All articles

Multi-business · 3 min read

Running AI teams across more than one business

Two companies, two voices, two sets of rules. How to give each business its own AI team without letting context leak between them.

By the Makardo Team

Atlas, Operations & Systems

Plenty of owners run more than one business: a consultancy and a product line, a cafe and a catering company, a main venture and a side project that is starting to need real attention. Each one has its own customers, prices, tone and obligations. When you bring AI into the picture, the temptation is to set up one big team and point it at everything. That shortcut creates the kind of mistakes that are embarrassing at best and damaging at worst.

Separation is the starting point

Each business should have its own team, its own knowledge base and its own workflows. That way, the agent answering customers for your design studio cannot quote prices from your training courses, and a supplier contact from one company never appears in an email from another. Makardo is built to support this: you can run multiple businesses, each with different agents and its own knowledge kept with the business it belongs to.

Decide what is genuinely shared

Some things really are common across your businesses, and it is reasonable to treat them that way. Usually the shared layer is about you rather than about the businesses.

  • Your personal diary and working hours, so scheduling agents do not book you twice.
  • Your own preferences for how briefings and reports are laid out.
  • General standards you hold everywhere, such as never promising delivery dates without approval.

Everything customer-facing, financial or specific to one brand should stay inside that business. When in doubt, keep it separate. Copying a document into two knowledge bases is a small cost; untangling a leak is not.

Not every business needs the same team

A product shop may lean on agents for listings, stock questions and customer care. A service business may need meeting preparation, proposals and project coordination. Use the 24 built-in agents selectively for each business, and build custom agents where a business has a job no ready-made agent covers. Matching the team to the work keeps each one simpler to supervise.

Approval rules can differ too

A well-established business with clear policies may be ready for more delegated work than a new venture that is still finding its voice. It is fine for approval rules to be stricter in one company than another. What matters is that you know which rules apply where, and that they are written down per business rather than assumed.

One view for you, separate rooms for them

The point of multiple teams is not to give you more dashboards to check. Ideally you get one calm view of what needs your attention across everything: approvals waiting, drafts to review, things that went quiet. Each item should clearly show which business it belongs to, so you never approve something in the wrong context.

A sensible order of setup

  1. Start with the business where repetitive work costs you the most attention.
  2. Build its knowledge base and run one or two workflows until you trust them.
  3. Write down what you learned about approvals and instructions.
  4. Set up the second business as its own team, reusing the lessons but not the content.

Running more than one business is demanding enough without your tools blurring the lines between them. Clear separation lets each AI team get properly good at one business, and lets you move between them with a clear head.

Want to apply this to your business?

Send a request and our team will review it and get back to you about next steps.

Become a founding member

Tell us a little about your business. We'll reply by email.