The Hidden Cost of the Spreadsheet Intake Interview

Many automation Centers of Excellence have one. It might live in SharePoint, Google Sheets, or a Teams folder. It often asks about process volume, handling time, exception rates, system names, and ownership. Its design and age vary by organization. For an explicit purchasing model rather than this organizational analysis, see IntakeOS versus forms and spreadsheets, including equivalent scope, residual human effort, and qualification assumptions.

The intake spreadsheet is the universal constant of enterprise automation. It's also one of the most overlooked sources of operational inefficiency in the CoE. Not because the questions are wrong - they're usually the right questions - but because the format of a static spreadsheet is fundamentally unsuited to the type of discovery that produces high-quality automation candidates.

What the Research Shows

We analyzed intake processes across 200+ enterprise automation programs, tracking time investment, data quality, and downstream project outcomes. The patterns were consistent enough to be striking.^[1]

The Incompleteness Problem

The most consistent finding was that spreadsheet intakes are systematically incomplete. In our sample, 31% of completed spreadsheet intakes were flagged as unusable or requiring significant re-work before development could begin. The most commonly missing data: exception rates (missing in 64% of intakes), system API availability (missing in 58%), compliance requirements (missing in 47%), and accurate volume data (off by more than 50% in 39% of cases where it could be verified).^[1]

The reason for this incompleteness is structural. A static form doesn't know what follow-up questions to ask. Illustrative example: if a business user answers '10% exception rate,' the form doesn't probe: 'What happens to those exceptions? Are they handled by a different team? Does the exception rate spike during month-end close?' That follow-up can be the difference between a clean development scope and a project that discovers additional process variants mid-build.

The Asynchronous Fill Problem

Rather than chase each spreadsheet owner for missing fields, use IntakeOS to conduct the initial business interview and organize the answers into a reviewable opportunity. Bring the existing row as a starting description, then inspect scope, exceptions, and workload assumptions with the participant. Keep the spreadsheet's useful history while moving new discovery into a connected qualification and mapping workflow.

In some organizations, spreadsheet intakes are completed asynchronously without an intake analyst present. Teams should validate locally whether that format produces optimistic or incomplete answers: users may describe how a process is expected to work rather than how it executes, and may not have verified exception rates, system availability, or average handling time.

The Prioritization Distortion

Because spreadsheet intakes produce inconsistent data quality, CoE teams that rely on them for prioritization end up with systematically distorted priorities. Processes submitted by sophisticated business owners with strong data literacy score higher on quantitative criteria - not because they're better automation candidates, but because they have better intake data. Genuinely high-value processes submitted with thin data get deprioritized or sent back for re-work, creating delays that reduce the overall ROI of the program.

The Alternative: Conversational, Adaptive Intake

The structural problems with spreadsheet intake aren't fixable by improving the questions. They're fixable by changing the format. A conversational intake - whether conducted by a skilled analyst or by an AI - naturally addresses the incompleteness and optimism bias problems. The conversational format enables follow-up. It can probe vague answers. It can ask the IDP sub-qualification questions when documents are mentioned, or the API availability question when a system integration is implied. It surfaces the exceptions and edge cases that a static form will never capture. This conversational discovery format is also the engine behind modern AI use case discovery - our guide on how to find agentic AI use cases shows the full 6-step method.

The value can be estimated with your own intake volume, analyst time, loaded labor cost, and measured rework. Avoid treating an illustrative calculation as a benchmark: any savings estimate should use local data and a stated automation-coverage assumption.