Launch by Lunch × ELIXR
Published June 2026. Link expires June 2027.
ELIXR × LBL · CareConscious Research Workspace

Everything your team needs to run the Research Workspace, in one place.

How the three skills work, how to run a cycle, where every file lives, and how to stand up a new client. Bookmark this page — it's your home base.

01 · Start here

The system at a glance

The Workspace turns raw census data into the inputs for your CareConscious report. The tool does the heavy lifting and lays everything out neatly, but your team makes every important call. Nothing moves to the report without human eyes.

1

Three skills, run in order

Themer codes the comments, PII Scrubber makes a safe-to-share copy, Analysis stages the findings. Each one ends in a human checkpoint.

2

Two streams of data

The comments (open-ends) and the numbers (Forsta crosstabs) travel separately and only meet at the final Analysis stage.

3

It all runs in Cowork

Everything runs in Cowork (Claude Desktop) on the ELIXR Team plan. Claude reads each skill and uses the Google Drive connector to read inputs and write outputs.

02 · The three skills

What each skill does.

Run them in order. Each confirms its inputs at Step 0, stops at its own human checkpoint, and emits a plain-language Results Summary at the end.

Skill 1

Themer

Reads the open-ended comments and tags each one with themes and a sentiment — what Nichole used to do by hand. It proposes the theme list, never finalizes it on its own.

  • In: the open-end export; the seed theme library
  • Does: Pass 1 drafts the taxonomy + sentiment rubric; Pass 2 codes every comment against the locked list
  • Out: Pass 1 draft, coded sheet, theme summary
Checkpoint · Nichole locks the theme list
Skill 2

PII Scrubber

Makes a safe-to-share copy by removing anything that could identify a person, while keeping the substance. The rule of thumb: redact who someone is, keep what they said. Over-redaction is the default.

  • In: the Themer's coded sheet; the redaction lists + per-client known-list
  • Does: Pass 1 mechanical scrub; Pass 2 catches identifying combinations and logs every removal
  • Out: a QA workbook, then the final client asset
Checkpoint · Nichole audits a ~30-row sample
Skill 3

Analysis

Brings the comments, the numbers, and the Intake Library together and stages the findings for the report. It does the bookkeeping so your judgment lands earlier and faster — it never replaces your call.

  • In: coded sheet, the 7 Forsta crosstabs, the Intake Library, Research Focus Guide, Cycle Adjustments Log
  • Does: labels each finding (Confirmation / Sharpening / Tension / Silence), scores confidence, stages the Top 10
  • Out: staged findings Sheet + topline brief Doc
Checkpoint · Jess, Elle & Nichole sign off

Always flagged for you: every Tension (a finding that contradicts what you knew) gets a mandatory review — the report cannot ship a contradiction without your stated point of view. Sentiment helps pull supporting quotes; it never ranks a finding.

03 · Run a cycle

How to run it, start to finish.

You don't paste instructions each time. Inside the client Project, run the three skills in order in one chat, finishing each checkpoint before starting the next. The lines below are all you have to say.

Before you start (prerequisites): the ELIXR Claude Team plan is live; Claude Desktop (Cowork) is installed and signed into the ELIXR account; the Google Drive connector is connected to the ELIXR Shared Drive with read and write access; and the three skills are published to the org skill directory. Each team member connects their own Drive — sharing the Project does not share Drive access.

Step 0 · each skill confirms its files → 1 · Themer → 2 · PII Scrubber → 3 · Analysis
01

Themer

"Run the Themer skill."

It drafts the taxonomy + sentiment rubric (Pass 1) and stops for Nichole to lock the theme list. Once it's locked, tell it to continue and it codes every response (Pass 2).

Or, for finer control:

  • "Themer: build taxonomy" — Pass 1 only
  • "Themer: code responses" — Pass 2, after the lock
  • "Classify the 'Other, specify' fields"
02

PII Scrubber

"Run the PII Scrubber on the Themer's coded output."

