When the spreadsheet becomes the system: signs it is time for case management

Nearly every business we work with has at least one spreadsheet that the company quietly depends on. It tracks complaints, claims, repairs, onboarding or orders. It started as a quick fix, it now has several owners, and nobody is quite sure which version is current. Now that AI tools are being layered on top of business data, that spreadsheet matters more than ever, because whatever sits in it is what the AI will work from.
Spreadsheets come up on the podcast more often than almost any other tool. Here is what we and our guests have said about when they stop being the right answer and what to move to.
Spreadsheets are great until they hold the process
In EP58 we went through ten signs that an organisation needs a case management platform, and the first was relying on spreadsheets for mission-critical work. They are on every machine and take minutes to set up. They also do not scale, and they are easy to break: several versions circulating, someone deleting a row by mistake, and an important decision made on the wrong numbers.
Our advice from that episode still holds as a starting point: "Go and find those spreadsheets. They will give you a clue as to where and what type of case centric automation you need to put in place." The spreadsheets people have built to keep track of work are a map of the cases your systems do not handle.
The other warning signs
Several other signs from EP58 tend to show up alongside the spreadsheets:
- Paper files for each case. Information on paper moves slowly and goes stale before it reaches the next person.
- Jumping between tools. Email, spreadsheets and other point solutions all open at once, with the risk of decisions made on incomplete information.
- Time spent searching. People hunting for the right document or the latest update instead of working the case.
- No consistent view for managers. A simple question such as how many cases are open, or who is working on what, becomes a research exercise.
- Weak audit trail. When information is scattered, showing that the right steps were taken on a case means digging through several systems. A proper case management system keeps an audit log as a by-product.
If three or four of these sound familiar, the spreadsheet has become the system.
Not every spreadsheet needs replacing
There is a middle ground. Henrique Cruz, who leads growth at Rows, described in EP106 how people use a modern, connected spreadsheet. He said the large majority use it for analysis, reports and lists. A small minority build real tools on top of it. His example was a manufacturing customer in South Africa that runs its operations through Rows: staff check parts inventory in the warehouse system and raise purchase orders through API connections, all from the spreadsheet. Henrique called this the frontier between spreadsheets and building internal tools.
That is a useful test. If the work is analysis and reporting, a better spreadsheet may be enough. If the spreadsheet is managing cases with owners, deadlines, approvals and a history, you are already building an application, just without the controls.
Henrique also told us that Rows took five years to reach 50,000 users, and that after launching an AI analyst that lets people talk to their tables, it passed a million users in under a year. People clearly want to ask questions of their data in plain language. The quality of the answers still depends on the data being in order.
What a proper platform makes possible
At the other end of the scale, Vijay Poondy of Yexle, an Appian partner, gave us an example in EP112 from financial services. Where firms had mis-sold products and the FCA required remediation, Yexle put a remediation case management platform live within 51 days, so the firm could triage and close cases for 250,000 customers. That volume, under that level of scrutiny, is far beyond what a spreadsheet can manage.
Most SMEs will never face a case load like that. The principle carries down, though: once work has to be tracked, assigned, evidenced and reported on, it belongs in a system built for cases.
The AI angle
In our Workflow Management 101 episode (EP31) we listed the signs that a business would be better off with proper workflow, and a team relying heavily on spreadsheets was near the top, alongside email-based workflows, paper still in use and staff copying data from one screen to another. Those same spreadsheets are what many businesses now want to point AI at, and an AI tool working from a stale or duplicated sheet will give confident answers from the wrong data. Moving case work out of spreadsheets into a structured system is one of the most useful things a business can do before it starts using AI on that work. It gives you one current record, clear ownership, and data that a person or an AI tool can trust.
A sensible first step
Start with an inventory. List the spreadsheets that track work, who owns them, how many people edit them and what would happen if one went wrong. Pick the one with the most people and the most risk, map the process behind it, and move that into a workflow or case management tool. Then do the next one.
Listen to the episodes
- EP58: Spreadsheets, Emails, Data Silos. 10 signs why you should consider a case management platform, Sascha Cutura and Arno van Rooyen
- EP106: Automation and Data, Spreadsheets on Steroids, with Henrique Cruz, Rows.com
- EP112: Inside Enterprise Automation, How Yexle Delivers Real Impact with Appian, with Vijay Poondy, Yexle
- EP31: Workflow Management 101 Part 1, Sascha Cutura and Arno van Rooyen
The free Automation Stack Field Guide helps you work out which kind of tool fits each process, from spreadsheets to full case management.

Pick the layer before you pick the product.
- An eight-layer map of the stack, plus the governance that sits across it
- When to reach for each layer, when to think twice, and what to ask the vendor
- A symptom table, vendor scorecard and quarterly checklist
Free PDF. Instant download.
More from the blog

Map the process before you hand it to an AI agent
Process mapping still decides whether automation and AI agents work. Six ways to capture a process, how to run the sessions, and where process mining helps.
Read article
AI agents in operations: where they earn their place
Where AI agents add value in operations and where ordinary automation is enough, with lessons from milo.ai, Garfield AI, DRUID AI and Turbotic.
Read article
Intelligent document processing: what invoice automation really involves
Intelligent document processing for invoices: why templates struggle, what good validation looks like, and how to build IDP into the wider workflow.
Read article