🛠️ The Practical Professional
The first lens is the person who works with the topic every day. It researches recent sources, case studies, discussions among operators, and data from people on the factory floor. Its one mission is to expose the GAP between what the hands-on operator knows firsthand and what academics and commentators simply don't see — workflow friction, what actually works, and where it breaks.
🟡 New here? — “case study” and “operator × commentator”
- Case study: a detailed account of ONE real situation—a company, a project, an implementation—including what was tried, what worked, and what didn’t. It’s more valuable than an opinion because it shows the concrete outcome.
- Operator × commentator: the operator is the person who uses the thing day to day and feels where it gets stuck; the commentator is the person who talks about it from the outside. This lens trusts the operator.
📐 The 3 required returns—identical for EVERY lens
No matter the perspective: every lens returns exactly these three blocks. Memorize the format now — topics 2 to 5 only change the content.
- 1CENTRAL POSITION in 2 sentences — the lens's thesis, short and direct.
- 2STRONGEST EVIDENCE — 3 to 5 bullets, each with a concrete data point + named source + URL.
- 3THE ONE THING that only this lens would say—the insight no other perspective would offer.
Você é O PROFISSIONAL PRÁTICO para: {TOPIC} ({TOPIC_FRAME}).
Faça pesquisa real na web. Retorne EXATAMENTE:
1. POSIÇÃO CENTRAL em 2 frases.
2. EVIDÊNCIA MAIS FORTE — 3 a 5 bullets, dado + fonte + URL.
3. A ÚNICA COISA que só um prático diria.
Menos de 400 palavras.
See the perspective in action. Using "Are AI voice agents worth the investment?", the Practical might return something like this:
Position: work well in narrow, repetitive workflows; they break down in long, ambiguous conversations, where handing off to a human becomes the rule.
The only thing: "80% of the effort in a real deployment isn’t the voice model—it’s handling exceptions and routing. Anyone who has never run a call center underestimates this."
Illustrative example — in a real run, each evidence bullet would include a verifiable source and URL.
✓ What the Practitioner sees
- ✓Where the tool breaks down in practice
- ✓The invisible friction in the workflow
- ✓What operators do but no one documents
✗ Its blind spot
- ✗May overgeneralize from their own experience
- ✗An anecdote is not always rigorous evidence
- ✗That’s why the next lens exists: the Academic
🎓 The Academic
The second lens cares about peer-reviewed evidence and effect size, not with anecdotes. It combs arXiv, journals, university reports, and research institutes to answer: what does rigorous evidence REALLY say, and where does it contradicts the hype? And, crucially, it signals when the evidence is weak and reports the status—published or preprint.
🟡 New here? — three academic terms
- Peer review: before publishing, other experts in the field read and critique the study. It passed peer review → more trustworthy. It’s science’s “quality control.”
- Effect size: how much something actually changes, not just "whether it changed." "Reduced the cost by 3%" and "reduced it by 60%" can both be statistically real—but only one matters in practice.
- Preprint: a published study before peer review (e.g., on arXiv). It may be correct, but it hasn’t been checked by others yet — so it ranks lower in the hierarchy.
The same 3 returns from Topic 1, now filled in by the academic lens on "AI voice agents":
The gains measured under controlled conditions are real, but smaller than the marketing suggests; most strong studies measure narrow tasks, not open-ended customer support.
Peer-reviewed studies with a stated effect size > recent preprints (labeled as such) > isolated reports. Each bullet links to a named study + URL.
"Half of the numbers in circulation come from preprints or a single commissioned study—treating them as established fact is the most common mistake."
📊 Strong × weak evidence
The value of the Scholar isn’t just bringing studies—it’s label the strength of each one. A peer-reviewed causal finding carries more weight than a preprint, which carries more weight than an analogy. This hierarchy becomes the basis for the reliability scores you'll see in Track 3.
🔍 The Skeptic
The third lens starts from the assumption that the dominant view is exaggerated and builds the strongest possible counterargument. It looks for criticism, documented failures, contradictory data, regulatory changes, and rebuttals—to answer: what's the strongest counterargument, and what do proponents conveniently ignore?
