Showrunner

gemini-4-argonrun 3 Oct 2026122,180 tokens

Instructions

Instructions — what the model was asked

Hi, I need to create a modified Claude code for beginners session, like run sheet basically, run of show, that has the following modifications. So I'm going to upload a few different examples of things I've done in the past, just so you have them to work from. But basically, there's a few modifications for this one which make it unique. One is that this is only using Claude Cowork, so we should...

Hi, I need to create a modified Claude code for beginners session, like run sheet basically, run of show, that has the following modifications. So I'm going to upload a few different examples of things I've done in the past, just so you have them to work from. But basically, there's a few modifications for this one which make it unique. One is that this is only using Claude Cowork, so we should not be using Claude Code, either in the desktop app or in the terminal, because this team is non-technical. The team are all working in financial services slash accounting for a education company. So what kind of workplace education company called Brightmoor Education. So the sorry that spelled b the uh team is very busy and um doesn have a lot of like time to explore speculative use cases of ai they want this to be incredibly practical and based on their existing workflows. I'm uploading some transcripts from a call we had with them and also just some notes on what I think could work. Specifically, they're interested in... creating the dashboards and specifically they mean just making the information look visibly better like like creating a html dashboard in a nice kind of style from from existing say data that contained in Excel They also interested in doing financial analysis So more of a, you know, like, I guess, a cloud coworker would use Python to do the analysis and then maybe output the analysis in Excel or the input would be in Excel. And then the third thing was like making a PowerPoint where they update the PowerPoint template automatically using Claude Cowork so that they can do kind of weekly or monthly presentations on the data. So that's the goal, is to kind of make it very focused on that. I still want to keep the way that we introduce Claude Cowork, where we explain to it that the difference between Claude Cowork and ChatGPT is that it has access to the file system. It can run code. I think that important But maybe the ASCII image idea that I normally do where I ask it to create an ASCII image based on my location isn like that useful I think we need to make it more practical for them So maybe we use an Excel implementation instead. The assumption is... oh, so they use Snowflake, by the way, for all their data, and that is powered by NetSuite, which is our accounting software. I guess that's by Oracle. So when we create synthetic data, we'll be, for these tasks that we're going to demo, we'll be focused specifically on solving that, solving those problems. So like the shape of the tasks, the shape of the data should be like NetSuite shaped, if that makes sense. So make it very tight run down, make it very condensed and practical. No real like wider theory or like, hey, we're excited about AI things. I think these guys just want to get stuff done.
So yeah, let me know what the run should be.

some very specific asks in here

  • definitely non-technical (claude cowork only + claude in excel)
  • core focus on accounting/finance function
  • use cases to demo: creating a dashboard, doing financial analysis, making a powerpoint

heavy focus on practical stuff rather than theory - they're just super busy and want to get stuff done, have heard from friends at other companies you can do it with claude

For accounting: put financials together, doing reconciliations, identifying trends in data sets, putting together repeatable compliance reports For planning & forecast: build financial models (make assumptions and pull & clean data from snowflake)create dashboards, and publish to executivesAutomate charts on slides

17 of 17 checks passed

ANSWER.mdthe model's output · 32.6 KBdownload

Claude Cowork for Finance & Accounting — Brightmoor Education Run of Show

Executive Overview & Session Parameters

  • Client: Brightmoor Education (Multi-Campus Workplace & Higher Education)
  • Audience (20–30 Attendees):
    • Accounting, Payroll & Bursar (~85% of room): Led by Marcus (Chief Controller / CAO). Mostly AI beginners ("rookies"). Focused on putting financials together, subledger-to-GL reconciliations, payroll/tuition tie-outs, identifying trends/anomalies across datasets, and building repeatable month-end close & compliance reports.
    • Strategic Finance / FP&A (~15% of room, ~3–4 people): Led by Dana Yoon (plus Greg Nerlin). Intermediate/advanced users who have experimented with Claude. Focused on pulling/cleaning data from Snowflake, building driver-based financial models with assumptions, creating interactive HTML dashboards published to senior leadership, and automating charts/tables/commentary in monthly PowerPoint decks.
  • Tooling Scope (Strictly Non-Technical):
    • Included: Claude Cowork (in the Claude Desktop App) + Claude in Excel (and Claude in PowerPoint where available).
    • Excluded: No Claude Code, no Terminal/CLI, no GitHub, and no volunteer peer TAs (per Dana's feedback that the team is too stretched to TA).
  • Data Architecture: Oracle NetSuite (Accounting GL, Payroll, Bursar subledgers) $\rightarrow$ Snowflake (Central Data Warehouse) $\rightarrow$ Excel & PowerPoint. Because IT provisioning for the live Snowflake connector takes ~2 hours per user, the workshop runs on NetSuite/Snowflake-shaped synthetic Excel/CSV files (with live Snowflake connection treated as an optional bonus during the buffer).
  • Duration: 3.5 Hours Core Workshop (210 minutes) + 30-Minute Buffer / 1:1 Clinic (4-hour total calendar block).
  • Tone & Philosophy: Zero fluff, zero AI hype, 100% practical execution. No theoretical lectures on "why now is the best time to build" or speculative toy apps (no ASCII maps, no vacation planners). Every minute is anchored in cutting hours off Brightmoor's actual monthly close, financial analysis, dashboarding, and executive reporting workflows.