It runs both passes, auto-redacts (over-redaction preferred), and produces the QA workbook for Nichole. After she signs off, tell it to generate the client asset.

Or, for finer control:

  • "Pass 1 — mechanical known-list scrub"
  • "Pass 2 — identifiability + proper-noun scrub"
03

Analysis

"Run the Analysis skill."

It reads the coded sheet, the seven crosstabs, the Intake Library, and the Cycle Adjustments Log, applies the quant rules and guardrails, and stops at the staging checkpoint for Jess, Elle, and Nichole.

Or, for finer control:

  • "Structure the Intake Library spine (Phase 1)"
  • "Classify themes and quant cuts, with confidence (Phase 2)"
  • "Stage classified findings to the bets framework (Phase 3)"
  • "Run all three phases sequentially to the checkpoint"

Mid-cycle facts: drop a note on the Cycle Adjustments Log and Analysis reads it for that run. At the checkpoint, the team decides which notes are keep-forever (promoted into the Intake Library) versus one-off.

04 · Where everything lives

The Drive is the system of record.

All client and build content lives in the ELIXR Shared Drive (org-owned, survives departures), not a personal Drive. The skills find their inputs by standard name inside one client folder, so you never edit a skill to onboard a client.

The engagement folder

Each client has one engagement folder — 2026 - <Client> Engagement 1/. Put this folder's URL in the Project instructions. It has two working subfolders:

Naming standard

Swap <CLIENT> for the client's short code (e.g. MSC) and <N> for the cycle number. These are the exact document titles.

Inputs — Advisory & Analysis/

Inputs — Research & Census Analysis/

Outputs — Research & Census Analysis/Claude Outputs/

How Claude resolves a file

A native Google Sheet wins over a same-named .xlsx (keep convert-on-upload ON) — the Forsta crosstabs are the exception, staying read-only .xlsx. Only the current version lives in the working folder; superseded ones go to Archive/. For dated files, use the most recent. At Step 0 every skill states what it found and stops if anything is missing or ambiguous.

05 · Set up a new client

Onboarding a new engagement.

You only set up a new Project and a new folder — you never edit a skill. Follow these in order.

01

Create the engagement folder

Reuse ELIXR's shape: two working subfolders, Advisory & Analysis and Research & Census Analysis. Inside Research & Census Analysis, add three subfolders: Data Tabs, Claude Outputs, and Archive.

02

Add the files, named to standard

Knowledge docs (Intake Library, Research Rules, seed Theme Taxonomy) into Advisory & Analysis. The Research Focus Guide and an empty Cycle Adjustments Log at the root of Research & Census Analysis. The OE export and the Forsta crosstabs into Data Tabs/.

03

Create a new Project (copy/paste the instructions)

Duplicate the Project in Cowork and paste the block below as its custom instructions. Change only the client name and the working-folder link. The Project keeps the cycle chats together and applies a standing orientation — it does not hold the skills or the knowledge.

