If you manage client portfolios, this is what Shibui looks like in practice. The walkthrough below shows how one user built a custom screening framework from scratch, layered in compliance constraints, checked existing holdings against it, and scanned for new candidates. All in plain English, with 64 years of US market data behind it.
Build a custom screening framework
Users on the platform have built multi-criteria scoring systems entirely through conversation. No code, no formula language, no predefined dropdown filters. The process typically unfolds over multiple sessions and follows a recognizable pattern.
Start with a few criteria
The first session is exploratory. You describe what you care about and see what comes back.
"Screen all US common stocks for revenue growth above 15%, EPS growth above 15%, and positive free cash flow. Show me what passes."
30 US stocks passing all 4 criteria (sorted by market cap):
Symbol Name Rev Growth EPS Growth FCF Debt/Assets
───────────── ───────────────────── ────────── ────────── ──────── ───────────
NVDA.NASDAQ NVIDIA Corp +73.2% +96.7% $34.9B 5.3%
GOOGL.NASDAQ Alphabet Inc +21.8% +82.0% $10.1B 12.9%
AAPL.NASDAQ Apple Inc +16.6% +22.4% $26.7B 22.8%
MSFT.NASDAQ Microsoft +18.3% +23.3% $15.8B 8.2%
META.NASDAQ Meta Platforms +33.1% +60.4% $13.2B 22.0%
MU.NASDAQ Micron Technology +196.3% +762.7% $5.5B 10.6%
V.NYSE Visa Inc +17.1% +35.8% $2.6B 25.2%
AMD.NASDAQ Advanced Micro Devices +34.1% +220.7% $2.3B 5.0%
NFLX.NASDAQ Netflix Inc +16.2% +83.8% $5.1B 23.5%
... 21 more
The results come back as a ranked table. Some names make sense, others don't. That's the starting point.
Add criteria, tighten the filter
In the next session, you add constraints based on what the first screen was missing. Too many speculative names? Add a debt-to-assets ceiling. Too many stale positions? Require proximity to the 52-week high. Each criterion scores 0 or 1, and the total gives you a ranked shortlist.
"Add two more criteria: insider ownership above 15%, and price within 10% of the 52-week high. Score each criterion 0 or 1 and rank by total score."
"Also add a P/E ceiling of 40 to filter out the most speculative names."
Layer on compliance constraints
If your mandate has sector restrictions, exclusion lists, or other compliance requirements, those layer directly into the same screen. You don't need a separate compliance tool or a second pass through the data. One prompt handles both the investment criteria and the compliance filter.
"Exclude Financials and any company where debt-to-assets exceeds 50%. Separate the results into a 'pass' group (below 33%) and a 'review' group (33-50%)."
Financials excluded. Results split by debt compliance:
PASS (debt/assets < 33%):
Symbol Name Mkt Cap Debt/Assets
───────────── ───────────────────── ──────── ───────────
NVDA.NASDAQ NVIDIA Corp $4,822B 5.3%
GOOGL.NASDAQ Alphabet Inc $4,666B 12.9%
AAPL.NASDAQ Apple Inc $4,113B 22.8%
MSFT.NASDAQ Microsoft $3,077B 8.2%
META.NASDAQ Meta Platforms $1,540B 22.0%
MU.NASDAQ Micron Technology $610B 10.6%
AMD.NASDAQ Advanced Micro Devices $587B 5.0%
NFLX.NASDAQ Netflix Inc $389B 23.5%
... 14 more
REVIEW (debt/assets 33-50% — needs manual check):
AVGO.NASDAQ Broadcom Inc $1,997B 38.9%
LLY.NYSE Eli Lilly $909B 37.2%
TXN.NASDAQ Texas Instruments $255B 40.9%
APH.NYSE Amphenol Corp $175B 44.5%
APP.NASDAQ AppLovin Corp $142B 48.8%
... 15 more
Check existing holdings
Once the framework is defined, you run your current portfolio through it. Which positions still score well? Which have deteriorated since you bought them? This turns a one-time screen into an ongoing monitoring tool.
"Score my current holdings (AVGO, MSFT, JPM, UNH, LLY, PG, HD, COST) against the same 6-criterion framework. Flag any that fail on two or more criteria."
8 holdings scored against 6-criterion framework:
Symbol Name Score Rev EPS FCF Insider 52wHi P/E
───────────── ──────────────── ───── ──── ──── ──── ─────── ───── ────
AVGO.NASDAQ Broadcom 4/6 pass pass pass fail pass fail
LLY.NYSE Eli Lilly 4/6 pass pass pass fail fail pass
MSFT.NASDAQ Microsoft 4/6 pass pass pass fail fail pass
JPM.NYSE JPMorgan Chase 3/6 fail pass fail fail pass pass ⚠
UNH.NYSE UnitedHealth 3/6 fail fail pass fail pass pass ⚠
COST.NASDAQ Costco 2/6 fail fail pass fail pass fail ⚠
HD.NYSE Home Depot 2/6 fail fail pass fail fail pass ⚠
PG.NYSE Procter & Gamble 2/6 fail fail pass fail fail pass ⚠
⚠ = fails on 2+ criteria. COST, HD, and PG are defensive
holdings that don't fit a growth-oriented screen — expected.
Scan for new candidates
With the framework locked in, run a full universe scan to find stocks you haven't considered. Filter to only perfect or near-perfect scores, cross-check against the latest earnings data, and you have a shortlist of new candidates that meet every criterion.
"Run the full 6-criterion screen across all US common stocks. Show only companies that score 6 out of 6 and had a positive earnings surprise last quarter."
3 stocks score 6/6 with positive earnings surprise:
Symbol Name Sector Mkt Cap Score EPS Surprise
───────────── ───────────────── ───────────── ──────── ───── ────────────
APG.NYSE APi Group Industrials $19.8B 6/6 +5.3%
FSBC.NASDAQ Five Star Bancorp Financials $878M 6/6 +10.1%
UNTY.NASDAQ Unity Bancorp Financials $534M 6/6 +1.4%
Full universe screened: ~4,800 US common stocks.
Only 3 pass all 6 criteria AND beat earnings estimates.
Contact
Have a use case you'd like to discuss? chris@shibui.finance