Part 1: Client-Facing Agenda (30–45 Minute Blocks)

Ready to send to Dana, Greg, and Marcus on Monday.

Time Block Duration Module What We'll Cover & Build
0:00 – 0:30 30 min 1. How Claude Cowork Works & First Excel Win Why Claude Cowork is different from ChatGPT/Gemini: it has direct access to your local file system and runs Python/Excel formulas behind the scenes so it doesn't guess math. Hands-on starter: Point Cowork at a folder of raw NetSuite exports, standardize multi-campus tabs into a clean Excel workbook, and audit formulas using Claude in Excel.
0:30 – 1:20 50 min 2. Automated Financial Analysis, Reconciliations & Modeling Using Claude Cowork (running Python under the hood) + Claude in Excel to do heavy analysis on NetSuite/Snowflake exports and output multi-tab, formula-driven Excel workbooks.• Accounting/Payroll/Bursar: Automated subledger-to-GL reconciliations, trend/flux detection, and repeatable compliance checks.• FP&A: Cleaning Snowflake exports, building a Revenue $\rightarrow$ EBIT $\rightarrow$ Adjusted EBITDA bridge (Payroll vs. Non-Payroll, Feb + YTD Budget vs. Actuals), and modeling assumptions.
1:20 – 2:05 45 min 3. Executive HTML Dashboards from Excel (includes 10m break) Turning flat Excel financials into a clean, modern interactive HTML dashboard for Senior Leadership—without touching code.• Build: Generate a multi-campus financial dashboard from Excel.• Configure & Tune: Swap chart types (e.g., vertical column to horizontal bar chart), add interactive filters (Campus sidebar, Department, Monthly vs. YTD toggle), and learn how to publish/share internally outside Finance.
2:05 – 2:50 45 min 4. Automating Monthly PowerPoint Reports & Charts Cutting the 8-hour monthly deck grind down to a repeatable workflow.• Quick Win: Take 1 existing chart and have Claude mirror/duplicate it across 5+ slides (one per Brightmoor campus).• End-to-End Flow: Translate 5 standard Excel tables (Revenue variance, EBITDA bridge, Payroll split) into a PowerPoint template, generate charts, and draft data-backed Feb + YTD commentary bullets.
2:50 – 3:30 40 min 5. Troubleshooting, Data Validation & Role-Specific Sprint How to verify numbers (Control Totals & Reconciliation Tabs) and fix formatting quirks after the workshop (ASK + Verify framework). Hands-on sprint split by role:• Track A (Accounting/Payroll/Bursar): Reconciliations, anomaly detection & repeatable compliance schedules.• Track B (Strategic Finance/FP&A): Snowflake data cleaning, forecast modeling & executive deck automation.
3:30 – 4:00 30 min 6. Optional Buffer: 1:1 Clinic & Live Workflow Setup Optional 30-minute clinic to apply today's prompts to your own sanitized Brightmoor spreadsheets, troubleshoot edge cases, or review Snowflake connector setup.

Part 2: Detailed Facilitator Run of Show (Run Sheet)

Pre-Session Setup Requirement (Sent 48 Hours Prior)

To ensure 0 minutes are wasted on IT troubleshooting during the call:

  1. Claude Desktop App installed with Claude Cowork enabled and logged into provisioned Brightmoor enterprise accounts.
  2. Claude in Excel add-in enabled.
  3. Workshop Folder (Brightmoor-Finance-Workshop.zip) downloaded and unzipped to Desktop/Brightmoor-Finance-Workshop/.

