- Home
- Skills
- Research & Analysis
- Research Synthesis Brief
Works with the AI tools you already use
Research Synthesis Brief
Many sources in; one brief showing themes, where sources agree, where they conflict, and what none of them answer.
Free
Research Synthesis Brief
Example session with this skill installed
Synthesize these three sources on whether we should move our team to a 4-day work week. I need one picture, not three summaries.
SOURCE A - Pilot program report (nonprofit research group, 2025, 61 companies, 6 months):
- 56 of 61 companies continued the 4-day week after the pilot; 18 made it permanent.
- Self-reported burnout fell for 71% of employees.
- Company revenue was "broadly unchanged" over the trial period (sample of 23 companies that shared financials).
- Participation was voluntary; companies self-selected into the trial.
SOURCE B - Blog post by a SaaS CEO who tried it for one year (2024):
- Their 40-person team moved to 4 days of 8 hours (32 hours total).
- Customer support response times got worse in the first two months, then recovered after they staggered days off.
- Engineering output "about the same" by their sprint metrics.
- They reverted for the sales team after two quarters because pipeline dropped; kept it for everyone else.
SOURCE C - Survey of 1,000 HR leaders (consultancy, 2025):
- 22% say they are "considering" a 4-day week; 4% have implemented one.
- Top concerns: customer coverage (63%), fairness across roles (48%), measuring productivity (41%).
- Survey does not report outcomes for companies that implemented it.
- Read your context and instructions
- Compiled the research synthesis brief
- Generated the document
Should we move the team to a 4-day week? Synthesis of Sources A, B and C
Research question: Should our team move to a 4-day work week?
Sources: A, the pilot report (2025, 61 companies). B, a SaaS CEO's blog post (2024, one company). C, the HR leader survey (2025, 1,000 respondents). The sources are about a year apart, so none is stale.
Themes
1. Output holds up for internal "build" work, but it's measured loosely.
Revenue was "broadly unchanged" across the pilot [A], and engineering output was "about the same" [B]. Two independent sources point the same way. Both measures are weak, though. A's figure covers only the 23 of 61 companies (38%) that shared financials. B's figure comes from the company's own sprint metrics. C says this is the general problem: 41% of HR leaders cite measuring productivity as a top concern.
2. Customer coverage is the main operational risk, and staggered days off fixed it in the one case we have.
Support response times got worse for two months, then recovered once days off were staggered [B, single-source on the fix]. C independently ranks customer coverage as the #1 concern (63%). A doesn't break results down by role, so it can't confirm or rule out this risk.
3. Results depend on the role, and the obvious fix creates a fairness problem.
B reverted to 5 days for sales after two quarters because pipeline dropped [B, single-source]. That leaves a two-tier workforce, which is exactly C's #2 concern: 48% worry about fairness across roles. The sources fit together here. B's answer to the sales problem is the fairness problem C describes.
4. The wellbeing gain rests on one source.
Burnout fell for 71% of employees [A, single-source, self-reported]. B doesn't report wellbeing, and C reports no outcomes at all. This may be the strongest argument for the change, and it has the least corroboration.
5. Few companies have adopted it, and the success data comes from volunteers.
Only 4% of companies have implemented a 4-day week and 22% are considering one [C]. A's companies chose to join the trial [A]. A's success rates therefore describe companies already inclined to make it work, not companies in general.
Consensus
- Output for non-customer-facing work doesn't collapse over 6–12 months [A, B]. The two sources are independent, but both use soft measures.
- Customer coverage and role differences are the real implementation problems [B saw them happen, C reports them as concerns]. These are independent, but C's evidence is opinion, not outcomes.
Contradictions
| Question | One side | Other side | Notes |
|---|---|---|---|
| Does revenue suffer? | Revenue "broadly unchanged" [A] | Sales pipeline dropped and the team reverted [B] | A's trial lasted 6 months. B's drop showed up by the end of two quarters, and pipeline drops reach revenue later. A 6-month revenue window could miss an effect that started in pipeline. This is a hypothesis, not established. |
| How well does it stick? | 56 of 61 continued (92%) [A] | One team reverted [B] | B would probably count as "continued" in A's framing, since most of its staff kept the 4-day week. A's 92% may include partial reversions like B's. Also, only 18 of 61 (30%) made it permanent. The other 38 are still trialing. |
| Is it a proven model? | High continuation rate [A] | 4% adoption [C] | Not a real conflict, because the populations differ (volunteers vs all firms). The gap does show how much self-selection inflates A's numbers. |
Gaps
Our own team's role mix. This is the biggest gap. None of the sources can tell us what share of our team is customer-facing or sells, and that is the variable most likely to decide the outcome.
