Executive Overview
The DFX5 Ops library is a well-structured, full-company operating system spanning roughly six years (2020–2026). It is organized into clear functional areas — Marketing, Finance, HR, Sales, Operations, and a large client/legal document store — and reflects an organized, consistently-maintained back office. Below is the high-level shape; later tabs go deep on the client portfolio, the people/contract roster, an activity-based productivity proxy, recommended next tasks, and a plan for the operations app.
~1,200
Total files
across 10+ functional areas
~26
Clients / projects
LatAm + Caribbean
~50
People under contract
8 offers + ~42 PSAs
2020–2026
Date span
most active 2023–2026
~80
Purchase orders
standardized & filed
~50
AWS fund requests
POC / MAP packages
Files by area
Library composition by file type
Image/video files dominate by count (event marketing); the highest business value sits in the PDF/XLSX/DOCX layer.
Headline: This is not a messy folder to clean up — it's a maturing operations backbone. The biggest opportunities are (1) de-duplication of event media and triple-filed POs, and (2) turning the static spreadsheets that already act as registers (clients, contracts, receivables, pipeline) into a single operations app with live status. See the Ops App Plan tab.
What's Here — and What's Interesting
Eight things stood out while mapping the library:
- 1A real legal/commercial backbone.
DFX5 FILESholds per-client SOWs, MSAs, CSOs and ~80 standardized Purchase Orders, plus anOfficial Agreementstree with a clean05_Purchase_Order/10_Fund_Request/12_Vendor_Services_NDAtaxonomy. This is the highest-value, most "auditable" part of the workspace. - 2~50 AWS Fund-Request packages (POC / MAP) under
10_Fund_Request— each with calculators and signed CSOs. This represents a significant amount of AWS co-funding activity being tracked manually. - 3An events engine. Marketing 2023→2026 documents a heavy cadence of summits and roadshows (CX Summit, AWS Community Days, reInvent, Partner Summits, Quito/Lima/Bogotá/Miami) with budgets, agendas, photo/video libraries and LinkedIn content plans.
- 4Finance is genuinely modeled, not just filed: Annual Revenue 2023/24/25, AR Pro Forma, Outstanding Receivables, AWS cost-vs-income analysis, QuickBooks data, and ~25 per-opportunity Financial Questionnaires.
- 5Forward-looking planning. A dedicated
Business Plan 2026with a 3-year partner projection, QBR/all-hands decks, and an updated org chart — the company is planning, not just recording. - 6Process maturity signals:
DFX5 PROCESSES(Process Triggers, "AI Agents DFX5"),DFX5 JIRAnotes, anOnboardingflow, an AWSSCORECARDtracker, and aLoopfolder of meeting transcripts + recordings (May–Jun 2026). - 7Several "register" spreadsheets already exist —
DFX5 Files database.xlsx,Projects Data Base,Employee Contracts.xlsx,Customer Tracker / Outstanding Receivables,Committed Pipeline. These are exactly the tables the ops app should absorb. - 8A global, blended workforce — contractor PSAs span Latin America, South Asia and Eastern Europe, which makes a single HR/contract register more valuable (multi-country renewal tracking).
Client Portfolio (~26)
Client/project folders under DFX5 FILES/Projects and Clients. Larger engagements use a Signed / Pending / PO / CSO sub-structure; smaller ones are flat (1–2 loose files).
| Client / Project | Folder maturity | Region signal |
|---|
Engagements are concentrated in banking, insurance and retail (Itaú, Interbank, Oriental Bank, BHD, BPD, Seguros Humano, Seguros Múltiples, Rappi, Evertec) around Amazon Connect contact centers, GenAI assistants, and cloud migration/modernization.
People & Contracts
~50
People under contract
8
Employment offers (PDF)
~42
Contractor PSAs
Employment offers
Contractor PSAs (sample of ~42)
Gap worth closing: contracts are stored as individual PDFs with no single view of start dates, renewal/expiry, country, or rate. The
Employee Contracts.xlsx partially does this — a small HR module (see App Plan) would make renewals and compliance trackable.Activity Proxy — Reading the Output
How to read this (please read first). This is a rough activity proxy built from folder structure, file naming, and document date-stamps — it measures volume and breadth of output, not quality, effort, or hours. File-system "modified" timestamps were not used because OneDrive sync rewrites them in clusters, which would be misleading. Treat this as "what got produced and where," not a performance score. For a true productivity measure, connect a task system (see Next Tasks → offer to wire up Jira/Asana/Slack).
