ListMatchGenie

The Genie

The AI persona that guides, explains, and reports across every stage of ListMatchGenie. Not a chatbot — a narrator that turns raw matching into something you can reason about.

The Genie is the AI layer that threads through every screen of ListMatchGenie. It is not a chatbot you have to prompt — it is an always-on narrator that translates the mechanical work of cleansing and matching into plain English, so you can make decisions without reading a spec.

If you've ever had to explain to a colleague why your VLOOKUP matched 60% of rows and what to do about the other 40%, the Genie is the tool that saves you that conversation.

What the Genie actually does

The Genie appears in four distinct roles across the product. Knowing which role you're interacting with helps you understand what to expect.

Profiler

When you upload a file, the Genie inspects every column and reports back: type detected, null rate, distinct values, distribution, and any data-quality anomalies (mixed casing, trailing whitespace, inconsistent date formats, possible PII).

This is what lets you catch problems before they break your match — for example, spotting that a "phone" column actually contains email addresses in 30% of rows, or that the "state" column has 60 distinct values when there should only be 50.

Cleanser

The Genie makes the call on which cleansing operations to run on each column and documents them in the cleansing report. You can override any decision, but the defaults are chosen based on what matching will actually need — there's no busywork.

Matcher narrator

While the match engine runs, the Genie produces a match narrative: a human-readable summary of what matched, why, and where the uncertainty lives. You'll see phrases like:

"Of your 4,812 source rows, 3,147 matched an existing master record with high confidence. Another 842 matched but fell into the review band — mostly address differences on otherwise identical names. The remaining 823 had no plausible master record, concentrated in your Pacific Northwest region."

This is the single highest-value output in the product. The raw numbers are in the stats cards; the Genie tells you what they mean.

Report writer

When you generate a report on a match job, the Genie produces a structured analytical document with an executive summary, data-quality narrative, and key findings. You can ask follow-up questions inline — the Genie answers in the same document with charts and suggested next prompts.

A note on voice

The Genie writes in third person as a narrator — "The Genie detected..." or "Your data shows...". Not "I detected" and not "we detected". This keeps reports readable as standalone artifacts when you share them with stakeholders who have never used ListMatchGenie.

What the Genie is not

The Genie is deliberately scoped. It is not:

  • A general-purpose assistant you can ask unrelated questions.
  • A replacement for your own judgment on borderline match decisions. The review queue exists because the Genie flags uncertainty rather than papering over it.
  • A place to upload your data to a third-party LLM. The Genie sends schema and aggregate statistics to the model — never raw rows, never PII. See PII and security for the exact boundary.

How the Genie uses your data

This is the most common question from customers evaluating the product, so we'll be specific.

PII never leaves your region

The Genie receives column names, data types, null rates, cardinality, and synthetic samples (never real values). The AI model produces narratives from this metadata. Your actual row data lives in regional S3 storage and is processed by the matching engine only — never sent to an LLM.

When the Genie writes a narrative, it is working from aggregate statistics: match counts per classification, score distributions, column-level cleansing summaries. Real values in your files are never transmitted to the AI provider. This is the invariant that lets ListMatchGenie stay GDPR-compliant while still delivering useful insights.

Where the Genie appears

  • Upload screen — column profiles and type overrides
  • Cleanse screen — per-issue counts and narrative summary of what's about to change
  • Run screen — live progress commentary as passes complete
  • Review screen — "The Genie's Take" summary card at the top of every completed match
  • Report pages — executive summary, data-quality narrative, key findings, inline Q&A
  • Dashboard — short summaries of your most recent matches

You don't need to learn a prompt syntax. Everywhere the Genie appears, it already knows the context of what you're looking at and what you most likely need to know.