2.
What the model is. B cut hours to 32 (4×8). A doesn't say whether its companies cut hours or compressed them (4×10). These are different policies.
3.
Outcomes for companies that didn't volunteer. There are none: A is self-selected, B is one company, and C reports no outcomes.
4. Results past one year. Only B reaches one year, and that's a single company.
5. Objective productivity, attrition, hiring and cost data. No source has it.
6. Revenue for the 38 pilot companies that didn't share financials [A].
Bottom line
The sources don't show that a 4-day week is the right move for teams in general. They support a narrower conclusion. For internal build work, output seems to hold for 6–12 months, though the measurement is weak and the companies were mostly volunteers. Customer coverage and sales are the specific places it breaks. Coverage looks fixable by staggering days off, but that rests on a single company. Sales may not be fixable, and making an exception for sales creates the fairness problem HR leaders already worry about. The answer depends on how much of our team is customer-facing or selling. If the evidence supports any step, it's a time-boxed pilot: staggered coverage from day one, an explicit decision about sales roles, and productivity and pipeline metrics fixed before it starts. An outright switch isn't supported.
Appendix: source inventory
| Source | Type | Date | Main evidence | Bias or interest |
|---|---|---|---|---|
| A: pilot report, nonprofit research group | Primary data, but self-reported and partial | 2025 | 61 companies over 6 months: continuation rates, burnout survey, revenue for 23 companies | Self-selected companies; financials from only 38% of them; the research group may be advocating for the policy (unverified) |
| B: SaaS CEO blog post | Opinion / case study | 2024 | One 40-person team over one year: support, engineering and sales outcomes | Single company (n=1); its own internal metrics; the CEO is reporting on a decision they made |
| C: survey of HR leaders, consultancy | Survey of attitudes | 2025 | 1,000 respondents: adoption rates and ranked concerns | No outcome data; consultancies often sell change-management work (unverified) |
If you tell me roughly how the team splits across support, sales and build work, I can turn this into a go/no-go recommendation with a pilot design.
research-synthesis-brief.pdf
PDF · document
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
For analysts, consultants, researchers, and product people sitting on a folder of articles, reports, and transcripts. It merges them into one cross-source brief that answers your research question, instead of a stack of per-source summaries.
What you get
- Themes: 3-6 recurring threads, each citing the sources that support it.
- Consensus, with a note when "agreement" is really one origin quoted many times.
- Contradictions shown side by side with each side's evidence. It does not quietly pick a winner.
- Gaps: what your question needs that no source answers.
- A bottom line stated at the confidence the evidence actually supports.
- A source inventory (type, date, evidence, bias) in an appendix.
What it refuses to do
- Pad thin evidence into a confident conclusion. "The sources don't establish this" is a valid bottom line, and often the right one.
- Present a claim from a single source as settled. Those are marked
single-source. - Average sources years apart without flagging which may be stale.
- Start without a research question. If you don't give one, it asks.
What's in the zip
SKILL.md: the skill.references/recipe.md: the full step-by-step recipe with examples and variations.evals/: test cases you can run to check its behavior.
Pure instructions: no scripts, no network calls. From the open-source Claude Code Recipes collection.
How to install
Works the same in every agent - Claude, Cursor, Codex, Copilot and 20+ more.
- 1
Download the ZIP
Free skills download straight away. Paid skills unlock right after purchase.
- 2
Unzip into your skills folder
Every agent reads skills from one folder on your machine. Drop the unzipped folder in there.
- 3
Ask your agent to use it
Restart the agent if it was already running. It picks the skill up automatically - no config needed.
Skills folder by agent
Click the path to copy it. Create the folder if it does not exist yet.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill passes our trust checks below.
Security scanned
Passed our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scanned
Verified clean 1 day ago
- Free to download with an account