Skip to main content

​
What is MagOneAI Hub?

MagOneAI Hub is the consumer-facing interface of MagOneAI — a conversational AI chat interface where end users interact with agents, trigger workflows, and track execution progress. If you’re an end user (not a builder or administrator), MagOneAI Hub is your primary interface. It’s designed to feel like chatting with a knowledgeable assistant, but behind the scenes, it’s orchestrating complex AI workflows.
Access to MagOneAI Hub is available to all user roles, including End Users who cannot access MagOneAI Studio. This makes Hub perfect for employees, customers, or partners who need to use AI workflows without building them.

​
Chat interface

MagOneAI Hub provides a modern, intuitive chat experience for interacting with AI.

​
Conversational AI interface

The chat interface works like any modern messaging app:
  • Type naturally — Ask questions or give instructions in plain language
  • Instant responses — Agents respond in real-time as they process your request
  • Contextual conversations — The AI remembers previous messages in the conversation, and keeps that context even if you step away and come back later in the same conversation
  • Multi-turn interactions — Ask follow-up questions and refine requests
No need to learn query syntax or structured commands — just chat naturally.

​
Natural language interaction with agents

Behind each conversation is one or more AI agents configured with specific capabilities:
  • Persona-driven responses — Each agent has a role and tone (e.g., a customer support agent, a compliance analyst)
  • Context-aware — Agents understand what you’ve said earlier in the conversation
  • Intelligent routing — If your message relates to a specific workflow, the system routes it to the right agent

​
Rich responses with formatted output

Agents don’t just return plain text — they can produce:
  • Formatted text — Bold, italics, bullet points, numbered lists
  • Tables — Structured data (e.g., comparison tables, reports)
  • Code blocks — If the agent generates code or configuration
  • Links — Hyperlinks to resources, documents, or external systems
  • Embedded media — Images, charts, or visualizations (when supported by the workflow)
This makes responses more actionable and easier to understand.

​
File upload and attachment support

Attach files inline, ChatGPT-style: pick a file, add your message, and the file rides along with that turn. Each file you send appears as a chip on the message you attached it to, so the conversation shows exactly which files went with which request. You can provide context by uploading:
  • Document uploads — PDFs, Word docs, text files
  • Image uploads — Screenshots, diagrams, photos
  • Spreadsheets — Excel or CSV files for data processing
Agents can analyze uploaded files using OCR, document parsing, or vision models (depending on how the workflow is configured). Spreadsheets and CSVs become queryable data the agent can run calculations over, and long documents are automatically indexed so the agent can retrieve the most relevant passages. How attachments behave in a conversation:
  • They accumulate — every file you attach stays active for the rest of the conversation, and all active files are available to the agent on each new turn, not just the message they arrived on. Your most recent upload is flagged as the latest, so “analyze the file I just sent” works as expected.
  • They show extraction status — a file moves through uploading → uploaded → extracting → ready (or failed), and large documents show indexing progress before they’re fully searchable.
  • You can detach — remove a file to free a slot and stop it from being sent to the agent on later turns.
A conversation can hold up to 10 active attachments at a time, each up to 50 MB. Detach files you no longer need to make room. Total content across all attachments is also bounded by a combined input limit, so a single conversation can’t overwhelm the agent.
When uploading documents, provide context in your message. Instead of just uploading a PDF and saying “process this”, say “Here’s a contract — can you extract the key terms and flag any unusual clauses?”

​
Files the agent creates

Agents don’t only read files — they can produce them. When a workflow generates a document (a report, a filled template, a slide deck, a spreadsheet, or a PDF), it comes back as a download chip on the agent’s reply. Click it to download the file. These generated files are streamed securely on demand rather than exposed as public links, so only people with access to the conversation can retrieve them.

​
Rate a response

You can tell the platform whether a response was helpful. Each agent reply carries a thumbs up / thumbs down control, and you can add an optional short comment (up to 2,000 characters) explaining what was good or what went wrong.
  • One rating per response, per person — your rating is yours; you can change it or clear it at any time.
  • Ratings attach to the run behind the reply — feedback is tied to the specific execution that produced the answer, so builders and admins reviewing a workflow can see which runs users found helpful and read the comments.
A quick thumbs down with a one-line comment (“missed the renewal date”) is the fastest way to help builders improve the agent — far more useful than silently re-asking.

​
Running workflows from chat

One of MagOneAI Hub’s most powerful features is the ability to trigger workflows through natural language.

​
Trigger workflows via conversation

You don’t need to navigate menus or fill out forms — just describe what you want: Example:
“Can you verify this business for KYB compliance?”
The system recognizes your intent, matches it to a KYB verification workflow, and starts execution.

​
Provide inputs through natural language

Workflows often require specific inputs (company name, document URLs, parameters). Instead of filling out a form, you provide them conversationally: Example:
“Run a background check on Acme Corp. Here’s their website: acme.com”
The system extracts the inputs (company name: “Acme Corp”, website: “acme.com”) and passes them to the workflow.

