Zack Gregory

Independent technical project · Signal North

NorthTracepowered by Veyra AI

An independent project exploring SQL Server migration assessment, data quality and AI-assisted reconciliation.

NorthTrace, powered by Veyra AI · 25-second walkthrough

How the assessment works

Read the video summary

The scanner inventories a fictional SQL Server schema and scores migration risks. The 11-object example has a complexity index of 54 out of 100. Two procedures are high complexity. The legacy customer sync reads a linked server and calls an operating-system command to export a file. When customer data is staged first, its object score changes from 31 to 6. Order ingestion and file export still require separate planning. The saved report estimates 58 hours for the original SQL scope.

NorthTrace, powered by Veyra AI · Migration assessment

Assess a SQL Server schema

Which objects should an architect review before moving this ERP to a lakehouse? Run the scoring rules on a fictional distributor’s schema, then inspect the SQL behind each finding.

The original Python scanner parsed the samples and resolved their dependencies. This preview calculates risk flags, complexity and estimated effort in your browser. No SQL statements are executed and no AI service is called.

Open the original sample report ↗

Saved assessment report · Original ERP schema · October 28, 2026

Interactive lab · synthetic data

Trace the revenue gap

Two systems report the same period in USD. Inspect the source rows, test the explanation and decide whether the evidence supports closing the case.

Revenue reconciliationDeterministic simulation
ERP ledger$12.4M
BI reporting$12.1M
ERP − BI$300K
Synthetic USD records. All calculations and review decisions run in your browser. No live model or external system is connected.

Reference design

Where an agent fits

A production design would let a model propose an explanation while SQL establishes the amounts. Access controls, evidence identifiers and review decisions travel with the case.

AI-assisted enterprise architectureConceptual design · select a stage
01Enterprise sources Provenance

Capture the source, business key, reporting period and extraction time. Align populations before comparing totals.

02Governed data platform Access control

Apply identity-aware access to curated evidence tables. Preserve source identifiers and lineage without forwarding unrestricted source data.

03Quality and semantic layers Definition checks

Test source completeness and compute measures under an explicit reporting contract. Record the measure version and calculation window.

04Evidence retrieval Scoped context

Select authorized exception rows, relevant metric definitions and execution metadata. Each evidence item carries a stable citation.

05Agent orchestration Bounded investigation

Choose from approved read-only checks and record tool calls. Stop on missing evidence, exceeded budgets or unauthorized requests.

06LLM reasoning Evaluation

Draft an explanation from retrieved evidence. Evaluate grounding, unsupported attribution and the ability to abstain against labeled cases.

07Validation and review Human oversight

Recalculate totals deterministically. A named reviewer examines the evidence and approves or rejects the proposed disposition.

08Business insights Resolution record

Store the decision, owner, evidence identifiers and residual variance. Keep pipeline latency, failed checks and model runs observable.

Current experiment

In my local prototype, DuckDB isolated one $300,000 synthetic invoice missing from BI. Qwen3 4B through LM Studio summarized the SQL results. The upstream cause remained unknown. This browser exercise uses a separate, two-exception dataset so visitors can test the controls.

Research directions

Agent orchestration, local model inference and evaluation are areas of ongoing exploration. The next technical question is whether a model can consistently cite the right evidence and abstain when a cause is ambiguous.