The case for a governed intelligence layer between language models and enterprise data. Written for the people being asked to approve it.
Jason Pugh · Chief Executive Officer and co-founder
Rayson Technologies, LLC · Huntsville, Alabama
12 sections · 4 figures · 25 references
Twelve sections. Use it on a vendor, or on the layer you are building yourself.
Accuracy does not degrade gradually as data gets more realistic. It falls off a cliff.
Five conditions drive the collapse, and only one of them gets better with a larger model: schema scale, encoded and dirty values, definitions that live outside the database, dialect variance, and no training signal for a warehouse that has never been public.
Supplying per-question domain hints moved one frontier model from roughly 35 to roughly 55 percent. That gain is the measured value of knowledge nobody wrote down.
Every error below produces a query that runs cleanly and returns a number that looks reasonable. This is the failure mode that should decide the architecture.
A fact table joined to a many-to-many bridge without deduplication multiplies rows and inflates every sum. Revenue doubles. Nothing errors.
Summing a monthly snapshot column across daily rows, or summing a pre-aggregated measure at the wrong level.
Omitting a soft-delete flag, a test-account exclusion, or a fiscal calendar restriction.
The mean of per-region averages rather than a weighted mean, which quietly reweights the entire result.
Joining a type-2 dimension without effective-date predicates, so historical figures shift every time a record changes.
Bucketing by UTC when the business reports on local fiscal periods, which moves revenue across period boundaries.
None of these are exotic. They are the ordinary hazards of analytical SQL, and they are precisely the hazards a data team spends years encoding into curated models so nobody has to rediscover them. Direct access discards that work on every query.
An integration standard solves interoperability, and interoperability is not trust. Four of the paper's eight concerns:
senti is Rayson's implementation of the governed layer this argues for. It holds the semantic contract, compiles and validates SQL before execution, executes read-only, and files the lineage for every request. 2
Ask them of a vendor, including us. Each one should have a demonstrable answer.
Ask the same question ten times. Is the generated SQL byte-identical each time?
Ask something it cannot answer from governed definitions. Does it refuse and explain, or does it guess?
Produce the lineage record for a single answer. Does it include identity, contract version, compiled SQL, and row count?
Show a case where the layer was wrong. A vendor that cannot produce one has not looked.
The remaining eight are in section 9 of the paper, together with the five-phase sequence for standing the layer up one domain at a time.
Two ways to start, both of them short.
One metric you cannot currently defend, run against your own warehouse. You keep the trace: the question, the SQL, the result, and the record.
One hour. Your architecture walked against the twelve questions, with a written note on where the evidence is thin.
1 Published accuracy figures are point-in-time; leaderboards move. The paper carries all 25 references with dates and links. 2 The paper describes architectural patterns and cites publicly available research and guidance. It is not legal advice.