Finance
21 research pieces on “Finance” — quoted word-for-word from the filing, with a link back to the source
Type the Theme, Not the Year: How to Ask AI About Thai Stocks So the Answer Comes From the Filings, Not From Your Question
Type the Theme, Not the Year: How to Ask AI About Thai Stocks So the Answer Comes From the Filings, Not From Your Question
Most people ask a filings AI the way they ask a search engine: a company, a year and a direction in one sentence — 'which companies were hurt by energy costs in 2568?'. On Boom Leverage AI that question quietly works against you. The year in the sentence does not filter anything, and the direction narrows the evidence before the AI has read a page. This is the workflow I use instead: type the theme, set years and scope in the filters, read the cited passage before the summary, and open the page before you rely on it. Three worked rewrites with the real passages each one returns, what every step costs in credits, and the traps that make a confident answer wrong. Figures and passages measured 11 October 2026.
Ask the Whole Thai Market a Question and Get an Answer You Can Audit — Boom Leverage AI, Where Every Line Ends in a Source You Can Open
Ask the Whole Thai Market a Question and Get an Answer You Can Audit — Boom Leverage AI, Where Every Line Ends in a Source You Can Open
Since 11 October 2026 every signed-in Boom Leverage Terminal account can ask a plain-language question across Form 56-1 filings, MD&A, auditor reports and Opportunity Day transcripts, and get a written answer in which every factual line ends in a numbered source — a highlighted PDF page, the original at the SEC, or the minute in the Opp Day recording. Lines that cannot be traced back are cut before you see them and the cut is counted on screen; if nothing survives, the answer is withheld and you are not charged for it. This is what it is, what it deliberately is not, how one question travels from scope to checked answer, what it costs, the five traps I would warn a model-validation committee about, and where it is still weak.
The Thai Earnings Call Nobody Reads: 5,755 SET Opportunity Day Sessions, Transcribed and Searchable Across the Whole Market
The Thai Earnings Call Nobody Reads: 5,755 SET Opportunity Day Sessions, Transcribed and Searchable Across the Whole Market
Every quarter, Thai listed companies stand in front of analysts at the Stock Exchange's Opportunity Day and answer the questions the Form 56-1 never has to address. The sessions are public, free, and effectively unread — hours of video with no transcript, no index, and no way to ask one question across every company that presented. We transcribed the last four years of them with AI, made them searchable beside the filings, and linked every passage back to the minute it was spoken. This explains what Opp Day is, why the format defeats ordinary research, exactly what our coverage number means and what it does not, the four rules our code enforces on every line we serve, and where the method is genuinely weak. Figures measured 21 September 2026.
Ctrl+F Doesn't Just Miss — It Finds the Wrong Thing, and on Some Files It Sees Nothing at All: Three Layers of Information Investors Lose in Thai Filings
Ctrl+F Doesn't Just Miss — It Finds the Wrong Thing, and on Some Files It Sees Nothing at All: Three Layers of Information Investors Lose in Thai Filings
Everyone knows Ctrl+F misses the euphemism. Fewer people know it also floods you with the right word in the wrong meaning — and almost nobody checks whether the PDF has any text in it at all. I opened the filings to count: 22.7% of Thai MD&A filings from FY2016–FY2020 (2,945 of 12,948) carry no searchable text layer, and in a 3,000-file sample from FY2021–FY2026 it is still 14.0%. This piece walks a real 18-page CENTEL filing where Ctrl+F finds zero characters, shows the OCR errors that make a naive fix dangerous, gives you a 10-second test for any PDF, and explains the rule the Terminal runs on: every result points back to the exact page, and anything that cannot be traced is dropped.
One Red Flag Was Never Enough: I Counted 104,153 Lines Thai Companies Wrote About Themselves, and All Six Lenses Lit Together Only 9 Times
One Red Flag Was Never Enough: I Counted 104,153 Lines Thai Companies Wrote About Themselves, and All Six Lenses Lit Together Only 9 Times
You found one danger signal in the accounts and sold — the stock ran for another two years. Next time you found the identical signal, told yourself that is just how the industry works, and eight months later it was carrying an SP sign. What separated the two is not in the signal. This article opens 104,153 lines of real management discussion and analysis from 896 Thai listed companies, FY2021–FY2026, through six operating lenses, and shows two places where a filing reader's instinct goes wrong: the most common topic holds the most severe material, and the rarest lens is the mildest of all (3.9%). Every line quotes the actual text with a page number and an SEC link you can check yourself.
The auditor warned for years and nobody read it — 68 Thai listed companies whose audit opinion was not 'clean' (FY2021-FY2025)
The auditor warned for years and nobody read it — 68 Thai listed companies whose audit opinion was not 'clean' (FY2021-FY2025)
One page in the Form 56-1 (One Report) is not written by the company. It is written by an outsider who saw the real books and signed under legal liability: the auditor's report. This piece hands you the Thai search strings to Ctrl+F straight into the Thai edition — lifted from real filings, not translated — then explains why knowing the strings still is not enough, because the report sits on a different page every year in every company and you need five years back to see direction. I scanned the whole market archive, 910 companies and 14,165 items, and counted how many did not get a clean pass, who was hit for consecutive years, and how far ahead of the news the warning landed. Every line quotes the actual text with page numbers and SEC links you can check yourself.
Latest
- Stop Guessing Which Sector Is Bleeding — Type the Symptom and See Who Raises a Hand
- Earnings Can Be Dressed Up. The Cash Has to Land Somewhere — 216 Disclosures Companies Wrote Themselves, on Pages Nobody Reads
- You lost the moment you didn't know what to type — institutions don't guess keywords, they run a checklist
- Funds Never Read a 56-1 From Page One — Here Is the Playbook for Where They Actually Open (a 4-Part, 9-Item Map + the 6 Places Everyone Skips)
- You Spend the Weekend on the Filings. The Funds Finished Trading Them Friday Afternoon — The Alpha Left for Retail Sits Where Machines Cannot Read
- The Launch Window Is Closed — Here Is What It Was For, and What the Free Tier Still Gives You
- The ฿400 Wage Headline Prints and You Dump Every Contractor — the Funds Went and Read Who Is Complaining and Who Quietly Swapped In Robots
- Your Portfolio Holds 8 Stocks and You Read the Whole Market's News — Fund-Desk Portfolio Surveillance, in One Click
- Reading Only the Latest Quarter Is Watching the Last Scene of the Film — How Institutions Trace a Management Narrative Across Years: PSL, FY2021–FY2026
- Retail Does Not Lose to Funds on Data Speed. It Loses on Memory — And Here Is What 4–6 Years of Memory Costs (Free / ฿799 / ฿1,399)
- Four Questions Institutional Analysts Use to Scan the Entire Market — and How to Point Them at Your Own Portfolio
- No Executive Ever Types 'The Company Is In Trouble' — Why Ctrl+F Has Always Lost Retail the Trade
- 80% of Financial Firms Now Run AI. Only 40% See Higher Profit — Is the AI Reading Your Filings Actually Improving Your Portfolio?
- AI Agents Are Now Moving Real Money Inside Banks — And It's the Same Risk Your Portfolio Runs Every Day
- I Built an NLP Bad-Debt Early Warning System for a Bank — the Lessons Retail Can Use on Stocks Today