Agents, Reusable Skills, Scheduled Automations + Brand Dashboard

I would like to see a proper operational Agent layer under every Brand or Workspace.

Remalt already supports AI workflows, but it would become considerably more powerful if Agents could have reusable Skills and run automatically on daily, weekly, monthly or custom schedules.

Persistent Agents

Each Brand should be able to create its own team of Agents, for example:

⦿ Research Agent
⦿ Market Intelligence Agent
⦿ Content Strategist
⦿ Copywriter
⦿ SEO Agent
⦿ Social Media Agent
⦿ Product Strategist
⦿ Developer Agent
⦿ Analyst
⦿ Custom MCP Agent

Each Agent should have a defined role, model, tools, permissions and access to the Brand Knowledge Brain.

Reusable Agent Skills

Skills should be reusable capabilities that can be assigned to one or several Agents.

Examples:

⦿ Competitor research
⦿ Deep web research
⦿ Trend monitoring
⦿ Customer pain-point analysis
⦿ LinkedIn content creation
⦿ Newsletter research
⦿ SEO analysis
⦿ Startup validation
⦿ Fundraising analysis
⦿ Campaign planning
⦿ Report generation
⦿ Code review

Each Skill could contain:

⦿ Instructions
⦿ Preferred model
⦿ MCP tools
⦿ Knowledge sources
⦿ Output format
⦿ Quality criteria
⦿ Permissions
⦿ Fallback model

This would be much easier to maintain than rebuilding large prompts inside multiple workflows.

Scheduled automations

Agents should be able to run automatically:

⦿ Daily
⦿ Weekdays
⦿ Weekly
⦿ Monthly
⦿ Custom days and times
⦿ Custom time zone
⦿ Webhook triggered
⦿ Event triggered
⦿ Manual trigger

Examples:

Every morning:
Monitor competitors, relevant industry news and emerging signals, then update the Brand Knowledge Brain.

Every Monday:
Create a weekly trend and content opportunity report.

Every Friday:
Analyse content performance and recommend what should be repeated, improved or stopped.

Monthly:
Create a market intelligence and competitor strategy report.

Human approval controls

Every automation should allow different levels of autonomy:

⦿ Fully automatic
⦿ Generate and request approval
⦿ Draft only
⦿ Notify only

A workflow could therefore be:

Research
→ Analyse
→ Generate output
→ Request approval
→ Publish, send or execute

Brand Dashboard

Every Brand should have an operational dashboard showing:

⦿ AI Team
⦿ Active automations
⦿ Scheduled tasks
⦿ Recent Agent runs
⦿ Pending approvals
⦿ Research signals
⦿ Knowledge updates
⦿ Content pipeline
⦿ Performance
⦿ API / coin usage
⦿ Failed runs and errors

For each automation:

⦿ Last run
⦿ Next run
⦿ Status
⦿ Agent responsible
⦿ Model used
⦿ Tools used
⦿ Knowledge accessed
⦿ Cost
⦿ Output
⦿ Run history
⦿ Errors

The structure could be:

Brand
→ Knowledge Brain
→ AI Team
→ Agent Skills
→ MCP Tools
→ Automations
→ Brainboards
→ Dashboard

A useful distinction would be:

Agent = who performs the work
Skill = how the work is performed
Automation = when and under what conditions it runs

That would make Remalt much easier to scale across multiple Brands without duplicating the same setup repeatedly.