​
Automatic parameter extraction

MagOneAI Hub uses NLP to extract structured data from your messages:
  • Entity recognition — Identifies names, dates, numbers, URLs
  • Intent classification — Determines which workflow you want to run
  • Slot filling — Extracts required parameters from your message
If the system needs more information, it asks follow-up questions: Example:
“I need to run a KYB check.” Which company would you like to check? “Acme Corp” Got it. Starting KYB verification for Acme Corp…

​
Execution progress tracking

MagOneAI Hub provides real-time visibility into workflow execution.

​
Real-time workflow execution visibility

Once you trigger a workflow, you see:
  • Workflow name and description — What’s running
  • Current status — Running, Completed, Failed, or Waiting for Input
  • Start time — When execution began
  • Estimated completion — Based on average execution time (if available)

​
Step-by-step progress indicators

As the workflow executes, MagOneAI Hub shows which activities are running: Example display:
This transparency helps you understand what’s happening and how long it might take.

​
Activity status updates as they complete

You receive real-time updates as each activity finishes:
Agent completed: Document analysis Found 3 potential issues requiring review.
This keeps you informed without requiring you to watch the screen continuously.
Workflows powered by Temporal are durable — If you close MagOneAI Hub during a workflow execution, the workflow continues running. When you return, you’ll see the latest status.

​
Human task handling

Some workflows require human approval or input before continuing. MagOneAI Hub makes this seamless.

​
Receive approval requests

When a workflow reaches a Human Task activity, you’re notified in MagOneAI Hub:
Approval needed: KYB Verification The compliance agent has completed its analysis and flagged this business as medium risk. Review the report and approve or reject the onboarding.

​
Review agent outputs before proceeding

Human Task activities present the agent’s output for your review:
  • Summary — What the agent found or did
  • Detailed results — Full data and reasoning
  • Recommendation — The agent’s suggested action
You review the information and make a decision.

​
Provide input when workflows require human decisions

Some workflows need your input to proceed: Example:
Input needed: Choose pricing tier Based on the customer’s usage estimate, which pricing tier should we propose?
  • Starter ($99/month)
  • Professional ($299/month)
  • Enterprise (custom pricing)
You select an option, and the workflow continues with your choice.
Human Tasks are perfect for governance, compliance reviews, and situations where human judgment is required. They ensure AI automation stays aligned with business policies.

​
Conversation history

MagOneAI Hub keeps a complete record of your interactions.

​
Full history of past conversations

Access all your previous conversations:
  • Chronological list — Most recent conversations at the top
  • Conversation titles — Auto-generated titles based on content
  • Timestamps — When each conversation started
Click any conversation to view the full history.

​
Search and filter conversations

Find past interactions quickly:
  • Search by keywords — Find conversations mentioning specific topics
  • Filter by date — Show conversations from a specific time range
  • Filter by workflow — See all conversations that triggered a specific workflow
This makes it easy to reference past results or resume previous discussions.

​
Reference previous workflow executions

Each conversation preserves:
  • Messages exchanged — What you said and what the agent responded
  • Workflow executions — Which workflows ran and their results
  • Attachments — Files you uploaded
  • Outputs — Reports, summaries, or data the agent generated
This creates an auditable record of your AI interactions.
Conversation history is user-specific — you can only see your own conversations, not those of other users (unless you’re a SuperAdmin or Org Owner reviewing audit logs).

​
Best practices for using MagOneAI Hub

Be specific in your requests — Instead of “check this company”, say “run KYB verification on Acme Corp and flag any compliance issues”. The more context you provide, the better the agent can help.
Upload supporting documents — If you have contracts, forms, or reports, upload them when you start the conversation. Agents can analyze documents to provide better answers.
Ask follow-up questions — If the agent’s response isn’t clear, ask for clarification or more details. The conversation is contextual, so the agent remembers what you’re discussing.
Review agent reasoning — When agents make recommendations, they often explain their reasoning. Read this carefully before approving actions, especially for high-stakes decisions.

​
Use cases for MagOneAI Hub

MagOneAI Hub is designed for end users across many scenarios:

Customer onboarding

End customers use Hub to submit onboarding information, upload documents, and track verification progress.

Internal support

Employees use Hub to get help from AI assistants (e.g., IT support, HR questions, policy lookups).

Document processing

Users upload documents (contracts, invoices, forms) and get summaries, extractions, or analysis.

Approval workflows

Managers use Hub to review and approve AI-generated outputs before they’re finalized.

​
Integration with other portals

MagOneAI Hub is part of a three-portal ecosystem:
  • Admin Portal — Platform administrators manage organizations, users, and security
  • MagOneAI Studio — Builders create the workflows that power Hub’s AI assistants
  • MagOneAI Hub — End users interact with deployed workflows through chat
Builders create in StudioWorkflows, agents, and tools are designed and deployed in MagOneAI Studio.
End users interact in HubOnce deployed, workflows become available as conversational assistants in MagOneAI Hub.
This separation ensures builders have the tools they need without cluttering the end-user experience.

​
Next steps

MagOneAI Studio

Learn how workflows are built in the Studio interface

Workflow builder guide

Step-by-step tutorial: build your first workflow