⚖️ Rigorous, not petty
The Skeptic's golden rule: be rigorous, not "contrarian for sport." It doesn’t reflexively deny — it finds the strongest evidence that weakens the thesis and presents it with a source. Well-practiced skepticism is the best gift the report can receive: it shows you how you could be wrong before that gets expensive.
✓ Rigorous skepticism
- ✓Looks for real, named failure cases
- ✓Brings data that contradict the narrative
- ✓Builds the strongest counterargument, with a source
✗ Petulance disguised as skepticism
- ✗Deny everything without evidence
- ✗Choosing only the data that confirms the "no"
- ✗Mistaking an irritated tone for an argument
About "AI voice agents", the one thing only the Skeptic would say:
The only thing: "Demos impress with short calls and happy-path scripts. Churn reports and human escalation rates—which rarely appear in sales presentations—tell a different story."
Illustrative example. The Skeptic always cites the strongest reliable opposing source they can find.
💰 The Economist
The fourth lens has a motto: follow the money. It researches revenue, valuation, market size, funding flows, unit economics, and incentives to answer: who profits from the current narrative, and what financial incentives shape what gets researched and what becomes hype?
🟡 New here? — “valuation” and “unit economics”
- Valuation: how much the market thinks a company is worth (e.g., "startup valued at US$ 2 bi"). Note: this is an expectation about the future, not cash on hand—it may be inflated by hype.
- Unit economics: the profit or loss in a single transaction or customer. If each voice interaction costs more than it generates, the unit economics are negative—and no amount of volume can fix that.
🔢 Why the Economist asks for NUMBERS
The evidence for this lens differs from the others: each bullet needs a real number — revenue, valuation, market size, or funding—with a named source and URL. Opinions about money don’t count; figures with sources do.
What comes in and what the market projects
The total pot at stake
Where the capital comes from and why
Who benefits from keeping the narrative alive
The one thing only the Economist would say about "AI voice agents":
The “follow the money” insight: "Much of the enthusiasm is funded by companies that sell voice infrastructure. The incentive to inflate cost savings is structural—it’s worth checking who paid for the study before believing the number."
🏛️ The Historian
The fifth lens has seen cycles of disruption before and looks for patterns. It researches genuine historical parallels—earlier technologies, fads, market shifts—to answer: which parallels really fit, and what can we learn from how they played out? Who won, who lost, and what stabilized?
🧭 How the Historian reads a cycle
Finds the right parallel
Not just any pretty analogy — one that actually fits the mechanisms, not just the vibe.
See how it ended
Concrete dates and outcomes: who won, who failed, what remained when the dust settled.
Points out the invisible pattern
The one thing only they would say: the pattern no one else notices because they're looking only at the present.
🕰️ Example of the pattern
About "AI voice agents": "We've seen this with IVR systems in the 2000s and chatbots in 2016—the technology handles the easy case, the public gets frustrated with the hard case, and the market settles on a hybrid human+machine model. The pattern is likely to repeat." (illustrative, with dates and sources from an actual run)
🧠 Why each lens finds a gap
The five lenses aren't redundant: each has a different priority, so each one stumbles on a fact the others would miss. The Practical feels friction the Academic doesn’t measure; the Academic demands rigor the Practical forgoes; the Skeptic finds the failure both ignore; the Economist reveals the incentive behind it all; and the Historian shows we’ve seen this movie before. Together, they produce complementary coverage — what one misses, the other finds.
Each cyan box is a lens researching the SAME topic from a different angle. All five converge into a single briefing (green): where they agree becomes a likely finding; where they disagree, it feeds the contradiction map (Module 1.3). No single lens sees the whole picture — coverage comes from the sum.
The actual friction of using it
The rigor of the evidence
The overlooked failure
The hidden incentive
The repeated pattern
💡 Tip — all five run IN PARALLEL
The five lenses aren't run one after another. STORM launches the five agents (subagents) in a single message, and they work at the same time — without talking to each other. Each researches independently; only at the end do the five results come together. You’ll understand exactly how this works in the Module 1.3.
Self-recovery (optional): which lens "follows the money" — revenue, valuation, and incentives?
📌 Module summary
Next module:
1.3 — Subagents, contradictions, and verification