How does Mosaic’s AI handle our data?
Mosaic, an AI project and resource planning tool, doesn’t train AI models on your firm’s data, and gives its AI no more access than your firm’s permissions allow.
Does Mosaic’s AI train on our project data?
No: the AI models Mosaic uses aren’t trained on your firm’s data, and under Mosaic’s plans with its model providers, they can’t be. Mosaic uses several AI models, including the ones behind ChatGPT and Claude, and specific models can be selected for a firm, for example to meet government requirements. Before signing, a buyer can check any AI planning tool the same way: ask in writing whether customer data trains any model, which AI providers process it, and whether AI access follows the permissions people already have.
For certifications and how Mosaic protects customer data, see Security at Mosaic.
What can Chello see and do?
Chello sees and changes only what your firm’s permissions allow, and its work is the upkeep of the plan, so the plan matches what is actually happening across projects and people. It brings the decisions that need a person to the person who makes them, so people stay in charge of the calls that matter.
What Chello can see
The plan data your firm’s permissions open to it, and nothing those permissions keep locked.
What Chello can do
Keep the plan current on its own, within the permissions your firm sets.
What goes to a person
The decisions that need a person, brought to the person who makes them.
Is it safe to give AI agents access to budgets and bill rates?
Agent access to budgets can be safe when the agent reaches only what the person it acts for can reach. Mosaic’s MCP server, the connection that lets an AI assistant such as Claude or ChatGPT work with Mosaic’s data, gives each assistant the access of the person who connected it. In Mosaic, access defaults to locked, and planning gets gated.
An assistant connected by a project manager sees what that project manager sees.
How should a firm limit which staffing changes an AI agent makes on its own?
Hold the agent to permissions, the same way the firm holds its people, rather than relying on it to judge its own limits. Chello keeps the plan current on its own, within the permissions your firm sets, and brings the decisions that need a person to the person who makes them. AI assistants connected through Mosaic’s MCP server act within the permissions of the person who connected them.
What an AI planning assistant does · Planning from Microsoft 365 and AI assistants
What else do firms ask about Mosaic’s AI and their data?
Does Chello change the plan on its own?
Yes, within the permissions your firm sets. Chello keeps the plan current on its own and brings the decisions that need a person to the person who makes them.
Can Chello be wrong?
Yes. Like any AI, Chello can misread a question, work from the wrong context or make a mistake. It answers from the live plan, works within the permissions your firm sets, and brings the decisions that need a person to the person who makes them.
What can an AI assistant connected to Mosaic see?
Only what the person who connected it can see. Mosaic’s MCP server is the connection that lets an AI assistant such as Claude or ChatGPT work with Mosaic’s data, and assistants connected through it act within the permissions of the person who connected them. They can change the plan as well as read it, and their changes go through Chello.
Which AI models does Mosaic use?
Several, including the ones behind ChatGPT and Claude. Specific models can be selected for a firm, for example to meet government requirements, and none of them is trained on your firm’s data.
Is Mosaic SOC 2 Type II?
Yes. Mosaic is SOC 2 Type II, and the report is on Mosaic’s Trust page, trust.mosaicapp.com, for customers and firms evaluating Mosaic.
Book a 20-minute demo
See how permissions shape what Chello and each AI assistant can see and do in Mosaic.