Module 1: How Claude Cowork Works & First Excel Win (0:00 – 0:30 | 30 mins)

1A. Welcome & The Core Mental Shift (0:00 – 0:10 | 10 mins)
  • Facilitator Framing (No Hype):
    • "We know everyone here is in the middle of close, payroll cycles, and monthly reporting. We aren't going to talk about AI theory or build toy apps today. Every exercise today uses NetSuite and Snowflake-shaped financial data for Brightmoor's campuses to solve three things: Excel financial analysis & reconciliations, executive HTML dashboards, and monthly PowerPoint automation."
  • Why Claude Cowork + Claude in Excel $\neq$ ChatGPT in a Browser:
    1. File System Access: In a browser chatbot, you copy-paste snippets back and forth. In Claude Cowork, you point Claude at a folder on your computer (Desktop/Brightmoor-Finance-Workshop/). It reads all your CSVs, Excel models, and PowerPoint templates directly, and writes brand-new files straight into your folder.
    2. Deterministic Code Execution (Zero Mental Math): You should never trust a chatbot to add up 10,000 rows in its head. When you give Claude Cowork a financial task, it quietly writes and runs Python scripts or native Excel formulas behind the scenes. You don't need to know a single line of Python—Claude uses Python as its calculator and hands you back a clean .xlsx, .html, or .pptx file.
1B. "Take It for a Spin": NetSuite Folder Consolidation & Excel Audit (0:10 – 0:30 | 20 mins)

(Replaces the standard "ASCII Map of Office Locations" starter prompt with an immediate Excel & file-system win.)

  • The Setup: Everyone opens Claude Cowork and selects the Desktop/Brightmoor-Finance-Workshop/01_Raw_Exports/ folder. Inside are 4 messy CSV/Excel files exported from NetSuite and Snowflake:
    • NetSuite_GL_TrialBalance_Jan2026.csv
    • NetSuite_GL_TrialBalance_Feb2026.csv
    • NetSuite_Payroll_Register_Feb2026.csv
    • Bursar_Campus_Tuition_Subledger_Feb2026.xlsx
  • Guided Starter Prompt (Everyone runs together in Cowork):

    "Look at all the files in this folder. First, give me a quick inventory of what files are here, their row counts, and any formatting inconsistencies (like inconsistent Campus names or date formats). Then, use Python to combine the January and February NetSuite GL Trial Balance CSVs into a single clean Excel workbook named 01_Brightmoor_Consolidated_GL.xlsx. Standardize the Campus names across all rows, map the GL account numbers into standard rollup categories (Revenue, Payroll Expense, Non-Payroll OpEx, D&A), and include a first tab called Control_Totals using live Excel SUM formulas to prove that the cleaned tab matches the raw CSV totals down to the cent."

  • What Participants Experience (10 mins):
    • Watch Claude Cowork inspect the local directory, write a quick script to clean messy NetSuite strings (e.g., "Bay Area - 01", "SF/Bay Area", "BayArea" $\rightarrow$ "Bay Area"), and deposit 01_Brightmoor_Consolidated_GL.xlsx right into their folder.
  • Hands-On with Claude in Excel (10 mins):
    • Participants open 01_Brightmoor_Consolidated_GL.xlsx in Excel.
    • Open Claude in Excel side panel and run two quick prompts:
      1. "Check the Control_Totals tab and explain how the reconciliation formula is verifying that zero dollars were dropped from the raw Jan and Feb exports."
      2. "Add a new column Cost_Type in the consolidated GL tab using an Excel formula that tags accounts 5000–5199 as 'Payroll' and 5200–5999 as 'Non-Payroll', and highlight any unmapped account codes in yellow."

Module 2: Build 1 — Financial Analysis, Reconciliations & Modeling (0:30 – 1:20 | 50 mins)

2A. The Repeatable Workflow (Talk → Jam → Make → Tune) & Trusting the Numbers (0:30 – 0:40 | 10 mins)
  • The 4-Step Loop for Finance:
    1. Talk (Plan First): Tell Claude what files to read, what business logic to apply, and what tabs/columns you want in the final Excel output—ask it to confirm its plan before generating the file.
    2. Jam (Align on Definitions): Clarify accounting/FP&A rules (e.g., "How should Adjusted EBITDA treat one-time campus relocation costs?" or "Which GL accounts map to Bursar tuition vs. fee revenue?").
    3. Make (Execute via Python $\rightarrow$ Excel): Let Cowork run the Python analysis and build the multi-tab Excel workbook with live formulas.
    4. Tune (Refine & Audit): Request specific adjustments (add variance % flags, adjust forecast assumptions, format currency/basis points).
  • Addressing the #1 Finance Question ("How do I know the numbers are accurate?"):
    • Rule 1: Always require live Excel formulas (SUMIFS, XLOOKUP, INDEX/MATCH) in the output workbook rather than hardcoded static numbers wherever possible.
    • Rule 2: Always ask for a Tie-Out / Reconciliation Check row at the bottom of every schedule (= Raw Total - Report Total, formatted to turn red if $\neq 0$).
2B. Guided Build 1: NetSuite/Snowflake Data $\rightarrow$ Multi-Tab Financial Analysis Workbook (0:40 – 1:20 | 40 mins)
  • The Scenario: It's February month-end close at Brightmoor Education. Both Accounting and Strategic Finance need a single source-of-truth Excel package (02_Brightmoor_Feb_Financial_Package.xlsx) built from the raw NetSuite GL, Payroll Register, Bursar Subledger, and Snowflake Budget export (Snowflake_FY26_Budget_Forecast.csv).
  • Step 1 — Talk & Jam (10 mins):
    • Participants prompt Cowork in Plan mode:

