Magic LibraryWhat this dashboard answers
- How much money and time has AI saved us so far this year, and what is that worth in headcount?
- What’s our ROI multiple — how many dollars saved per dollar spent on Thread?
- Is performance improving month over month, or flattening?
- Which AI features are driving the savings, and which moved backwards last month?
- Where do the numbers sit at the board and feature level when someone asks us to prove one?
Filters and time windows
This dashboard is organized around fixed windows rather than a filter bar:What’s on the dashboard
Year-to-date summary
A single scorecard covering the year so far: total estimated dollar savings, hours saved, FTE equivalent capacity freed, monthly recurring revenue, ticket volume, AI coverage, triage deflection rate, ROI multiple, voice call volume, and Messenger deployments. This is the section to screenshot for a QBR opening slide.Month-over-month comparison
The same headline measures trended across each month of the year — savings, hours saved, ticket volume, AI coverage, deflection rate, response times, sentiment, CSAT, and after-hours coverage — so you can see the direction of travel rather than a single snapshot.Month-over-month feature comparison
Breaks savings, session volume, and FTE-equivalent time saved down by individual AI feature — categorization, triage, prioritization, and so on — comparing the current month against the prior month and flagging whether each feature moved up or down. Use this when overall savings look flat: it’s usually one feature slipping while others hold.Insights
A detailed table that puts the ROI figures (dollar savings, hours saved) next to the service-quality figures (AI coverage, triage deflection, response times, sentiment, ticket volume) at the month × board × feature grain. This is where you go to answer “which board is that number coming from?”Metric definitions
Things to watch
Dollar savings are an estimate, not an invoice. They’re derived from AI activity valued at a standard fully-loaded technician cost benchmark — not from your billing, and not from logged time entries. Present them as modeled savings.
- FTE equivalent is capacity, not headcount removed. It answers “how much technician time did we get back”, which most MSPs redeploy rather than cut.
- A flat savings line rarely means AI stopped working. Check the feature comparison — savings usually shift between features before they fall overall.
- Response-time and sentiment figures here are service-quality context, deliberately sitting next to the savings so a QBR doesn’t tell a cost story without a quality story.
- Data refreshes daily, so the current month is always partial. Compare complete months when you’re trending.
- Numbers should track your PSA but can differ by roughly a day. See Things to know & gotchas.
Next steps
Automation ROI Report
Turn these figures into a written ROI narrative.
Outcomes and analytics
The leadership view of what to measure and when.