From tool chaos to marketing symphony: With marble work 2.0, AI agents and people work in the same task chains for the first time. Agents research, write, analyse and produce. People review, decide and approve. Behind it is not another chat window, but a complete workspace for projects, clients, teams, files and time.
Contents
-
1.
One task becomes a shared chain
-
2.
Agents instead of one-off prompts
-
3.
Work that starts by itself
-
4.
Text, images, video, audio and PDFs
-
5.
Not just AI. Complete project management.
-
6.
The tools your team uses become part of the workflow
-
7.
Your own AI keys. No artificial credits.
-
8.
Productive AI without losing control
-
9.
For teams ready to move AI beyond experiments
The core
One task becomes a shared chain
In marble work 2.0, a task is no longer just a title and one responsible person. It can contain any number of steps. For every step, the team decides whether an AI agent or a person takes responsibility.
One agent researches a topic and gathers reliable sources. A second turns them into an article in the company’s voice. An editor reviews and approves the content. An image agent creates the cover, while another agent publishes the post.
When the chain reaches a human step, it pauses and notifies the responsible person. If a result needs revision, the task can restart precisely from that step. Earlier versions remain available.
The agent researches. You review. The agent publishes. Everything remains one traceable task.
Your AI team
Agents instead of one-off prompts
In marble work, agents are not anonymous chat windows. They are created like team members. Each agent gets a role, clear instructions, a suitable AI model, connected tools and reusable skills.
Shared company knowledge provides context about products, audiences, voice and processes. What agents learn through collaboration remains visible, editable and deletable. Valuable insights can be promoted to company knowledge.
This turns a general AI model into a specialist that understands both its task and your business.
Automated, but controlled
Work that starts by itself
Recurring work should not depend on someone remembering it at the right moment. With the new Auto-Start feature, tasks can begin daily, weekly, monthly or on a custom schedule. Alternatively, a task can start once at its due date.
A competitive report can now begin every Monday at 7 a.m.: an agent researches new developments, connects them with existing data, creates a summary in the company’s voice and turns it into a PDF with charts. At 8 a.m., the report is waiting for human approval.
Automation handles the routine. Decisions remain with the team.
Results instead of chat histories
Text, images, video, audio and PDFs
Agents deliver their work directly inside the relevant task step. Text is displayed with formatting, files are stored permanently and generated media is captured automatically.
marble work manages more than text. Agents can create images, video and audio, analyse existing images and PDFs, and produce finished PDF documents, including charts when needed.
Files can be assigned to a project, a task or an individual step. Time-limited preview links can be created for clients and other external people.
The operational foundation
Not just AI. Complete project management.
An AI platform has limited value in daily work if projects, clients and time still have to be managed in separate systems. That is why operations are not an add-on, but the second foundation of marble work 2.0.
- Tasks in board and list views with status, owners and due dates
- Projects and clients with budgets, hourly rates and custom fields
- Subtasks, comments, files, tags and access permissions
- Project time through a timer or manual entries
- Work hours, breaks, leave, sick days and hour accounts
- Exports and branded PDF reports
- Multiple workspaces with one login
The “AI versus human” project analysis also shows which share agents completed and what remained with the human team.
Connections
The tools your team uses become part of the workflow
Agents become truly useful when they can access the tools where information and work already live. marble work connects with services including Notion, Google Workspace, Google Analytics, Search Console, Google Ads, Higgsfield, ElevenLabs, Nano Banana, Brevo and many more.
The system selects the specific capability an agent needs for each step. Credentials are stored encrypted in the vault and are never exposed to the agent or the frontend.
Through MCP and the REST API, marble work can also become part of open AI and developer workflows.
Your model, your costs
Your own AI keys. No artificial credits.
Many AI products sell credits, impose monthly ceilings and add a markup to the actual model costs. marble work takes a different approach.
Organisations connect their own OpenAI or Anthropic accounts. Model usage is billed directly by the respective provider. marble work does not impose an artificial credit limit or make money from a markup on AI costs.
Each agent can use the model that fits its task. Requests and token consumption remain transparent by model and period.
Control remains built in
Productive AI without losing control
Bringing AI into real client and business processes requires more than a convincing demo. marble work logs executions and changes, keeps result versions and explains errors clearly. Failed steps are retried automatically.
Workspaces are separated from each other in the database. AI keys and connector credentials are stored encrypted in the vault. The infrastructure is operated in the EU and hosted in accordance with the GDPR.
Human review is not an obstacle to automation. It is a built-in part of it.
Who is it for?
For teams ready to move AI beyond experiments
marble work 2.0 is built for marketing teams, agencies, small and medium-sized businesses, and solo professionals who want to accelerate recurring work with AI without piecing their operation together from separate tools.
No technical expertise is required. Agents are described in natural language rather than programmed. Guided onboarding helps set up the workspace, company knowledge, AI connection and first agent.
marble work 2.0 turns isolated AI capabilities into a shared workflow, and tool chaos into a marketing symphony.