      "I need to build our February Month-End Financial Analysis workbook (02_Brightmoor_Feb_Financial_Package.xlsx) from the files in this folder. Before you build it, propose a plan for a 4-tab Excel workbook containing:
      1. EBITDA_Bridge: Full bridge from Tuition & Fee Revenue $\rightarrow$ Gross Margin $\rightarrow$ Payroll Expense vs. Non-Payroll OpEx $\rightarrow$ EBIT $\rightarrow$ Adjusted EBITDA (adding back D&A and one-time campus restructuring costs). Columns must show Feb Actual, Feb Budget, Feb Variance ($ and %), and YTD (Jan + Feb) Actual, YTD Budget, YTD Variance ($ and %).
      2. Campus_Summary: The same P&L and Adjusted EBITDA breakdown segmented across all 6 Brightmoor Campuses.
      3. Reconciliations_&_Flux: (For Accounting/Payroll/Bursar) Reconcile the Feb Bursar Tuition Subledger and Feb Payroll Register against the NetSuite GL by campus and invoice/employee batch, listing every unmatched variance > $100, plus a MoM Flux table flagging GL accounts that changed >10% from Jan to Feb.
      4. Forecast_Model: (For FP&A) A Q2 rolling forecast tab where Enrollment Growth %, Tuition Discount %, and Payroll Merit % are editable assumption cells at the top that dynamically drive projected Apr–Jun EBITDA via live Excel formulas.
      Tell me how you will map the accounts and verify control totals before writing any code."

    • Review Claude's questions/plan together (2–3 mins), confirm account mappings, and say "Go ahead and build it."
  • Step 2 — Make (15 mins):
    • Claude Cowork writes and executes Python to join the NetSuite GL, Payroll, Bursar, and Snowflake Budget tables, runs the subledger-to-GL matching logic, and outputs 02_Brightmoor_Feb_Financial_Package.xlsx with clean number formatting ($#,##0;($#,##0);"-"), bold subtotal rows, and live Excel formulas.
  • Step 3 — Tune & Inspect in Excel (15 mins):
    • Open 02_Brightmoor_Feb_Financial_Package.xlsx.
    • Accounting check: Inspect Tab 3 (Reconciliations_&_Flux)—see how Python instantly identified 4 timing differences between the Bursar subledger and NetSuite GL and 2 duplicate payroll accruals.
    • FP&A check: Inspect Tab 1 (EBITDA_Bridge) and Tab 4 (Forecast_Model)—change the Bay Area Enrollment assumption cell in Excel from 3.0% to 4.5% and watch Q2 Adjusted EBITDA update automatically.
    • Tune Prompt: Ask Cowork (or Claude in Excel) to add an Automated Variance Commentary column on the EBITDA_Bridge and Reconciliations_&_Flux tabs explaining which campus and sub-account drove each variance.

Module 3: Build 2 — Executive HTML Dashboards from Excel (1:20 – 2:05 | 45 mins)

3A. Guided Build 2: Excel Financials $\rightarrow$ Interactive HTML Dashboard (1:20 – 1:40 | 20 mins)
  • The Scenario (Directly from Dana's Call): Formatting dashboards inside Excel is clunky and fragile. Senior Leadership wants a clean, modern web-style view where they can see Consolidated financials and drill down by Campus on the left-hand side.
  • Step 1 — Talk, Jam & Make (20 mins):
    • Participants prompt Claude Cowork:

      "Using 02_Brightmoor_Feb_Financial_Package.xlsx, create a single-file, self-contained interactive HTML dashboard named Brightmoor_Executive_Dashboard.html designed for Brightmoor Education's Senior Leadership.
      Include:
      - A left-hand navigation sidebar listing all Brightmoor Campuses (All Campuses Consolidated, Bay Area, Austin, Chicago, New York, Boston, Online) so clicking a campus dynamically updates all KPI cards, charts, and tables.
      - Top-level KPI Summary Cards for Revenue, Payroll % of Revenue, EBIT, and Adjusted EBITDA, showing Actual vs. Budget variance badges (green for favorable, red for unfavorable).
      - Four interactive charts: (1) Revenue vs. Budget by Campus, (2) Revenue-to-Adjusted EBITDA Waterfall/Bridge, (3) Payroll vs. Non-Payroll Expense Trend, and (4) Campus Margin Comparison.
      - A clean, executive-formatted Financial Summary Table below the charts showing the exact numbers from the Excel file."

