Feature
AI Assistance
AI assistance in EShopSet (prototype): summaries, drafts, and proposed changes that help teams communicate and execute faster.
Overview
EShopSet includes assistive AI surfaces to help teams move faster with clearer communication.
- Context-aware across organizations, projects, and stores.
- Assistive by default—review and approve before anything changes.
- Connected to the platform via MCP tooling, so it can propose and perform real actions.
Portal: draft + Q&A
In the customer portal, AI helps stakeholders write better requests and get quick answers without needing to learn your internal process.
- Drafts a request from a short description.
- Asks clarifying questions based on the current project/store.
- Answers basic “what’s the status / what’s next” questions using current context.
Outcome: fewer back-and-forth messages, clearer intake, and faster start.
Grounded context
AI responses are grounded in the page you’re on—so the same question gets a more accurate answer when it knows the current project, tasks, artifacts, and store connections.
- Project: tasks, owners, blockers, timeline.
- Store: connections, environments, backups, operational signals.
- Artifacts: docs, files, and run outputs.
Agency chat
In the agency workspace, the assistant can answer questions across a broader scope—customers, projects, and stores—while staying permission-aware.
- Summarize activity, changes, and conversations.
- Answer “where are we on this” across many customers.
- Help draft customer updates with the right details.
Recommendations
When a request comes in, AI can recommend what to use next.
- Assets: scripts, templates, checklists, reusable deliverables.
- Apps: workflow tools to execute common operations reliably.
MCP actions
The assistant communicates with EShopSet through MCP tools. That means it can do more than chat—it can help execute work.
- Propose actions and show what will change before execution.
- Perform a wide range of platform operations (scoped by permissions).
- Keep an audit trail by attaching outputs to runs, tasks, or artifacts.
AI is designed to keep teams in control: review first, then run.
