How You Got Here: Tackling CFO Tech Legacy II
Why this is in the vault
Part 2 of the Tech Legacy arc. The Secret CFO walks the Seven Eras of CFO Technology - Manual Ledgers (1500+), Electronic Calculators (1960s), Spreadsheets (1979), ERPs (1992), Patchwork (2000s), Automation (2010s), AI (2020s) - and shows that each era's solution became the next era's constraint. Spreadsheets unlocked predictive finance but created shadow systems that persist today. ERPs promised unified backbone but most implementations stopped 90% of the way. Patchwork created integration chaos. Automation targeted symptoms not root causes. The load-bearing thesis: tech debt lives where business reality doesn't match system design, not in reconciliation processes or integrations - so layering more automation on broken process foundations multiplies the debt instead of paying it down. AI inverts the historical contract: instead of humans conforming to system requirements, technology can finally meet humans where they operate.
Mapping against Ray Data Co
This is the historical version of the harness-engineering thesis - it explains WHY the moat is the harness assembly rather than the platform underneath. Specific RDCO maps:
- MAC content arc: the 7-eras frame is portable straight into MAC content. Data engineering has its own version (Excel → Hadoop → Cloud DW → Modern Data Stack → Data Mesh → AI-native), and each era's solution became the next era's constraint. MAC can write the data-engineering 7-eras piece with this article as the structural template.
- The "tech debt lives at business-reality vs system-design intersection" frame is the diagnostic MAC's buyer needs. It reframes data quality from a technical problem (clean the data) to a process problem (the system can't capture how the business actually behaves). This is the language that gets data quality budget approved at the CFO level.
- RDCO's own infra discipline: each RDCO MCP / agent / workflow choice is an "era" decision and the article is the warning against optimizing for the current era's pain without seeing how it constrains the next. The agent stack I'm building today (MCP wrappers around legacy tools) is the equivalent of Patchwork - good enough for now, but the next era's constraint. The discipline question: am I building the harness or building the patchwork?
- Sanity Check angle: "the iPhone emerged despite the finance team" energy is in this piece's bones too - the right SC re-frame is "every era's solution becomes the next era's constraint" applied to AI agents. The original SC angle would be: when does the AI-agent era's solution itself become a constraint?
⚠️ Sponsorship
Sponsored by Maximor AI (recurring). Topic-matched. Editorial-vs-sponsor line is clean - the article's history doesn't bend toward Maximor's product specifically.
Related
- [[06-reference/2026-03-07-cfosecrets-shelfware-shenanigans-tech-legacy-i]] - Tech Legacy I, the diagnosis this part historicizes
- [[06-reference/2026-03-21-cfosecrets-bad-data-where-to-start-tech-legacy-iii]] - Tech Legacy III, the data-quality diagnostic
- [[06-reference/2026-03-28-cfosecrets-unbundling-the-erp-tech-legacy-iv]] - Tech Legacy IV, the unbundled-stack architecture
- [[06-reference/concepts/2026-05-10-harness-moat-two-layers-portability]] - the harness-moat thesis this history explains
- [[06-reference/2026-04-28-cfosecrets-finance-stack-of-the-future-unbundled-erp]] - the Mailbag-format unbundled-ERP companion