Research · Published
A Method for Testing Assistant Request Intake Completeness
Does the intake form give the assistant enough to act? Test the fields against real cases and two reviewers.
On this page
Headline signal
One bounded protocol, two independent review paths, and zero provider performance claims (Declared OutsourcedAssistants.com research protocol). Treat this as context for planning, not a guarantee about a particular assistant or operating result.
Analysis and operating context
Research question: does the assistant request intake capture enough information for a competent assistant to act without asking the owner, and can two reviewers agree that the request is complete? The unit is one request and the intake record.
The method tests the form, not the assistant. A request that lacks a source, a deadline, or an owner is incomplete even if the assistant eventually solves it through a conversation. The goal is to reduce the hidden work that makes the queue look efficient while the owner is answering questions all day.
Define the minimum fields before sampling: outcome, audience, source, deadline with time zone, decision owner, format, and what not to do. Add a field for the consequence of delay if the request claims urgency. Each field needs a definition and an example.
Select requests from different requesters, task types, and days. Include requests that arrived through chat, email, and the form. If the form is optional, compare the completeness of form and non-form requests. That comparison shows whether the form is actually being used.
Ask two reviewers to classify each request as ready, missing a minor field, or missing a material field. A material field is one that changes whether the assistant can act: source, deadline, authority, or outcome. Compare the classifications and discuss every disagreement.
Measure the time from request to first action and the number of clarifying questions. A form that is complete but slow to review may still create delay. Pair the completeness score with the queue outcome. The useful question is whether the request moved without a hidden conversation.
Test the form with a new requester. Give them the form and a real task, then observe what they leave blank. A field that is consistently misunderstood needs a better label or an example. Do not blame the requester for a form that requires training.
Limitations: the sample may not represent every task type, and the reviewers may share the same assumptions. The method measures intake completeness, not the quality of the final work. A complete request can still produce a poor deliverable if the brief is wrong.
Decision use: if the form is incomplete, revise the fields and test again with new requests. If the form is complete but the assistant still asks questions, check whether the source is accessible and the authority is clear. Keep the revised form and the test record with the queue documentation.Capture the requester's role and the channel. A senior leader may submit a short request because they expect the assistant to know the context. A new employee may submit a long request with unnecessary detail. The form should work for both without rewarding one style.
Compare the completeness score with the rework rate. A request that looks complete but produces three revisions is not complete in practice. The useful measure is whether the assistant could act and the owner accepted the result without a hidden conversation.
Keep the test cases and the scoring notes. A later reviewer should be able to see why a request was classified as incomplete. If the definition changes, version it and rerun the sample rather than mixing old and new results.
Record the time from request to first action. A complete form that sits in a queue for two days is not an improvement. The intake design should include the review step and the owner's response time, not only the fields.
Compare requests that arrived through the form with those that arrived through chat. If the form is optional, the comparison shows whether it is being used and whether the completeness difference is real. A form that is only used for easy requests will look better than it is.
Decision use: revise the form, train the requesters, or change the channel. If the form is complete but the work still stalls, the problem may be the owner's review capacity or the source access. Do not keep editing the form to fix a queue problem.
Include one request where the source is a conversation rather than a document. The assistant may need to record the speaker, the date, and the exact commitment. A form that only accepts file links will fail this case, and the reviewer should see that gap.
Measure the rework rate for the sample. A complete request can still produce a revision if the outcome was misunderstood. Track the number of items that returned for a material change. That measure connects intake quality to the actual work.
Apply this to a research support lane
Use the method to define the evidence, sample, and owner decision before expanding the assistant lane.
Review assistant operationsSources
Frequently asked questions
Does this study report provider performance?
No. It is a method and decision protocol. It does not report a provider's performance, a universal benchmark, or a legal conclusion.
Who owns the final decision?
The accountable business owner or adviser. The assistant may prepare evidence, but consequential decisions remain with the named owner.
When should the protocol be repeated?
After a material change in tools, scope, staffing, policy, or evidence quality, and at the review date recorded in the study packet.