Skip to main content

YingClaw Multi-Account Setup Guide: Running Your Whole Team on One Platform

The first management headache a team hits after adopting digital employees is accounts: does everyone register their own? Or share one? If sharing, how are tasks divided? Do memories get mixed up? If these questions aren't answered before launch, a month later the platform turns into a black hole where "everyone uses it and everyone says it's a mess."

YingClaw supports multi-account collaboration. Configured properly, one platform manages the entire team's automation work. This is a follow-along tutorial — step by step, you'll have the account system sorted in ten minutes.

What multi-account collaboration needs to solve​

Clarify the goal before configuring. Multi-account collaboration boils down to three things:

  1. Identity isolation — each person's tasks, memory and configuration never interfere; A's work records never leak into B's
  2. Resource sharing — team-level skills, knowledge and common instructions are pooled so nobody reinvents the wheel
  3. Controllable permissions — who can see, modify and execute what, with an admin who has full visibility and can adjust anytime

When all three are in place, the platform becomes a true collaboration hub for the team — not just a pile of accounts.

Step one: plan the account system — one account per person, power organized by role​

The first step isn't creating accounts, it's planning. The recommended pattern: each person gets an independent account, and capabilities are organized into roles within accounts.

Concretely:

  • Person = account — each member logs in with their own account; task records, memory and personal configuration follow the person
  • Role = capability group — configure roles (managed by the admin or within accounts) such as "sales assistant," "admin assistant," "operations helper," each with its own instruction set and default behaviors
  • Team = organization — the admin manages members, roles and shared resources centrally; onboarding and permission changes all happen in one place

The benefit: personal accounts keep memory and tasks from cross-contaminating, while the role system keeps capabilities and permissions clear and controllable. Avoid "one account per department" — it saves configuration effort but throws away traceability and isolation. Not worth it.

Step two: configure roles and permissions​

Once accounts exist, focus on roles and permissions. In YingClaw this usually splits into three parts:

Create roles. Create roles by team function, each with a written description of its scope and default behavior. For example: "Sales assistant: handles customer list organization, follow-up reminders, sales data summaries; outputs in tables."

Assign members. Attach members to the corresponding roles. One person covering multiple functions gets multiple roles; permissions travel with the role, so a job change just means adjusting role assignments, not touching accounts.

Set permission boundaries. Define each role's data scope (which folders, spreadsheets, websites it may access) and action scope (which operations run automatically, which require confirmation). For deletion, sending and external publishing, always produce a result for human confirmation first.

The principle is "conservative by default, relax on demand": start tight, and loosen what proves safe during actual use.

Step three: share resources and knowledge​

The second value of multi-account collaboration is resource sharing. Assets the team uses frequently get built once and reused by everyone:

  • Shared skills — package common automation flows into skills like "weekly report generation," "customer follow-up reminders," "data cleaning," callable by every team member
  • Shared knowledge base — put company terminology, standards and templates into shared memory or a knowledge base, so the digital employee follows one standard no matter who assigns the work
  • Shared instruction templates — turn excellent instructions for frequent tasks into templates; newcomers produce qualified output just by following them

The key to shared resources is clear ownership and an updater: assign a responsible person to each shared skill/template so content stays maintained and doesn't rot.

Step four: configure notifications and task ownership​

The final link in team collaboration is result delivery and task traceability:

  • Notification channels — route notifications by role. Sales assistant results go to the sales group, admin assistant's to the admin group, avoiding group-wide spam. Push to WeChat, DingTalk, Feishu or other IM tools the team already uses
  • Task ownership — when assigning work, mark "whose task, which role, based on what materials"; execution records are kept automatically and queryable anytime
  • Review cadence — periodically revisit task records; identify which task types fail often and which roles need tuning, optimizing based on records rather than gut feeling

Scenario: a sales + marketing + admin trio​

Let's tie it all together with a concrete example. Say the team is three people: sales rep Zhang, marketer Li, admin Wang.

Account planning — each person gets an account. Zhang is assigned the "sales assistant" role, Li "market analysis," Wang "administrative affairs."

Permission boundaries — the sales assistant may only access the customer data folder; market analysis only marketing assets and campaign data; administrative affairs only attendance and office documents. All external sending requires human confirmation.

Shared resources — make "weekly report template" and "industry news monitoring" shared skills all three can call. Each produces their own weekly data, but in a unified format.

Notification routing — Zhang's customer follow-up reports push to the sales group, Li's competitive updates to the marketing group, Wang's attendance summaries to the management group. When leadership needs data, they go straight to the right person — and the records are there to back it up.

Can one person have multiple roles?​

Yes. Many members wear multiple hats — a marketing specialist handling both competitor monitoring and content publishing, for example, gets "market analysis" + "content operations" roles. Roles stay independently configured with isolated memory, and tasks are attributed by role without interference.

How do you onboard a new teammate?​

Three steps: create the account → assign roles (based on job responsibilities) → unlock the relevant shared skills. Within ten minutes a newcomer is using digital employees productively, following team-wide standards with assured output quality.

The pre-launch checklist​

After configuring, quickly verify with this checklist:

  1. Account isolation — are each person's task records and memory interference-free?
  2. Complete roles — does every team function have a role, and are members attached to the right ones?
  3. Clear permissions — are each role's data scope and action scope declared, with confirmation on sensitive operations?
  4. Sharable resources — are high-frequency skills and templates pooled and directly callable?
  5. No notification spam — are notifications routed by role so each group only receives what it cares about?
  6. Traceable tasks — can every task be traced to its requester, role and execution record?

One platform, a whole team under control​

The point of multi-account configuration is upgrading YingClaw from "one person's productivity tool" to "the team's collaboration hub." The philosophy at Yingying Zhineng is that AI should genuinely do work for the team — and team-level AI adoption always starts with basics like accounts and permissions.

Follow the steps above — account system, roles and permissions, shared resources, notification routing — and everyone on the team can use digital employees with confidence while the admin keeps full control. That's the value of an AI agent platform like YingClaw: the platform provides capability, configuration provides order, and the team delivers productivity.