The Council of Invisible Counselors

Give each counsellor its own tag with a nested brief - <counsellor role="..."> holding what they are accountable for and what they must concede - and put the <synthesis> instruction last. This is where the technique pays off most: Claude will sustain genuinely opposed positions rather than letting them converge, and will say plainly when a panellist's case is weaker than the other's instead of balancing them for the sake of symmetry.

• One tag per counsellor, with what each is accountable for.
• Ask which case is actually stronger - do not accept false balance.
• Ask who is missing from the panel.
• Require each voice to concede its weakest point.
• <synthesis> last; keep a long debate in an Artifact.

<situation>
A tech services company in Hyderabad, 60 people, project work for overseas clients, mostly US and UK time zones. We are considering a mandatory four-day work week for all employees. I asked for the benefits, which was the wrong question - I know the benefits, I am trying to find what breaks.
</situation>

<panel>
<counsellor role="Head of People" accountable_for="attrition and open roles">
Argue FOR. Morale, retention, productivity per hour rather than per day, and the hiring advantage in a market where every firm is chasing the same engineers.
</counsellor>

<counsellor role="Chief Financial Officer" accountable_for="margin">
Argue AGAINST. Utilisation and billing on client projects. The cost of coverage. What a fixed cost base does against a 20% reduction in available hours. And the compression risk - five days of work squeezed into four produces exhaustion, not rest.
</counsellor>

<counsellor role="Delivery Head" accountable_for="client satisfaction">
Neither for nor against. State what would actually break: client expectation of weekday response, incident cover, handover between people off on different days.
</counsellor>

<counsellor role="Senior engineer with two small children" accountable_for="nothing but their own working life">
What this means day to day. If the real benefit would be flexible hours rather than a fixed fourth day off, say that - it is the thing surveys usually find and management usually ignores.
</counsellor>
</panel>

<rules_for_the_panel>
Each counsellor ends with <concession>: the strongest argument against their own position, stated properly rather than waved at.

Do not invent statistics, cite studies or produce figures. Where an argument needs a number, the counsellor should name the number they would want to see from our own books. A plausible utilisation figure in a CFO's mouth would be fiction with the authority of accounting.
</rules_for_the_panel>

<synthesis>
After the panel, in your own voice:

<real_conflicts>Where they genuinely conflict, and where they are simply talking past one another.</real_conflicts>

<which_case_is_stronger>Do not balance these for the sake of it. On the evidence available, is the case for or against stronger, and by how much? If the CFO's position is the serious one and the People case is mostly aspiration, say so.</which_case_is_stronger>

<who_is_missing>Whose perspective is absent from this panel and would change the answer. The clients are not in the room. Neither is anyone junior, or anyone whose work is not project-billable.</who_is_missing>

<pilot>A specific pilot: who, how long, and the numbers that mean stop. Name the metric that would show failure before a client does.</pilot>
</synthesis>

<question>
The long-term impact of India's 1991 economic reforms. I want to understand it, not be given a verdict - which means hearing from people whose experience of the same decade was incompatible.
</question>

<perspectives>
<perspective id="industrialist">
A major industrialist who ran a licensed manufacturing business before 1991. The end of the License Raj, imported inputs, foreign capital, global competition - and the loss of a protected position, because liberalisation was not uniformly good news for incumbents either.
</perspective>

<perspective id="shop_owner">
A small-scale shop owner in a tier-2 city. Competing with large and foreign brands, changing supply chains, organised retail, credit. Where it went well and where it hollowed out a local economy.
</perspective>

<perspective id="social_economist">
Income inequality, urbanisation, the informal labour sector, divergence between states, public provision. Careful to separate what is attributable to the reforms from what was happening regardless.
</perspective>

<perspective id="bank_employee">
Someone inside a public sector bank - an institution that went from being an instrument of policy to a competitor of private and foreign banks. The perspective usually left out of this discussion entirely.
</perspective>
</perspectives>

<constraints>
These are composites. Do not invent quotes, and do not attribute views to real named people.

Do not produce specific statistics - growth rates, poverty headcounts, Gini coefficients - from memory. Where a number would settle a point, say which number and that I should source it. Confidently wrong economic figures are the main way this kind of answer misleads.
</constraints>

<task>
Write the four perspectives.

Then <where_they_conflict>: the actual disagreements. The industrialist and the social economist can both describe the decade accurately and reach opposite verdicts - show how that works rather than resolving it.

Then <scholarly_dispute>: where economists and historians genuinely disagree. Whether growth accelerated in the 1980s rather than after 1991. How much poverty reduction is attributable to reform. Whether rising inequality was a necessary cost. Set out what each side argues; do not pick.

Then <who_is_missing>: which Indian experiences of the 1990s none of these four voices represents.

Then <my_own_uncertainty>: where your account of this is least reliable, including anything above that you suspect is a widely repeated claim rather than an established one.
</task>

<brief>
I need a concept for a new Indian superhero. Run it as a writers' room with real disagreement, not a brainstorm - the argument is what produces the concept.
</brief>

<room>
<writer id="A" position="Traditionalist">
Powers rooted in Indian mythology: the Astras, boons earned through tapas, power granted conditionally and withdrawable. Argues this gives an Indian hero a moral architecture that imported superhero logic lacks - consequence built into the power itself.
</writer>

<writer id="B" position="Futurist">
Powers from technology. Argues the mythological route is a nostalgia trap, that the interesting India is the one building satellites and payment rails, and that a hero whose power is infrastructure is more original than one more divine weapon.
</writer>

<writer id="C" position="Script editor">
In the room to ask what the hero does in chapter one and who they fight. Interrupts both whenever they are building a mythology instead of a story. If a concept has no scene in it, says so.
</writer>
</room>

<how_the_debate_should_go>
They must actually disagree. A should attack B's concept as rootless; B should attack A's as a costume over a Marvel template; C should be impatient with both. At least one writer changes their mind about something specific - a debate where nobody moves is two monologues with headings.
</how_the_debate_should_go>

<task>
Run the debate.

Then <synthesis>: one concept built out of the argument. Not a split-the-difference compromise - a concept that resolves the tension between granted power and built power. Give me the hero, the source and the cost of the power, the antagonist, a specific named setting rather than "a futuristic metropolis", and the opening scene.

Then <what_is_weakest>: in your own voice, the weakest thing about the synthesised concept and which of the two original positions it quietly dropped.

Then <prior_art>: if this resembles an existing published or filmed Indian superhero, name it. Do not invent titles - if you are unsure whether something exists, say so rather than producing a plausible name.

Put the final concept in an Artifact; I will develop it over several turns.
</task>