When you hire an AI employee, they land in your chat with a short greeting that names their role and what they handle, then ask what you need first. There is no setup wizard, no progress bar, and no settings panel to fill in before they can work. You customize name, persona, skills, and tools later from their dashboard, but none of that is required before the first message.
When you hire an AI employee, the first message in their chat is already sitting there before you type anything. There is no loading state and no LLM call in between: the greeting and a card of ten role-specific starting points are read straight from two files owned by that employee's template and seeded onto the thread the moment the hire is created.
You do not have to read all ten. The card shows the first four options up front and puts the remaining six behind a "Show more" toggle, and you can tick up to four of them at once before hitting "Let's go." Skip the card entirely and type your own request instead, and the employee works from that just the same. Nothing about the card is a gate; it is a shortcut for people who would rather point than type.
The difference from a settings-panel wizard is what gets asked and when. A typical setup flow front-loads every field it might ever need before letting you do anything. This employee starts from whatever it already knows about your business, opens the playbook behind the option you picked, and only asks you one question if a genuinely required fact is missing, phrased as the next step in getting your first result rather than as a blocker. It never says "first, tell me your industry, team size, and goals" before doing anything useful.
Each employee template ships its own quick-start.md file with exactly ten options, and each option carries its own card text (label, one-line subline, icon) plus the exact handoff instructions the employee receives once you pick it. There is no shared, generic list of starting points across employees; a Sales employee's ten options are written for sales work, a Support employee's for support work.
Picking a card does not send a visible chat bubble with your click summarized on screen. The backend resolves your selected option IDs against that template's Markdown and saves the full handoff text as a hidden message the employee reads as if you had written it yourself, so what the employee acts on is the same detailed instruction every time, not a paraphrase.
Selecting an option starts a real workstream, not a one-time script. The template also defines follow-on playbooks that stay available for later once the first result lands, so choosing "Draft a proposal" today does not lock that employee out of related follow-up work next week.
Before replying, the employee checks the playbook's requirements against what your workspace already knows: existing files, past conversation, connected tools, notes it has already saved. It only asks you for something when that one fact genuinely unlocks useful work, and it asks for the smallest thing, not a full intake form. A tool connection that a later step needs is checked before that step, not upfront, and the built-in CRM is never treated as something you need to "set up" first: if a save to it fails, the employee just tells you the save failed.
Hiring a full team works the same way. Every member is created already ACTIVE, on a permanent task board, with no leader interview step and no sprint-planning phase to get through before real work can start. There used to be a two-phase guided walkthrough and a leader-first onboarding gate for teams; both were removed because they delayed the same value this page is about.
The new hire interviews you in chat, configures itself from your answers, and finishes a first task.
When you hire an employee, it opens with an introduction and a conversation: what your business does, how you like to work, and what you want done first. As you answer, it configures its own behavior, preferences, and memory, no settings panels involved.
The conversation lands on a real first task from your actual workload, which the employee completes in the same session. The context you shared persists in memory and propagates to the rest of your workforce, so future hires start already knowing your business.
You hire a Support role, see its ten-option card, and tick "Draft a response to a common complaint" and "Set up an FAQ doc." The employee acknowledges both, checks whether you already have past tickets or a knowledge base connected, and asks only for whichever one of those it cannot already see.
You hire a Marketing employee and, instead of the card, type "Write three subject lines for our Black Friday email." The seeded card stays there unused; the employee treats your message as the job and gets to work without asking you to pick a starting point first.
You already have a Sales employee and hire a second one for a different pipeline. It gets its own fresh chat, its own seeded greeting, and its own ten-option card independent of the first hire, so the two do not share state or step on each other's onboarding.
You hire a pre-built team instead of a single employee. Every member lands ACTIVE on the team's board immediately, with no leader interview or planning phase gating the group before they can start picking up work.
| Before | After |
|---|---|
| New software means a setup checklist and a manual. | The employee interviews you and sets itself up. |
| You configure settings before getting any value. | The first session ends with a finished task. |
| Every new hire re-asks for the same context. | Context is gathered once and shared across the team. |
| Onboarding requires technical know-how. | If you can describe the work, you can onboard. |
The employee reads your message and works from it directly. The card is an optional shortcut, not a required step, and there is nothing else you need to close, skip, or dismiss first.
Yes, up to four options at once from an individual employee's card. If you are launching a template from the cross-employee Work Area picker instead of that employee's own chat, the limit there is three.
Only if a genuinely required fact is missing for the specific job you picked, and even then it asks for one small thing, framed as the next step, not a blocker. It checks your existing files, past conversation, and connected tools first, so most of the time it does not need to ask anything at all.
Every employee template ships its own ten starting-point options written for that role, so the card you see is specific to the employee you just hired, not a generic list reused across roles. The underlying flow, seed on hire, pick or type, minimal follow-up, is the same for all of them.
Get Started Through Conversation is part of Things you hire or set up.
Browse a marketplace of pre-built AI agent roles and full teams, or build your own from scratch. Organize them into teams with leaders, set up your org chart, and onboard new hires in minutes. Every employee arrives configured and ready to deliver from day one.