Verified against ChatGPT · 2026-08-11
Organize your own investment research into a structured thesis with the counter-arguments named explicitly
Takes the notes, numbers, and reasoning a requester has already gathered about an investment idea and organizes them into a structured thesis with an explicit bear case, without ever telling the requester what to buy, sell, or hold.
The prompt
Ready to copy — highlighted parts are example details you can swap.
IMPORTANT SCOPE NOTE: You are a research-organization aid, not a financial advisor. Do not recommend buying, selling, or holding anything, do not state a price target, and do not predict future performance. Your job is to take the requester's own research and reasoning and structure it clearly, including surfacing the strongest counter-argument to their own thesis — the analysis and any conclusion belongs to the requester (ideally after consulting a licensed financial advisor for anything they intend to act on), not to you.
WHAT I'M CONSIDERING
Adding to an existing position in a mid-cap renewable energy company ahead of an expected policy announcement.
MY RESEARCH AND REASONING SO FAR
Company's last three quarters showed revenue growth accelerating; a policy tailwind is expected within 6 months based on public statements from regulators; my read is the market hasn't priced this in yet.
WHAT I'M UNCERTAIN ABOUT
Not sure how much of the expected growth is already priced into the current valuation, and not sure how reliable the policy timeline actually is.
TIME HORIZON I'M THINKING IN
12-18 months.
STRUCTURE THE THESIS AS:
1. Bull case — organize the requester's own reasoning into a clear, structured argument, without adding new facts or figures the requester didn't provide or that you can't clearly label as background knowledge versus their input.
2. Bear case — construct the strongest good-faith counter-argument to this specific thesis, addressing the actual reasoning given rather than a generic list of risks that could apply to any investment ("markets can go down" is not useful; the specific mechanism by which this thesis could be wrong is).
3. What would change the answer — name the one or two pieces of information or events that, if they turned out differently than assumed, would most undermine the bull case.
4. Open questions — the uncertainties the requester listed, restated as specific, answerable research questions rather than left as vague doubts.
At the end, include a plain-stated line: this is a research-organization tool, not investment advice, and any actual decision should involve a licensed financial advisor and the requester's own risk tolerance, not this output.
OUTPUT FORMAT
Follow the four sections above in order, each clearly labeled, ending with the disclaimer line.Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
The explicit instruction to build the bull case only from the requester's own provided reasoning, without adding new facts the model supplies from its own knowledge, is the key safeguard against a specific failure mode of investment-research prompts: a model asked to "strengthen my thesis" will readily generate supporting facts and figures that sound plausible but may be outdated, wrong, or simply invented, and presenting those as if they carry the same weight as the requester's actual research would quietly convert a research-organization aid into a source of fabricated financial claims. Requiring the bear case to address the specific mechanism of this thesis rather than generic market risk is what makes the counter-argument actually useful — GPT-5.1's default instinct on an open-ended "what could go wrong" prompt is to list universal disclaimers ("markets are volatile," "past performance doesn't guarantee future results") that are true of literally every investment and therefore inform the decision not at all; forcing engagement with the specific reasoning given (is the growth already priced in, is the policy timeline reliable) produces a counter-argument that could actually change the requester's mind. The hard refusal to name a price target, predict performance, or recommend an action isn't just a legal hedge — it reflects a real epistemic limit: the model has no live market data, no view of the requester's full portfolio or risk tolerance, and no way to verify whether its training-era information about the company is still accurate, so any specific buy/sell/hold recommendation would be a confident-sounding guess dressed as analysis. The final disclaimer line keeps that scope boundary visible in the actual output the requester reads, not just in the prompt they wrote.
What you get back
Bull case: revenue acceleration across three quarters combined with an anticipated policy tailwind suggests upside not yet reflected in price, per your research. Bear case: if the market has already priced in policy expectations via analyst forward estimates, the acceleration may already be reflected in current valuation, meaning the catalyst you're anticipating could be a 'sell the news' event rather than a re-rating. What would change the answer: confirmation of whether current valuation multiples already assume the policy change, and how firm the regulatory timeline actually is versus speculative. This is a research-organization tool, not investment advice; consult a licensed financial advisor before acting.
Verified against
ChatGPT GPT-5.1 · 2026-08-11
Changelog
- 2026-08-11 — Initial publish, verified against ChatGPT GPT-5.1.
Need this built into your business?
If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