Output volume by area (file count)
Content presence by year (date-stamped docs)
| Area | 2022− | 2023 | 2024 | 2025 | 2026 |
|---|
High Medium Low · intensity = relative volume of dated files, not exact counts.
What the activity profile suggests
- ▸Consistent, broad coverage. Output appears across every functional area in every recent year — a sign of a back office that's kept current rather than letting areas go stale.
- ▸Sustained pace into 2026. The newest-authored material (pipeline meetings Apr–May 2026, Loop transcripts Jun 2026, Business Plan 2026, Marketing 2026) shows the workspace is actively maintained right now, not a historical archive.
- ▸Heaviest production in Marketing & Client Files. Events and client legal docs are the two largest output streams — consistent with a partner-led, event-driven GTM motion.
- ▸Strong filing discipline. Standardized PO/SOW/CSO naming and a consistent client sub-folder convention indicate process is being followed, not improvised.
Recommended Next Tasks to Assign
A concrete, prioritized list you can hand over. Grouped by effort. (Check items off as you assign them — resets on reload.)
Quick wins (hours)
Cleanup & hygiene (days)
Higher-value / sets up the app (1–2 weeks)
Want this as a shareable checklist for her? I can export it to a Word doc or, better, wire these into a real task board (Jira/Asana/Slack) so completion is tracked automatically — which also gives you the genuine productivity metric the Activity tab can only approximate.
Operations App — Proposed Plan
You asked the app to cover four areas: Projects & Clients, Tasks/EA work, HR & Contracts, and Finance & Invoices. The good news: the folder already contains the "tables" for each as spreadsheets — the app turns those static registers into one system with live status, ownership, and dates. Recommended stack matches your standard: Python FastAPI + React.
1 · Projects & Clients
Source: DFX5 FILES + Projects Data Base + Files database
clientengagementSOWPOCSOstatusvaluestart/end
Pipeline → signed → delivery view; links to the actual SOW/PO PDFs; "pending signature" alerts.
2 · Tasks / EA Work
New — the productivity layer
taskownerareaduestatuslinked_doc
Assign, track, and measure work. This is what converts the "activity proxy" into a real, fair productivity metric.
3 · HR & Contracts
Source: Contracts/ PSAs + Employee Contracts.xlsx
persontype(offer/PSA)countryratestartrenewaldoc
Renewal/expiry tracking across countries; one roster instead of ~50 loose PDFs.
4 · Finance & Invoices
Source: Finance/ + Customer Tracker + Expenses/AWS
invoiceclientamountduepaidAWS_costmargin
Outstanding receivables, AWS cost-vs-income margin, fund-request status in one place.
Suggested build approach
- 1Phase 0 — Ingest. A one-time script reads the existing register spreadsheets (clients, contracts, receivables, files DB) into the app database, so we start with real data, not an empty shell.
- 2Phase 1 — Read-only dashboard. FastAPI backend + React front-end surfacing the four modules as views over that data. Lowest risk, immediate value.
- 3Phase 2 — Tasks + status editing. Add the EA task layer and editable statuses (pending-signature, paid/unpaid, renewal-due) with simple alerts.
- 4Phase 3 — Integrations. Optional connectors to Jira/Asana, QuickBooks, and DocuSign (all available) so the app reflects live reality instead of manual entry.
Per your "report first, then decide" choice, this is the plan — say the word and I'll scaffold the FastAPI + React project and load it with the real register data as Phase 0/1.
Method, Confidence & Caveats
How this was built
The library was enumerated exhaustively by file type and by sub-folder. Counts are best-effort and shown with "~" where the enumeration hit listing limits. Areas, client list, contract roster, and date ranges are derived from folder/file naming.
Known limitations (please weigh these)
- !Spreadsheet/Doc contents were not opened this session. The analysis sandbox failed to start, so
.xlsx/.docxinternals (actual revenue figures, salaries, pipeline values) were not read. Everything here is from structure + naming. A follow-up pass can extract the numbers once the sandbox is available. - !File counts are inflated by duplication. Event photos are largely duplicated between per-event folders and a parallel
MARKETING/Photostree (true unique images ≈ half). Many POs exist in 3 places. Real unique-document count is meaningfully lower than ~1,200. - !The activity proxy is not a performance review. It reflects output volume/breadth and filing discipline — not quality, difficulty, or hours. Don't use it to score an individual; use it to spot where work is and isn't flowing.
- !"Modified" dates were deliberately excluded because OneDrive rewrites them on sync. Year buckets come from date-stamps inside file/folder names, which is more reliable but not complete.
Net confidence: High on structure, clients, contracts and areas; Medium on exact counts; Not yet assessed on financial/quantitative contents (needs the follow-up extraction pass).