This user's frustration is exactly what happens when traditional spreadsheet tools hit their natural limit. Filters, IF statements, and nested lookups are the standard workarounds, but they weren't designed for the kind of relational matching this capacity-planning task demands. The problem isn't that the formulas are wrong; it's that the tool itself is asking the user to manually bridge gaps that software should handle automatically.
What this user needs is an AI-native approach that can interpret the intent behind the request: "Give me each employee's primary and secondary customers, along with their ticket numbers, and skip anything that doesn't match cleanly." A human would understand that request in seconds. A traditional spreadsheet, however, requires the user to translate that intent into a brittle chain of nested functions, and any misalignment, like missing variables or partial matches, produces errors rather than helpful feedback. The user's attempt to generate a list of customers with "false" values shows they were close, but the tool couldn't help them take the next logical step.
This is where AI transforms the experience. Instead of fighting with syntax, the user could describe the desired output in plain language: "Match customers to employees, split by primary and deputy roles, and bring ticket values along for the ride. Exclude any rows that don't match." An AI-native spreadsheet would parse that instruction, handle the matching logic, and return a clean table ready for further calculations. No error messages about missing variables, no manual filtering of false rows. The technology exists to make this workflow feel as natural as asking a colleague for help.
The takeaway is straightforward: When your spreadsheet starts fighting you instead of working for you, the solution isn't to learn more obscure functions. It's to use a tool that understands what you're trying to accomplish. This user's capacity-planning task is exactly the kind of multi-condition, relational problem that AI can simplify in seconds. The real question isn't whether it's possible; it's whether you're willing to explore an approach that puts the intelligence where it belongs, in the tool, not in the formulas you have to debug.