    • Participants double-click Brightmoor_Executive_Dashboard.html in their folder—it opens immediately in their web browser.
3B. Tuning, Customizing Charts/Filters & Publishing Internally (1:40 – 1:55 | 15 mins)
  • Hands-On Tuning (Addressing Dana's Exact Questions from the Prep Call):
    • Dana specifically noted: "I can already generate a basic dashboard, but the most important thing is: how do we change a specific chart (e.g., convert a default vertical chart to a horizontal bar chart), how do we add filters, and how do we publish it?"
    • Everyone runs 3 targeted Tune prompts in Cowork:
      1. Changing Chart Types & Sorting:

        "Look at the Campus Revenue vs. Budget chart. Convert it from a vertical column chart into a horizontal bar chart sorted from highest positive budget variance at the top to lowest at the bottom, and display the exact $ Variance label at the end of each bar."

      2. Adding Multi-Dimensional Filters & Toggles:

        "Add three filter controls in the top header bar that work alongside the left-hand Campus sidebar: (1) a toggle to switch the entire dashboard between February (Month) and YTD (Jan + Feb), (2) a Cost Category filter (All, Payroll Only, Non-Payroll Only), and (3) a 'Highlight Material Variances (>5%)' checkbox that filters the bottom table to only line items exceeding our materiality threshold."

      3. Making It Refreshable Every Month Without Rebuilding:

        "Add a small 'Load Monthly Excel File' button (or create a companion refresh_dashboard.py script I can trigger via Cowork) so that in March, when I drop 02_Brightmoor_Mar_Financial_Package.xlsx into this folder, the dashboard updates with March + Q1 YTD numbers in one click while keeping this exact layout."

  • How to Publish & Share Internally with Senior Leadership (5 mins):
    • Because Brightmoor_Executive_Dashboard.html is a single self-contained file with zero external server dependencies:
      1. Zero-IT Internal Share (Immediate): Drop the .html file into a shared OneDrive / SharePoint / Google Drive / Microsoft Teams folder or email it to Greg and the executive team—anyone who clicks it opens the full interactive dashboard locally in Chrome/Edge.
      2. Internal Intranet / Password-Protected Link: Show how IT or Finance can host the static file on an internal SharePoint page or password-protected internal site (e.g., Netlify/Cloudflare Access or internal company portal) in 2 minutes so leadership can log in from a link.
3C. Break (1:55 – 2:05 | 10 mins)
  • Quick bio/coffee break.

Module 4: Build 3 — Automating Monthly PowerPoint Reports & Charts (2:05 – 2:50 | 45 mins)

4A. Low-Hanging Fruit: Mirroring & Scaling Charts Across Slides (2:05 – 2:20 | 15 mins)
  • The Scenario (Directly from Dana's Call): Every month, someone spends ~8 hours copying tables and building repetitive charts across a 15-page PowerPoint deck. First, let's tackle Dana's "low-hanging fruit" ask: taking an existing slide/chart style and automatically replicating it across multiple pages.
  • Hands-On Exercise:
    • Inside Desktop/Brightmoor-Finance-Workshop/03_PowerPoint_Templates/ is Brightmoor_MBR_Template.pptx, which has Slide 1 (Title) and Slide 2 (a formatted Consolidated Revenue & EBITDA chart + variance callout box).
    • Prompt (in Claude Cowork / Claude in PowerPoint):

      "Look at Slide 2 in Brightmoor_MBR_Template.pptx, which shows our Consolidated Revenue vs. Budget and Payroll/Non-Payroll chart layout. Mirror the exact layout, colors, and chart structure of Slide 2 and automatically generate 6 new slides right after it—one for each Brightmoor Campus (Bay Area, Austin, Chicago, New York, Boston, Online)—using the campus data from 02_Brightmoor_Feb_Financial_Package.xlsx. Save the result as 03_Brightmoor_Campus_Charts_Deck.pptx."

