Markhub

Hand work to an AI agent

Ask the way you would ask a colleague, and the agent does the work.

This guide isn't translated yet. Showing the English version.

محتويات الصفحة

Markhub’s AI agents read the context of the conversation and take work off your hands. No separate tool to open, no background to re-explain: you hand the job over inside the conversation you are already having.

What you can hand over

Agents cover everything from recurring admin to document analysis and research.

Recurring work

  • Managing to-dos
  • Managing schedules
  • Tracking inventory
  • Checking the status of work
  • Writing regular reports

Document analysis

Upload Excel, PDF, Word and other files and the agent reads them and pulls out what you need.

  • Compare and summarise several Excel files
  • Distil a PDF report down to its core
  • Extract the key clauses from a contract
  • Analyse sales data and draw out the insight

Research

Agents research using the internet together with your own material.

  • Market research
  • Competitor analysis
  • Technical research
  • Gathering and comparing sources

Supporting the team

Agents carry work forward from the context of the conversation.

  • Summarising a meeting
  • Sorting out the follow-up work
  • Drafting a plan
  • Writing and revising documents
  • Organising ideas

Asking an agent

Type / in the composer and the list of available agents opens. Choose one and ask the way you normally would.

① Typing / switches into asking an agent. ② Write the request after it. Press ESC to back out.
The composer switched into agent mode by typing /
/MAKi Summarise today’s meeting and create the follow-up to-dos.
/Task Manager List only the work that has to be finished this week.

The agent does the work and shares its progress and results back into the same conversation.

Working with several agents

One request can involve more than one agent. Each agent’s progress and results continue in the same thread.

Have the market research agent analyse the competitors, and the planning agent write a proposal from that.

Role-based agents

Split agents by role and you can call the right one where it is needed. Agents with a clear remit, like Task Manager, Inventory Manager, GTM Manager or Infra Manager, behave less like a chatbot and more like a teammate doing the job.

The context comes with it

Agents work from the current hub and conversation. You do not repeat the background or re-gather the material, and the results stay as record attached to the conversation.

هل كان هذا مفيدًا؟