Phrase Health
Request demo Login
QUERY CONCIERGE

Answer every data request with the rigor of your best analyst

An AI intake concierge that gathers requirements, applies your institutional definitions, and drafts SQL — at the speed your organization actually demands.

Request a demo Follow a request
01The front door

Anyone can ask, in plain language.

No SQL. No ticket template. People describe what they need — “how many CHF patients were seen by cardiology within 2 weeks of discharge?” — and Concierge takes it from there.

02Discovery

The bottleneck was never writing the query.

It’s the requirements — what a population really means, the back-and-forth, the semantics meeting. Concierge runs that discovery right in the conversation, the way your best analysts do.

03Refinement

It reshapes vague asks into answerable ones.

Concierge reads the real intent and proposes a sharper question — turning “how many” into the cut that actually drives a decision.

04Your institutional knowledge

Every request carries your definitions and sources of truth.

Concierge learns from your query library and cloud ETLs, then applies your accepted definitions automatically — so a population means what your team means.

05The handoff

Then it hands off to your analysts for final approval

The requestor gets an instant confirmation, and a fully vetted question and fully-drafted query land in your analysts' queue — discovery already done, ready to validate and run. Weeks become the same afternoon.

See it on your data
Query Concierge
app.phrasehealth.com / concierge
Query ConciergeNew request

What can we help you find?

Describe the data you’d like. We’ll walk you through a few clarifying questions and produce a SQL draft for the analyst team.

How many CHF patients were seen by cardiology within 2 weeks of discharge?

Cmd/Ctrl + Enter to submit
Query Concierge Gathering requirements
How would you like to define ‘seen by cardiology’ — by provider specialty, or department specialty?
Department specialty.
Checking for the common gotchas…
Query Concierge Refining the request
It sounds like you’re working to improve discharge continuity. Would it be more helpful to tweak this to “what proportion of CHF patients were seen by cardiology within 2 weeks of discharge, by discharge service?”
Yes, use the refined question.
Query Concierge Confirming definitions
It appears your organization defines Congestive Heart Failure (CHF) as ICD-10 I50.* present on the problem list. Do you want to modify that?
No, that’s correct Also require reduced EF Add hypertensive HF (I11.0)
Learned from your query library & cloud ETLs
Query Concierge Routed to analytics
A query has been generated. We’ll send it to your analytics team to validate and run — your results will appear right here.
DR ML Handed to analytics
New data request
Routed from Query Concierge
chf_cards_followup.sql
SELECT svc.name AS discharge_service,
       COUNT(DISTINCT e.pat_id) AS n_seen
FROM   ehr.encounter        e
JOIN   inst.chf_definition    chf  -- inst. def.
         ON chf.pat_id = e.pat_id
JOIN   ehr.appointment      c
         ON c.pat_id = e.pat_id
WHERE  c.dept_specialty = 'Cardiology'
  AND  DATEDIFF(day, e.disch_date,
            c.appt_date) <= 14
GROUP BY svc.name;
Two ways to deploy

Fit Concierge to the team you have.

Concierge + your analysts

Amplify the team you already have. Every analyst works with the discovery done and your best query as a starting point.

  • Clear the backlog without adding headcount
  • Analysts focus on the work that needs judgment
  • Your standards applied to every request

Concierge + Phrase’s analysts

A managed option to absorb overflow — the consultant spend you likely already have, redirected to cleared requests.

  • Overflow handled without hiring
  • Runs on your definitions and standards
  • Scale up and down with demand
“But we already have…”

Where Concierge fits alongside what you’ve already built.

SLICER DICER

Great — until the query needs a complex join

It works when the user is savvy, the question is simple, and the model is mapped to Caboodle. That covers less than half. Let's face it, custom SQL reporting will never go away. Concierge handles exactly that gap.

Epic’s AI tools

Helpful — but not end-to-end

One feeds SlicerDicer with the same constraints; the other augments analysts writing SQL. Neither does intake and discovery from the requestor, and neither injects your institution-specific definitions.

Warehouse · Databricks · Caboodle

It builds on that investment

Concierge learns from your query library and ETLs — so the work you’ve already done rides along in every request, rather than being rebuilt from scratch.

The outcome you can take credit for

Same-day answers. A backlog that finally moves the right way.

Same-day turnaround

Requests that took weeks come back ready to validate — the same day they’re asked.

A backlog that clears

The queue stops growing — and your team gets ahead of demand instead of behind it.

Consistent rigor

Every query carries your best analyst’s judgment — no matter who’s on call.

Give your organization same-day answers — and give your analysts back the work that matters.

See Query Concierge run on your own definitions in a personalized walkthrough.

Request a demo Built for analytics leaders. No backlog required.