    • Open the resulting PowerPoint deck: 6 campus slides generated in 45 seconds with native editable PowerPoint charts/tables.
4B. Guided Build 3: Translating 5 Excel Tables into a Full Monthly Deck + Executive Commentary (2:20 – 2:50 | 30 mins)
  • Step 1 — Talk & Jam (10 mins):
    • Moving from single charts to the full monthly reporting workflow: taking 5 core financial tables in 02_Brightmoor_Feb_Financial_Package.xlsx and populating a 10-slide Monthly Business Review (MBR) presentation (04_Brightmoor_Feb_Monthly_Review.pptx).
    • Prompt:

      "Using 02_Brightmoor_Feb_Financial_Package.xlsx and Brightmoor_MBR_Template.pptx, generate our complete February Monthly Financial Review deck (04_Brightmoor_Feb_Monthly_Review.pptx).
      Translate these 5 Excel schedules into formatted slides with both tables/charts and a 'Key Takeaways' commentary box on every slide:
      1. Executive Summary Scorecard (Feb & YTD Revenue, EBIT, Adjusted EBITDA vs. Budget).
      2. Revenue Variance Breakdown by Campus (Feb & YTD table + horizontal variance bar chart).
      3. Revenue-to-Adjusted EBITDA Bridge (showing Payroll vs. Non-Payroll OpEx, D&A, and adjustments).
      4. Payroll vs. Non-Payroll Deep Dive (highlighting overtime, benefits, and campus staffing variances).
      5. Accounting Close & Bursar AR Summary (reconciliation status and key month-end flux items).
      For the 'Key Takeaways' bullets on each slide, write concise, CFO-ready commentary citing the exact $ and % variances from the Excel file (both February and YTD). Do not invent any reasons not supported by the variance tab."

  • Step 2 — Make & Realistic Iteration / Tuning (20 mins):
    • Setting Realistic Expectations (As discussed with Dana on the prep call):
      • Show the generated deck (04_Brightmoor_Feb_Monthly_Review.pptx). Point out that while the numbers, tables, charts, and commentary are 90% there in 60 seconds, PowerPoint automation always has small visual quirks on the first pass (e.g., a table font is 14pt instead of 10pt, or a 4-bullet takeaway box overlaps a chart).
    • How to Fix PowerPoint Formatting Fast (The Screenshot Tune Loop):
      • Instead of manually reformatting 15 slides or writing long descriptions, participants take a screenshot of Slide 3, paste it directly into Claude Cowork, and say:

        "Look at this screenshot of Slide 3. The table font is too large and overflows the bottom border, and the Key Takeaways box is too narrow. Update the script to set table cells to 9pt Arial, widen the commentary box to 3.5 inches, and re-export 04_Brightmoor_Feb_Monthly_Review.pptx."

    • Locking In the Repeatable Monthly Automation:
      • Once the slide layout looks right, ask Cowork:

        "Save the exact script and formatting rules you just used into our folder as build_monthly_deck.py along with a simple checklist in MONTHLY_RUNBOOK.md so next month I can just drop in March data and say 'Run the monthly deck update using our saved script.'"


Module 5: Troubleshooting (ASK + Verify) & Role-Specific Sprint (2:50 – 3:30 | 40 mins)

5A. Troubleshooting When We're Not in the Room: The Finance ASK + VERIFY Ladder (2:50 – 3:00 | 10 mins)

What do you do next week when a script throws an error, a chart looks wrong, or a number doesn't tie?

  1. A — Ask with Visuals: Paste the exact error message or a screenshot of the broken Excel tab / HTML chart / PowerPoint slide into Cowork and say "Fix this specific issue without changing anything else."
  2. S — Simplify the Step: Don't ask Claude to clean 5 raw NetSuite CSVs, build a 6-tab model, and output a 15-slide PowerPoint in one massive prompt. Break it into discrete checkpoints: Step 1: Clean & reconcile in Excel $\rightarrow$ Verify totals $\rightarrow$ Step 2: Build HTML Dashboard or PowerPoint from the verified Excel tab.
  3. K — Kill Polluted Threads: If you've gone back and forth 8 times and Claude is getting confused, close the chat thread and start a fresh session. Because your files are saved on your computer in your folder, the new session starts clean using your latest Excel file.
  4. V — Verify Control Totals: Always include the phrase "Add a reconciliation check row using live Excel formulas comparing the output back to the raw input total" in every analysis prompt.
5B. Role-Specific Applied Workflow Sprint (3:00 – 3:20 | 20 mins)

Participants choose one high-value workflow tailored to their exact role (using provided NetSuite/Snowflake synthetic data packs in folder 04_Role_Tracks/, or their own sanitized exports):

