Gradial’s global mention system lets users type @ in any rich text surface — chat, tasks, artifact comments, and skills — to insert a durable, clickable reference to a person or platform entity. Reviews happen in context, Grady receives mention context automatically, and handoffs that used to require a Slack loop can happen in a single prompt.
Mentions must be explicitly enabled for your organization. Contact your Gradial team to get access.
What You Can Mention
The @ picker searches across seven entity types:
| Entity | What it references |
|---|
| People | Teammates in your Gradial organization |
| Workspaces | Project workspaces in Gradial |
| Tasks | Existing tasks — useful for reusing structure or passing context to Grady |
| Integrations | Connected systems like AEM, Marketo, or Salesforce Marketing Cloud |
| Environments | Specific environment configurations within an integration |
| Rules | Org-level rules governing agent behavior |
| Skills | Reusable agent skill definitions |
How It Works
Typing a mention
Type @ anywhere in a supported surface to open the picker. Start typing to filter by name — results are grouped by entity type. Use arrow keys to navigate, Enter or Tab to select, and Escape to dismiss.
When you open the picker with no query, your recently used entities appear first so you don’t have to retype common references.
How mentions look
Mentions render as rich interactive pills — styled badges with entity icons and links. They are not plain blue text. Deep-links for environments and integrations open inline.
How mentions stay accurate
Mentions reference entities by their internal ID, not their display name. If someone renames a workspace or task, the mention label updates automatically at render time — the link stays intact.
Grady context hydration
When you mention a task, workspace, or integration in a task request, Grady automatically receives that entity’s structured context before it starts working. You don’t need to paste URLs or re-explain what you’re referring to.
Use Cases
Collaboration
Mentions make Gradial the place where review, approval, and handoffs happen in context — without forcing anyone to copy task links into Slack.
Ask for review or approval
“@Maya please review the updated hero copy on @Nordstrom Holiday Campaign Page before launch.”
The reviewer is pulled into the exact work item with the page and task context already attached.
Debug with a teammate
“@Alex, can you help debug why this task failed? It looks like the AEM update succeeded but the email asset did not refresh. It’s using @AEM Sandbox integration.”
The person, task, and failure context are all connected in one place — no separate “send me the task link” loop.
Pull the right people into a workflow
“Bring @Maya, @Priya, and @Jordan into @Nordstrom Holiday Campaign Launch. Maya should review visuals, Priya should approve copy, and Jordan should give final launch approval.”
Cross-functional handoffs become explicit without a separate coordination thread.
Execution Context
Mentions tell Grady what prior work, decisions, chats, or tasks should guide the next piece of work.
Use a task as source-of-truth guidance
“Use @Nordstrom Holiday Campaign Guidance for every follow-up task in this campaign. Make sure product images are not repeated, sale language stays consistent, and CTAs match the campaign pattern.”
Campaign-level consistency without repeating the same guidance in every prompt.
Reference a task as a content pattern
“Reference @Anniversary Sale Page Task and create a new page that follows the same structure for the Kids campaign.”
Grady reuses a proven structure instead of starting from scratch.
Carry forward decisions from a chat
“Before making changes, read @Q4 Launch Planning Chat and preserve the decisions about image rotation, headline tone, and CTA placement.”
Decisions made in prior conversations become reusable execution context.
Governance
Mentions make brand, legal, compliance, and policy requirements explicit at the moment of work.
Apply rules directly in the prompt
“Rewrite this email using @FINRA Copy Rules and @No Guaranteed Returns Rule.”
The right constraints go in the prompt — reducing back-and-forth review cycles.
Convert feedback into governed changes
“Address @Priya’s feedback on @Holiday Landing Page Review, but make sure all changes still follow @Premium Brand Voice Rules.”
Feedback loops stay grounded in the right rules, not just individual preferences.
Systems
Mentions tell Grady which connected systems, integrations, or operating environments to inspect or act through.
Use an environment as an operating bundle
“Use @North America Campaign Ops and @Launch QA Skill to prepare this campaign for publishing.”
The environment tells Grady which connected systems are relevant for the work.
Debug across a system bundle
“Within @Campaign Ops Environment, check whether the lead routing issue is coming from the CMS form, @Marketo, or the ticket workflow.”
The environment provides the system map; individual integration mentions narrow the investigation.
Prepare a campaign across connected tools
“Use @Campaign Ops Environment to prepare @Nordstrom Holiday Campaign Page for launch. Confirm the AEM page is updated, the Marketo email references the right assets, and the ticket workflow is ready for approval.”
Multi-system work in one natural language instruction.
Skills
Mentions let users tell Grady which playbook, capability, or workflow mode to apply.
Use a skill as the playbook
“Use @AEM Page Editing Skill on the work from @Product Launch Task.”
Tell Grady exactly which workflow or capability should guide the task.
Combine a skill with prior task context
“Use @AEM Page Editing Skill and reference @Anniversary Sale Page Task to build a new Kids campaign page with the same structure.”
The skill tells Grady how to operate; the task tells Grady what pattern to follow.
Run a QA skill on demand
“Run @Launch QA Skill on @Nordstrom Holiday Campaign Page and summarize anything that needs to be fixed before publishing.”
Repeatable review workflows become explicit, reusable one-line actions.
Where Mentions Work
| Surface | Supported |
|---|
| Chat composer | Yes |
| Task / ticket editor | Yes |
| Artifact comments | Yes (including dark-mode styling) |
| Skills | Yes |
| Plain text fields | No — mention pills require a rich text editor |
Configuration
Enabling mentions
- Contact your Gradial team to enable the
global-mentions-picker feature flag for your organization.
- The picker activates immediately across all supported surfaces — no per-surface configuration needed.
Permissions
Available to all Gradial roles with organization access. No additional role restrictions apply.
Limitations
- Artifact mentions not yet supported — the ability to @ an artifact directly in the picker ships in a future release.
- External ticket mentions not yet supported — Jira, Asana, Workfront, and Wrike mentions are later scope.
- No mention dates in V1 — @today / @tomorrow style date mentions are designed but not yet shipped.
- No backlinks or mention analytics — a backlink graph and mention history are on the roadmap.
