Momenta believes the key to AI-ready analytics lives in query history. Because Neuron reads it, we get a close look at how analytics teams actually work: the patterns that recur, how knowledge moves, and where an AI agent is finally positioned to help. These are ten readings from that work.

Momenta Analytics · The Health Assessment

Ten Vital Signs of a Query History

The most valuable thing your analytics team produces is the record of every analysis they have designed and run. It is sitting in your query history, and nothing in your stack is counting it.

That record is the definitions your team settled on, the rules they applied, the joins they worked out, and the questions someone decided were worth asking. It already exists, in full. It is written in SQL, in a log nobody reads.

dbt models the tenth of the work that gets asked on repeat. Everything asked once sits unread: the cohort a client needed by Friday, the washout window one analyst set at six months and another at twelve. That untouched majority is what these ten readings measure.

They tell you how much your team produces, how much can be recovered, how much is stuck in one person's head, and how much has earned enough trust to build on. Taken together they answer one question: is your history becoming an asset, or just getting longer? It becomes an institutional knowledge asset, a set of files your people and your machines can both read, once it is available, searchable, instructive, fresh, and comprehensive.

01 · ExtractedPulled out of one-off queries and individual heads, into view.

02 · RefinedThe real logic separated from the one-off pulls around it.

03 · PooledThe same rule merged out of forty silos into one shared source.

04 · ProvenReuse marks which definitions earned enough trust to build on.

The last stage is the point. Trust through use is where exhaust becomes gold.

You do not know your ten numbers yet. You can know all ten today, off your own query history, for free.

ExtractionWhat are you pulling into view?

01Intelligence Volume8,792 / qtr Whether your team is breaking new ground or re-running last quarter's report. The breadth here is the breadth your AI can cover. 02Analysis Rate47% How much of your query volume is real analysis, and how much is exploration and housekeeping.

RefinementIs there logic worth refining out?

03Business-Rule Density4,317 The logic your business actually runs on, living in WHERE clauses with no model or documented column anywhere. 04Hidden-Join Rate109 / 184 The share of your data model that exists only inside the queries, in no dbt model and no catalog.

PoolingIs it merged out of the silos?

05Evaporation90%+ How much reasoning leaves nothing behind, and what that costs the next analyst who needs it. 06Definition Concentration94-97% How much of your shared vocabulary lives in one person's SQL, with no second copy to check it against. 07Key-Person Risk34% How much of the team's output rides on one person's judgment, and whether that judgment is credited anywhere.

ProvingIs use proving what to trust?

08Reuse / Build-On80% How much work stands on work the team already trusts. Usage is a vote. 09Standardization0% Whether any method longer than a single query is shared across the team. 10The Feedback Loopcompounds Whether corrections carry forward, and the full definition of what an institutional knowledge asset is.

What you get

A semantic model paired with an agent you can rely on, so every metric resolves to one definition, every join is documented, and the answer comes from what your team actually trusts. Built in hours, portable to whatever stack you run next, and yours to keep.

Ten numbers describe the record your team has already built. One run on your own query history returns all ten, and the semantic model that fixes them. The other way to learn your numbers is from a client who found one of them first.

Implementing AI in your analytics team? .