The Problem with Spreadsheets
You rely on spreadsheets because they work well, mostly. But everything that surrounds them - data, versioning, testing - is a real pain.
Spreadsheet Sprawl Is a Real Hassle
Spreadsheets are ubiquitous because they're so useful, and yet "spreadsheet sprawl" – the work surrounding them – is a productivity killer. A recent survey of finance professionals uncovered the leading frustrations are:
- data consolidation
- manual entry/updates
- version control
- collaboration
- hunting for errors
The majority of business people spend at least six hours per month fixing Excel errors.
Your Team Has a System of Record But You Still Need Excel
Most professionals have a finance or accounting or CRM system that they're required to use as the source of truth and yet they use Excel to make reports and answer questions.
For example, a CFO may have a simple question:
- What were sales by region this quarter versus last quarter, excluding discontinued products, and which regions are more than 10% below forecast?
Conceptually, that’s straightforward. But the actual work may become:
- find data → export → clean → join → VLOOKUP/XLOOKUP → add calculated columns → pivot → filter → calculate variance → format → chart → check totals
The question took ten seconds to formulate. Producing the answer could take an hour.
The Rise of AI for Analysis
Into the sprawl has now arrived a huge time-saver: Claude, Gemini or ChatGPT. Suddenly we have "someone" who can do the grunt work – that's why enterprise AI is here to stay.
You simply drop your XL into AI chat and assign tasks or outcomes. And what you've put off for days is magically complete in minutes.
Where AI Falls Short
There's Always a Catch! AI does take work off your plate but, in exchange, AI introduces a new set of problems:
- AI chat sessions are hard to share with others
- Mostly it works but sometimes it fails
- AI chat runs up token usage costs quickly
- When the AI vendor "upgrades", it breaks things
So getting the most out of AI requires formulating tactics to address these. For example, emerging tactics include:
- Turn your AI Chat into an App
- Ask Claude or ChatGPT to document your process
- Build "agents" to dynamically self-correct