Value Dashboard

The Value Dashboard is the first screen a customer sees when they log in. In one glance it answers the key question: is this worth what I'm paying for? Nine widgets surface the real return — in USD and in time — that Rela AI generates on your operation.

Executive summary

The game-changer: your dashboard no longer shows how many Gemini tokens you consumed (useful metric for us). It now shows avoided downtime, predictive interventions triggered and your real ROI vs the plan cost.

What does it do?

With per-machine pricing, tokens became an internal problem for us (now lives in the admin panel). The customer pays a fixed fee and needs a quick answer to four questions:

1. Am I using what I paid for? (Plan utilization)
2. Is the system working? (Ops health + Agent impact)
3. Where is money being left on the table? (Lost-value alerts)
4. How do I compare with the market and am I on the right plan? (Industry benchmark + Plan fit)

The Value Dashboard answers all four in under 30 seconds.

How does it work?

flowchart LR
  A[Real tenant data] --> B[Cross-collection<br/>aggregations]
  B --> C[Tenant economic<br/>benchmarks]
  C --> D[Hero ROI]
  C --> E[Lost-value alerts]
  B --> F[Plan utilization]
  B --> G[Ops health / Agent impact /<br/>Proactive vs Reactive / Inventory]
  B --> H[Industry benchmark]
  B --> I[Plan fit]
  D & E & F & G & H & I --> J[Value Dashboard]

Each widget hits a backend endpoint that aggregates cross-collection data: _assets for active machines, _machine_events for month events, _asset_kpis for OEE/MTTR/MTBF, _tenant_benchmarks for customer costs, _subscription_plans for price and limits.

USD figures are computed by multiplying physical metrics (avoided downtime hours, saved technician hours) by the economic assumptions the tenant configured in /settings/benchmarks. If they haven't been set, widgets show physical metrics only (no USD) plus a CTA to complete them.

Panel widgets

1. "Value generated this month" Hero

The most important card: above the fold, full width.

Configured mode (benchmarks ready):

• Big net savings number (total savings − plan cost).
• Delta ▲/▼ vs previous month.
• 6 metrics: total savings, plan cost, ROI multiplier, downtime avoided (h), technician hours saved, "Adjust assumptions" link.
• Current plan badge (Essential/Professional/Enterprise).

Fallback mode (no benchmarks): amber card with physical metrics (downtime avoided, technician hours) and CTA "Configure assumptions" → /settings/benchmarks. We never fabricate USD figures without real customer data.

Endpoint: GET /api/v1/roi/hero.

2. Lost-value alerts

Full-width widget right below the Hero. Turns passive data into rentable actions:

Each card has an inline CTA that takes the user to the relevant path (/assets, /parts, /tasks?filter=overdue).

Endpoint: GET /api/v1/roi/lost-value.

3. Plan utilization

Replaces the old tokens view. Shows:

• Current plan (Essential / Professional / Enterprise) + monthly price.
• Progress bars: Machines (assets), Agents (WhatsApp + Email), Interactions (month events).
• Next invoice date.

Endpoint: GET /api/v1/billing/overview (new composite service).

4. Operational health

6-metric cockpit in a 3×2 grid using canonical severity tokens:

• Active critical alarms (emergency + critical).
• Active warning alarms.
• Events in the last 24 h.
• Maintenance compliance (%).
• SLA completion (%).
• Average resolution time.

Endpoints: GET /api/v1/dashboard/kpis + GET /api/v1/dashboard/sla.

5. AI agent impact

Metrics of the core subscription value:

• Active agents vs total (WhatsApp + Email combined).
• Handled conversations (last 30 days).
• Messages handled.
• Last activity with relative time.
• WhatsApp vs Email split in the footer.

6. Proactive vs reactive

Cultural-shift narrative toward planned maintenance:

• Percent of maintenance planned on time.
• Overdue WOs.
• Tasks in progress.
• Completed tasks.
• Segmented bar with priority mix (urgent/high/medium/low).

7. Inventory + procurement

• Total inventory value (sum of unit_cost * quantity_on_hand).
• Registered parts.
• Parts below minimum (with link to /parts?filter=low).
• Out-of-stock parts (with link to /parts?filter=zero).
• Footer with open POs by status + committed value.

8. Benchmark vs industry

Compares tenant KPIs against curated baselines for 7 sectors:

Tolerance ±5 %: better / worse / on_par / no_data.

9. Am I on the right plan?

Heuristic that detects tier-gated features in use and recommends upgrade or downgrade:

• Upgrade: tenant uses prognostics, anomaly_detection, energy_monitoring or extraction but is on Essential. Or uses oee, sso, audit_trail but not on Enterprise.
• Downgrade: tenant is on Professional or Enterprise but doesn't touch the exclusive gates.
• Fit: plan matches the real usage.

Confidence: high (≥2 features), medium (1), low (downgrade).

Shared endpoint with benchmark: GET /api/v1/roi/insights.

How to use it

1. Go to home (/dashboard). It's the first screen you see.
2. Check the Hero for the month's headline number.
3. If you see amber lost-value alerts, click any of them. Each has a direct CTA to the rentable action.
4. Review Plan utilization to know if you're approaching a limit.
5. Ops Health + Proactive/Reactive summarize the physical state. If you see active critical alarms, jump to the operational inbox with the "Open operations" button.
6. Benchmark vs industry and Plan fit are contextual — review them every 2-3 months to calibrate expectations and confirm you're on the right tier.

Economic assumptions setup

To unlock the USD Hero the tenant must fill in the assumptions at /settings/benchmarks:

While unset, relevant widgets show physical metrics (hours, counts) only, and the Hero runs in amber fallback mode.

Endpoint: GET/PATCH/POST /api/v1/tenant-benchmarks (+ /reset).

Use cases

1. Maintenance supervisor, Monday 7 AM. Opens the Value Dashboard, sees the Hero with "Net savings $33,200 this month, +12% vs previous month". Scans the lost-value alerts, finds "3 degraded assets — potential $12,500", clicks and creates WOs for the week.
2. Plant manager, monthly review. Opens the panel, confirms the ROI stays above 15×. Reviews Benchmark vs Industry to confirm the 78% OEE is still above the food-sector average (65%). Exports a screenshot for the steering committee.
3. New customer onboarding. During the first week they see the Hero in amber fallback mode with physical metrics only. Friday they finish the benchmarks wizard and from Monday the Hero displays real USD.

Limitations and assumptions

• USD figures depend 100% on the tenant having set their benchmarks. Inaccurate inputs mean inaccurate ROI.
• Benchmark vs industry uses fixed public baselines (not a live sector average). Use it as directional reference, not an exact ranking.
• "Plan fit" is a heuristic based on tier-gated feature detection. It doesn't consider volume (event count, calls, etc.).
• Widgets cache per request — no websockets. A reload fetches the freshest data.
• No dedicated mobile layout yet; tablet and desktop look good.

Key benefits

• The customer understands in 30 seconds what Rela AI is giving them this month.
• Proactive alerts turn the dashboard from passive to active — every card has a direct CTA to the rentable action.
• Plan fit prevents churn from under-utilization and surfaces organic upgrades.
• The Hero with real ROI justifies annual renewal with a verifiable number.
• The Value vs Operational split respects one-responsibility-per-screen.

Endpoints summary