Platform · the Effort Graph
Lumopath reads the tools your team already uses and extracts what was done, for whom, by whom, how long it took, and what it drove. Including what didn't happen.
Measured, not assumed
Most signals capture what someone scheduled, logged or was assigned. The Effort Graph looks at the work itself — across every tool, all day — and only then decides what happened and how long it took.
Algorithms assign a specific amount of time to each action, based on the work itself.
A block of time that may or may not have gone to the QBR.
Three bursts of edits to the deck, mostly outside the block. The algorithm turns the edit stream into active working time.
An activity only counts once several independent signals agree.
Depends on someone remembering to log it — and logging it the same way every time.
Capacity comes from each person's full day of activity, not the size of their book.
Same book, very different weeks: Maya's accounts generate constant escalations across email, chat and tickets; Jordan's run smoothly.
Under every number
Nobody logs anything — the Effort Graph reads the work itself. Getting from there to a trustworthy number takes four stages of engineering, and each one has to hold up every day, for every team.
Email, chat, calendar, ticketing, CRM, docs and project tools — connected once, then kept current through outages and silent API changes.
The same person appears under different IDs in every tool. Get this wrong and every number quietly double-counts or misses people.
"Meeting time" and "focus time" sound simple. Each hides a dozen ways to be wrong.
Every KPI is produced by a pipeline that runs in strict order, every day, and is watched while it runs.
Four years of engineering
Consistency by design
A model calculating on the spot gives a slightly different answer each time — fine for a chat, not for a board deck, a headcount plan or a performance review. So every number in the Effort Graph is computed by the pipeline, stored and versioned. The AI explains it. It never does the math.
Four years of edge cases
A failure that throws an error is easy to fix. The ones that matter look fine. Every definition in the Effort Graph has been tuned against real teams until it matched how the work actually felt.
Nothing was broken — the raw data honestly looked like seven and a half hours. Then someone who knew the work said "that can't be right." We found the cause, corrected the definition, and checked the same pattern everywhere else it could appear.
A model can read data; it can't tell when the data is wrong. That loop — people who know the work, a definition that changes, history re-applied so numbers still line up — is built into every KPI.
Quiet failures we've caught — and now check for automatically
A one-line timezone default blanked a metric for everyone west of Eastern time, every afternoon — no error, just "—".
Metrics that return nothing are treated as failures, not displayed as empty.
A weekly briefing said a rep owned 191 accounts; the dashboard said 8. Two pieces of code counted differently.
The same metric computed in two places is reconciled on every run.
A sync showed green while millions of updated rows sat behind a stuck checkpoint.
Freshness is measured on the data itself, not on the job's status.
"Meeting time" counted vacation days, because two code paths disagreed on what to exclude.
Exclusion rules are validated identically across every code path.
Trust
The Effort Graph is designed so your team only ever sees finished, role-scoped metrics.
Seven tiers of access control, so managers only ever see their own org.
Strict per-customer data isolation, enforced in every single query.
No screen capture, no keystroke tracking, no browsing history and no private messages. Most metrics come from who, when and for which account.
Independently audited controls, audited admin access and a verified deletion flow.
What it powers
The Effort Graph is the shared record the rest of Lumopath reads from. Because everything draws on the same definitions, the number on the dashboard, the trigger on a play and the answer from an agent always agree.
Coverage, capacity, cost to serve, responsiveness and account health — per person, team and account, every day, with previous-period comparison built in.
Explore the KPI Catalog → PlaysWhen the graph sees a pattern — an account gone quiet, a review overdue, adoption stalling — the right play fires with its context attached, and its outcome is measured against NRR.
See the Plays → Build with LumopathYour RevOps and data teams keep the logic and the interface. Every KPI and play is available over MCP, so whatever you build inherits the same trusted numbers.
Build with Lumopath →A 30-minute review of where your team's effort is going, and what it would take to point it at retention and expansion.
