'll keep investing in it.
Want to meet yours?
Open a Vezra workspace. The Manager will be waiting. Tell it what you're trying to do.
Want an AI team that runs your work?
Vezra agents each get their own computer, log into your tools, and get things done. Start free — no card required.
Start free