Inspiration
Enterprise business, marketing, and finance teams drown in disconnected dashboards while decisions stall in chat threads and email chains. AI can summarize data, but knowing what to do next takes organizational context: what campaigns are live, which channel is leaking budget, what was already decided last week, and who owns execution in Jira and Confluence.
We built ACME Decisions Layer & Command Center to bridge live cross-channel business intelligence with Atlassian Forge, Rovo Agents, Atlassian MCP, and Teamwork Graph — turning raw organizational signals into one-click Approve / Hold decisions that automatically move work forward inside Jira and Confluence.
What it does
ACME Decisions Layer & Command Center unifies 6 live signal streams (Brandwatch, Meta / Instagram, Google Keyword Tool, TikTok, Press & PR Monitor, and TV & Radio Logs) into an executive decision workspace integrated with Atlassian:
- Decisions Layer (
We propose. You approve. One click each.):- Surfaces high-confidence recommendation cards (Budget Shift — 88% CONF, Creative Push — 91% CONF, Regional — 84% CONF, Journey Fix — 79% CONF) with quantified projected impact (
+6% approvals · ~310 cards,+1.9M USD projected AUM,Save 18K USD / month). - Clicking APPROVE invokes a Forge Rovo Action (
rovo:action) that automatically creates the linked Jira execution issue, assigns the channel owner, and logs the decision into a Confluence Decision Page and Atlassian Teamwork Graph. - Clicking HOLD queues the recommendation into the weekly Confluence review agenda.
- Surfaces high-confidence recommendation cards (Budget Shift — 88% CONF, Creative Push — 91% CONF, Regional — 84% CONF, Journey Fix — 79% CONF) with quantified projected impact (
- Live Command Center (
Every signal. Every rial.):- Real-time cross-channel telemetry tracking Approved Cardholders (12,000), ROE (13.8%), New Customers (52,000), NPL (2.8%), Share of Voice (32.4% #1 in market), and Search Velocity (+318%).
- Associative Decision Log & Bilingual Voice Agent (
EN / عربي):- Tracks every historical decision alongside its downstream outcome in Teamwork Graph so teams and Rovo Agents always know what has already been decided and why.
- Forge MCP Server:
- Exposes live campaign telemetry, Share-of-Voice anomalies, and the Decision Log over Atlassian MCP so Rovo and external AI agents can query real-time business context before proposing actions.
How we built it
- Atlassian Platform & Forge Integration: Built around Atlassian Forge (
manifest.ymlwithrovo:agent,rovo:action, Custom UI panel for Jira & Confluence, Teamwork Graph connector, and Forge MCP server endpoint). - Frontend & Decision Engine: Responsive bilingual (English / Arabic RTL) executive workspace with Day / Night theme parity, live signal tape, command-palette search (
Cmd+K), and 7 modular intelligence views (Decisions Layer, Command Center, Executive Overview, Campaigns, Consumer Intel, Brand Watch, and Insights Overview). - AI & Context Routing: Multi-signal anomaly detection and confidence scoring paired with a bilingual Arabic/English Voice & Decision Assistant.
Challenges we ran into
- Connecting External Telemetry to Organizational Decisions: Raw marketing and financial metrics lack ownership context until they are linked to the people, Jira epics, and past Confluence decisions that govern them. Designing the schema mapping between recommendation cards and Teamwork Graph entities ensured every approved card produces a traceable execution artifact.
- Zero-Noise Executive UX: Business leaders need a decisive signal-to-action ratio rather than 50 charts. Structuring the Decisions Layer around four prioritized cards with explicit confidence scores, projected ROI, and a persistent Decision Log kept the workflow fast and auditable.
Accomplishments that we're proud of
- Delivering a complete 7-view Executive & Marketing Intelligence Suite (Decisions Layer, Command Center, Executive Overview, Campaigns, Consumer Intel, Brand Watch, and Insights Overview) with sub-second interactive navigation and bilingual English / Arabic support.
- Closing the loop between signal detection, executive approval, and Jira / Confluence / Rovo execution so approved decisions never get lost between dashboards and project management tools.
What we learned
- AI recommendations only drive real business velocity when paired with historical decision memory (knowing what was already approved or held last week) and one-click workflow execution inside the tools teams already use to track work.
What's next for ACME Decisions Layer & Command Center
- Expanding bidirectional Teamwork Graph outcome tracking to automatically measure actual 7-day and 30-day ROI against each approved decision card's projected impact inside Jira and Confluence.
Built With
- ai-agents
- atlassian-forge
- confluence
- jira
- mcp
- nextjs
- react
- rovo
- typescript
- vercel