This Project runs the ELIXR CareConscious Research Workspace for {{CLIENT NAME}}. Three skills run in order, each ending in a human checkpoint: Themer (codes the open-ends), PII Scrubber (makes the client-safe copy), Analysis (stages the report content). The system of record is the client's engagement folder in Drive, not this Project: read inputs from Drive each run, do not paste their contents here. The engagement folder has two working subfolders: Advisory & Analysis (the knowledge we curate: Intake Library, Theme Taxonomy, Research Rules) and Research & Census Analysis (the raw data and the cycle-level steers: the Research Focus Guide and the Cycle Adjustments Log at the root, the open-end export and the Forsta crosstabs in its Data Tabs subfolder, plus the skills' outputs in its Claude Outputs subfolder). Resolve every input by its standard <CLIENT>_ name (the skills do this); where a Google Sheet and a same-named .xlsx both exist, use the Sheet (the Forsta crosstabs are the exception). Use the current version of each file; superseded versions live in Archive. Before running, confirm the client's engagement folder is shared with this account so the Drive connector can resolve files. Engagement folder for this Project: {{ENGAGEMENT FOLDER URL}} Canonical inputs (the skills resolve these by name; no IDs to maintain): - Open-end export (Themer input): <CLIENT>_OE Export in Research & Census Analysis/Data Tabs/ (use the latest dated) - Forsta crosstabs (Analysis input): Tab_<cut>_<CLIENT>.xlsx in Research & Census Analysis/Data Tabs/ - Locked theme taxonomy (seeds each cycle): <CLIENT>_Theme Taxonomy in Advisory & Analysis/ - Intake Library, native Sheet (Analysis reads it and writes durable facts back on approval): <CLIENT>_Intake Library in Advisory & Analysis/ - Research Focus Guide (the bets): <CLIENT>_Research Focus Guide in Research & Census Analysis/ (root, .docx) - Research Rules (PII known-list + false positives): <CLIENT>_Research Rules in Advisory & Analysis/ - Cycle Adjustments Log: <CLIENT>_Cycle Adjustments Log in Research & Census Analysis/ (root) Team and checkpoints: Nichole locks the taxonomy (Themer) and signs off the PII sample; Jess, Elle, and Nichole sign off the Analysis staging. Over-redaction is preferred in the client copy. Always confirm the inputs are in place before running, and stop and flag if any are missing.
04

Connect Drive & run in order

Each team member connects their own Google Drive and confirms access to the folder. Then run the skills in order, confirming the files Claude lists at Step 0 before proceeding.

06 · Maintain a skill

Updating a skill (maintainer only).

Skills are edited only by the maintainer (Jess), on purpose. Adjusting how a skill behaves inside a chat changes only that chat — to change it for everyone, the maintainer edits and republishes.

1 · Edit the skill

Edit the SKILL.md directly in Claude (Customize → open the skill), keeping the YAML frontmatter (name, description, when_to_use) intact at the top. Keep the name lowercase-with-hyphens.

2 · Save & re-share

Save the skill and re-share it with the team. The team will auto-receive the update.

07 · What's built in

Safeguards & conventions.

These run automatically, so the deliverable is never built on partial or unsafe data.

Validation gate

A run will not build its deliverable until every response is processed. A failed or missing item is left for an automatic re-run, never silently coded or under-redacted. Code 0 = blank / non-response, kept separate from a real theme and from a processing failure.

Privacy defaults

Over-redaction is the default in the client copy (analysis is unaffected — it runs on the internal text). Business-unit and location demographic fields are excluded from the client file and kept in the internal files.

Quant rules

Percentages tab only; any group of 5 or fewer people is suppressed; a single 95% significance gate; cells that are true by definition are skipped.

Analysis guardrails

Hourly vs Salaried is the master cut; a structural absence can itself be the finding; wins worth protecting are flagged as such; every Tension carries a mandatory-review flag; sentiment is context for retrieval, never a ranking weight.

08 · The full documentation

The handoff documents.

The complete references behind this page. Open these for the line-by-line detail.

01

Claude Team Account — Quick-Start Guide

New to Claude? Start here. The whole-team onboarding from Karen: the desktop app, Chat vs Projects, building skills, setting up routines, which model to use, and your first-week homework. Open the quick-start →

02

Phase 1 Research Workspace — overview

The one-page picture of what was built and how it fits together. Open the overview →

03

Cowork Setup & Run Guide

The practical "how do I stand this up and run it" companion: prerequisites, publishing the skills, the Project setup, the per-skill run prompts, and the new-client checklist. Open the setup & run guide →

04

How the Process Works (End to End)

The full walkthrough of how raw census data becomes the report inputs — what goes in, what the tool does, where a human reviews, and what comes out at each stage. Open the end-to-end doc →