FROM SCOPE TO A USABLE DELIVERABLE
Complete examples of what a deliverable can look like. Read, download and discuss what your own context needs.
All samples are fictional and free to read and use for learning; they are not client work or measured outcomes. Tailored work requires a context review and agreement on outputs, timing and fees. A sample does not establish intervention success or decision validity.
One challenge, three options and a bounded trial with an owner, measure and review date.
How can a small team submit complete weekly reports on time without assuming training is the answer? This is a fictional team; every number and situation below is synthetic.
Decision owner: team lead. Review roles: reporting coordinator and one report preparer. Scope: the template and collection process for two weeks; this does not assess individual performance.
In the synthetic sample, 3 of 6 reports arrived on time and 4 needed at least one field completed. The template changed a week earlier; responsibility for collecting final inputs was undocumented. These observations do not establish a cause.
Check understanding of the template with two preparers, review available time and tools, and ask the team where delays occur. There is insufficient evidence to conclude there is a skill gap.
A - Train everyone: potentially useful if a specific skill gap emerges, but it uses team time and cannot by itself resolve a conflicting template or unclear responsibility. Defer commitment until the need is clearer.
B - Clarify the template and handoff: a short review of one example, agreed roles and an input deadline. Lower effort and easier to reverse, but feasibility still needs checking.
C - Automate collection now: it might remove steps later, but needs an approved tool, setup and permissions. Unclear responsibility could simply migrate into a new system.
Provisional choice: test B while filling the evidence gaps. It is not a proven solution; the lead decides after discussion.
Day 1: the coordinator and team review a completed fictional report and agree the source, preparer role and deadline for each field. Document exceptions rather than hiding them.
Weeks 1 and 2: apply the agreed template to 6 expected reports each week. The coordinator records completeness, timeliness, rework and approximate effort without names.
Illustrative review criterion: 5 of 6 complete, on-time reports each week, without an unacceptable increase in collection workload. This is a fictional target, not an achieved result.
Risks: cosmetic completeness that hides missing information, or shifting work onto one role. Mitigate with an explicit missing-information field and a brief workload check.
Day 14: the lead and coordinator review the evidence. Continue if criteria are met with credible data; adapt if gaps remain; explore focused training only where evidence supports it.
A complete activity linked to published episode 8, with facilitator guidance, a participant sheet and a worked example.
View published episode 8 and its script → · Episode companion
Suggested audience: university students or educators in groups of 3-4. Aim: distinguish a claim from its evidence and identify checks before sharing. This activity alone does not measure long-term behaviour change.
Source: published episode 8, “Did a human write what you saw?”. Preview it and choose an excerpt of no more than two minutes, or the full video if shorter. Prepare the synthetic cards below and a sheet with four boxes.
Card A - Fictional promotional post: “This tool raises every student’s concentration by 80% in one week.” It has polished copy and an attractive image, but no study link, definition of concentration or measurement method. Do not attribute it to a real organization.
Card B - Synthetic methods note for this exercise: ten volunteers rated their perceived focus after using a tool. There was no comparison group, equivalent baseline or performance test. This is not a published study.
Card C - Verification prompt: where is the original source, what does the number mean, who was observed, is there a comparison, and what interest might the publisher have? Polished language cannot answer these questions.
0-2: ask “What makes this post believable?” and collect two reasons without revealing a verdict. 2-4: show the selected episode excerpt.
4-7: groups read Card A and isolate the exact claim. 7-12: distribute B and C; each group completes its participant sheet.
12-16: two groups exchange sheets and challenge an inference that exceeds the evidence. 16-19: discuss whether to share or pause, and what is needed before relying on the claim. 19-20: complete the exit card.
Box 1 - Claim: every student’s focus improves by 80% in a week. Box 2 - Available evidence: promotional wording and a synthetic self-report from ten volunteers; the cards provide no inspectable original source.
Box 3 - Missing: a definition of focus, suitable measure, baseline, comparison and participant-selection method. Box 4 - Provisional action: do not rely on or share the number as fact; request the source and methods and explain that current evidence is insufficient.
For use: give groups a fresh sheet with the four box headings and ask them to complete it in their own words. More than one justified action is possible; guessing that content is AI-generated is not the learning target.
Facilitator checks: is the claim precise, is evidence separate from impression, is there a feasible verification question, and is the action proportionate? Record descriptive feedback for each criterion without ranking individuals.
Exit card: “Before sharing a similar claim I will check…; if the source is missing I will…”. At a later meeting, discuss a public example selected by participants without collecting their accounts, messages or personal data.
Two use cases, data and human-review rules, and explicit continuation and stop criteria.
This fictional team prepares reports and training material. The sample is not a security audit, compliance certificate or approved institutional policy. Decision: should the team trial two bounded uses with an institution-approved tool? Approval by the authorized owner comes first.
Scope: two weeks, two user roles and a human reviewer for every output. Use public or synthetic information only. Do not upload personal files, service enquiries or private operational information.
Input: a synthetic list of six public activities with descriptive dates and no names. Task: draft one page separating completed actions, open questions and next steps. Inventing figures, causes or outcomes is prohibited.
Check: the preparer compares every sentence with the input, removes unsupported material and records edits. The coordinator checks wording and figures and approves outside the tool. Suggested internal disclosure: “A tool helped draft this text; the team reviewed it before approval.”
Input: a public text permitted for reuse about source verification, a learning aim, audience age range and time limit. Task: propose a discussion question, exercise and participant sheet, linking every factual claim to a source passage.
Check: the facilitator reviews language, timing and accessibility, and independently checks claims and references. Rehearse instructions with a colleague before use. Do not enter learner responses, names or assessment data into the tool.
Before input: the user confirms that material is public or synthetic and that the tool and use are permitted. If uncertain, stop and ask the owner. Removing a name is insufficient if context still identifies a person.
The pilot owner approves scope and stopping; the user checks inputs; the reviewer verifies outputs; the coordinator keeps a brief log of time, edits and issues without copying sensitive inputs.
Outputs trigger no automatic action or decision and are not sent externally without review and approval. The pilot is not used to evaluate people or decide their eligibility.
Baseline: prepare three synthetic examples using the usual method and record drafting and review time. During the pilot, record equivalent time, checked factual claims, required edits, and accessibility or workload issues.
Fictional planning criteria: human review of every output, no unsupported claim left in approved text, no prohibited input, and acceptable total team effort. These are not achieved results or a promise of time savings.
Stop immediately if prohibited data is entered, a claim cannot be sourced, permissions exceed scope, or no human reviewer is available. Quarantine the output, inform the owner and follow institutional procedures; deleting a chat does not establish provider-side erasure.
End of week 2: review the log and three outputs from each use case. Decide whether to continue within scope, change the rules or stop. Hand over one use-case card each, input/output checklists and a decision record with evidence gaps and next steps.