Track A: Accounting, Payroll & Bursar (Majority of Room)
  • Option A1 — Automated Three-Way Reconciliation (Bursar / Payroll $\leftrightarrow$ NetSuite GL $\leftrightarrow$ Bank):
    • Task: Take 600 rows of Bursar student tuition/fee transactions and Payroll batch postings, match them against the NetSuite GL Detail, isolate timing differences vs. true discrepancies, and output a ready-to-upload NetSuite Correcting Journal Entry CSV + an audit-ready Excel reconciliation binder.
  • Option A2 — Multi-Period GL Anomaly & Flux Detector:
    • Task: Scan 12 months of NetSuite trial balances across all Brightmoor campuses to flag missing recurring accruals, vendor payments coded to the wrong department/campus, and expense lines spiking $>15%$ above their trailing 6-month average.
  • Option A3 — Repeatable Compliance & Accreditation Reporting Pack:
    • Task: Pull financial ratios, instructional vs. administrative expense splits, and tuition revenue schedules required for education compliance reporting into a locked, repeatable Excel template with automated tie-out checks.
Track B: Strategic Finance / FP&A (Dana's Team)
  • Option B1 — Snowflake Dump Cleaner & Multi-Scenario Driver Model:
    • Task: Take a raw, unpivoted Snowflake query export (campus enrollment cohorts, credit hours, tuition tiers, faculty headcount), clean and structure it via Python, and generate a dynamic 3-scenario (Base / Upside / Downside) Excel financial model with live sensitivity tables.
  • Option B2 — Customizing Brightmoor's Real Monthly Deck & Dashboard Pipeline:
    • Task: Adapt the HTML dashboard and PowerPoint generator scripts built in Modules 3 & 4 to mirror Brightmoor's exact monthly reporting calendar, campus hierarchy, and executive slide layout so it is production-ready for next month's close.
5C. Rapid Group Showcase & Monthly Close Playbook (3:20 – 3:30 | 10 mins)
  • Quick Demos (7 mins): 2 volunteers (1 from Accounting/Payroll/Bursar, 1 from Strategic Finance) share their screen for 3 minutes each to show what they built and one prompt trick that worked well.
  • Wrap-Up (3 mins): Everyone now has a complete local folder containing:
    • Cleaned & reconciled Excel workbooks (01 & 02)
    • An interactive, filterable Executive HTML Dashboard (Brightmoor_Executive_Dashboard.html)
    • Automated PowerPoint deck generators (03 & 04)
    • Reusable prompt templates (FINANCE_PROMPT_PLAYBOOK.md) they can copy-paste on Monday morning.

Module 6: Optional Buffer — 1:1 Clinic & Snowflake Setup (3:30 – 4:00 | 30 mins)

(Built into the 4-hour calendar block; attendees who need to get back to work can drop cleanly at 3:30.)

  • 1:1 Real-File Clinic: Michael and Natalia stay on to help team members run Cowork on their actual Brightmoor templates (assuming NDA is signed prior to Thursday).
  • Snowflake Connector Walkthrough (for Strategic Finance): For Dana's team (or anyone whose IT Snowflake provisioning is complete), walk through connecting Claude directly to Snowflake to pull and clean query results without manually downloading CSVs.

Part 3: What We Changed (Compared to Standard "Claude Code for Beginners" / Nvalo Run of Show)

  1. 100% Claude Cowork + Claude in Excel (Removed All Claude Code, Terminal & GitHub):
    • Completely removed CLI/Terminal setup, GitHub accounts/forks/commits, and developer-oriented framing. Everything happens inside local folders on their desktop using Claude Cowork and Claude in Excel (plus PowerPoint).
  2. Cut All Theory, Hype & Speculative Toy Projects:
    • Removed the "Why Now Is the Best Time to Build" / "Total Builders" philosophical intros and removed speculative builds (Vacation Planner, NPS survey, PRD-to-prototype).
    • Replaced the ASCII Map of Office Locations starter prompt with an immediate, practical NetSuite Multi-Campus CSV Consolidation & Control-Total Audit in Excel.
  3. Built Around Brightmoor's Exact 3 Requested Workflows:
    • Build 1 (Financial Analysis): Python-powered analysis in Cowork outputting native, formula-driven Excel workbooks covering both Accounting needs (Bursar/Payroll-to-GL reconciliations, MoM flux/trend detection) and FP&A needs (Revenue $\rightarrow$ EBIT $\rightarrow$ Adjusted EBITDA bridge, Feb + YTD Actual vs. Budget, rolling forecast assumptions).
    • Build 2 (Executive HTML Dashboards): Directly addresses Dana's workflow—turning Excel financials into a clean HTML dashboard with a left-hand Campus selector, plus explicit hands-on practice changing chart types (vertical to horizontal bar chart), adding interactive filters, and publishing/sharing internally with Senior Leadership.
    • Build 3 (Monthly PowerPoint Automation): Covers both Dana's "low-hanging fruit" ask (mirroring an existing chart across 6 campus slides) and the full workflow (translating 5 Excel tables into slides with Feb + YTD executive commentary), while setting realistic expectations around visual iteration using the screenshot-tuning loop.
  4. Tailored for Both Sub-Teams in the Room (85% Accounting/Payroll/Bursar vs. 15% Strategic Finance):
    • Removed peer TA volunteering (respecting Dana's comment that her team is too busy to TA).
    • Designed Build 1 and the Module 5 Sprint with explicit Accounting/Payroll/Bursar and Strategic Finance/FP&A tracks so both rookies in Accounting and advanced users on Dana's team get immediate value for their day jobs.
  5. Added Finance-Specific Data Validation (ASK + VERIFY):
    • Upgraded the standard ASK Ladder to include mandatory Control Totals & Reconciliation Check Tabs so accounting and finance staff know how to verify 100% mathematical accuracy without reading Python code.

Part 4: Synthetic Data & Pre-Work Package Specification (For Tuesday Internal Prep)

To make the exercises feel immediately authentic without requiring live Snowflake/NetSuite credentials on Thursday, prepare Brightmoor-Finance-Workshop.zip with the following NetSuite/Snowflake-shaped files:

1. Entity & Dimension Structure ("Brightmoor Education")

  • 6 Campuses / Cost Centers: 101 - Bay Area, 102 - Austin, 103 - Chicago, 104 - New York, 105 - Boston, 200 - Online Division.
  • Departments: Academic Delivery, Student Services & Bursar, Campus Facilities, Admissions & Marketing, G&A / Corporate.
  • Chart of Accounts (NetSuite-Shaped):
    • 4000 - Gross Tuition Revenue
    • 4100 - Student Lab & Technology Fees
    • 4200 - Scholarships & Tuition Discounts (Contra-Revenue)
    • 5010 - Faculty Salaries & Wages (Payroll)
    • 5020 - Admin & Bursar Salaries (Payroll)
    • 5030 - Overtime & Temporary Staffing (Payroll)
    • 5040 - Payroll Taxes & Employee Benefits (Payroll)
    • 5210 - Campus Lease & Utilities (Non-Payroll OpEx)
    • 5310 - Student Acquisition & Marketing (Non-Payroll OpEx)
    • 5410 - Instructional Software & Licensing (Non-Payroll OpEx)
    • 5510 - Bad Debt / Bursar AR Provision (Non-Payroll OpEx)
    • 6010 - Depreciation & Amortization (D&A)
    • 7010 - Campus Restructuring / One-Time Relocation Costs (EBITDA Add-Back)

2. Files to Include in Brightmoor-Finance-Workshop.zip

  1. 01_Raw_Exports/NetSuite_GL_TrialBalance_Jan2026.csv & NetSuite_GL_TrialBalance_Feb2026.csv
    • Columns: Period, Transaction_Date, Subsidiary, Campus_Code, Campus_Name (intentionally messy in raw file, e.g., "Bay Area", "101-BayArea" for Module 1 cleaning), Department, Account_Number, Account_Name, Journal_ID, Debit, Credit, Net_Amount, Memo.
  2. 01_Raw_Exports/NetSuite_Payroll_Register_Feb2026.csv
    • Columns: Pay_Period_End, Batch_ID, Campus_Code, Department, Employee_Category, Regular_Pay, Overtime_Pay, Benefits_&_Tax, Total_Payroll_Expense, GL_Account_Mapped.
    • Planted Anomalies for Module 2: Include a $+$38,000$ overtime spike in Austin and 2 unposted accrual batches totaling $$14,250$ so the reconciliation script catches them.
  3. 01_Raw_Exports/Bursar_Campus_Tuition_Subledger_Feb2026.xlsx
    • Columns: Student_Cohort_ID, Campus_Code, Program_Track, Gross_Tuition_Billed, Financial_Aid_Discount, Net_Tuition_Recognized, AR_Aging_Bucket, GL_Posting_Batch.
    • Planted Anomalies for Module 2: Include 3 late-month student billing batches in Chicago ($$22,400$) recorded in the Bursar subledger on Feb 28 that didn't hit the NetSuite GL until March 1.
  4. 01_Raw_Exports/Snowflake_FY26_Budget_Forecast.csv
    • Columns: Fiscal_Month, Campus_Code, Department, Account_Number, Budget_Amount, Prior_Forecast_Amount, Enrollment_Headcount_Assumption.
  5. 03_PowerPoint_Templates/Brightmoor_MBR_Template.pptx
    • Clean 3-slide master template in Brightmoor brand colors (navy, slate, teal) containing a Title Slide, an ExecutiveConsolidated Chart/Table reference slide (for the Module 4A "mirror across campuses" exercise), and a blank section layout.
  6. FINANCE_PROMPT_PLAYBOOK.md
    • Copy-pasteable prompt reference sheet containing all workshop prompts plus the ASK + VERIFY checklist so attendees don't fall behind typing during live builds.