Home › Research

The research library

Every study below is computed from the same corpus: 165,000+ earnings-call transcripts our AI read and scored with an identical battery of questions, joined to a 22,449-call sample of 12-month market outcomes. We publish what the data shows — including when the answer is "no edge."

How we grade 165,000 earnings calls — and why we don't publish buy/sell

The methodology behind every page on this site: the 37-question battery, the A–F grade, the expected-move model, and the honest story of 1,600 tested hypotheses.

Call Signals

Critic Ammunition: What 165,000 Earnings Calls Show
Most earnings calls invite criticism, and most contain it: of 165,182 transcripts read by an AI, 145,051 — about 88% — show what we call critic ammunition, language that hands skeptics material to use against the company…
Confidence Proportionate to Evidence: What 165,000 Earnings Calls Show
Most executives claim their confidence tracks the evidence, but does it? This study isolates a pattern Artul.ai's models flagged across 165,182 earnings-call transcripts: calls where stated confidence appears proportiona…
Calls That Resolve Doubts: What 165,000 Earnings Calls Show
Most earnings calls leave questions dangling, but some do the opposite: they directly resolve the doubt an analyst just raised. When an AI read 165,182 earnings-call transcripts and answered the same battery of questions…
Guidance Worth Underwriting: What 165,000 Earnings Calls Show
Of the 165,182 earnings calls in Artul.ai's library, 118,164 — about 71.5% — contained forward guidance worth underwriting: statements specific enough to be checked later. That matters because guidance is where managemen…
Skeptic Reassured: What 165,000 Earnings Calls Show
When a skeptical analyst pushes hard on an earnings call and the management team answers with substance instead of deflection, something measurable changes. We call this pattern "Skeptic Reassured." An AI read 165,182 ea…
Results Worse Than Direction: What 165,000 Earnings Calls Show
Some earnings calls describe weak results but still point forward with confidence — a gap between what happened and where management says things are going. We call this pattern "results worse than direction," and the AI …
The Question Left Hanging: What 165,000 Earnings Calls Show
On nearly half of the 165,182 earnings calls in our corpus — 79,206 calls, or 48% — at least one analyst question goes unanswered or deflected without a real answer. We call these hanging questions. They matter because h…
Underused Fixed Costs: What 165,000 Earnings Calls Show
When a company says its fixed costs are "underused," it is arguing that extra volume would flow almost straight to profit. This study asks how often that framing appears in earnings calls and what tends to come with it. …
Calls That Read Rehearsed: What 165,000 Earnings Calls Show
Some earnings calls sound like they were written before the quarter happened. Across Artul.ai's corpus of 165,182 AI-read transcripts, 63,830 calls — about 39% — scored high on a "rehearsed" pattern: polished, promotion-…
Early Products Growing Fast: What 165,000 Earnings Calls Show
Some earnings calls describe a product that is still small but growing quickly — an early-stage offering scaling off a modest base. Artul.ai's AI read 165,182 earnings-call transcripts and asked the same battery of quest…
A Tiny Fraction of the Market: What 165,000 Earnings Calls Show
This study examines the roughly 30% of earnings calls — 49,623 of 165,182 transcripts — where the AI reader flagged a specific pattern, and asks what distinguishes these calls from the rest of the market. For anyone who …
Volume About to Step Up: What 165,000 Earnings Calls Show
Some calls sound like a company about to say more. Across a corpus of 165,182 earnings-call transcripts read by an AI, 47,036 calls — roughly 28% — carried the linguistic markers of a volume step-up: management leaning h…
Consolidation Among Peers: What 165,000 Earnings Calls Show
When a company's industry consolidates — mergers among peers, roll-ups, shrinking competitive fields — the language on its earnings calls changes. We asked an AI the same battery of questions across 165,182 earnings-call…
Pricing Recovering: What 165,000 Earnings Calls Show
Pricing recovering — calls where management signals that price increases are sticking or coming back after a weak stretch. It is one of the most-watched signals on any call, because pricing power is the cleanest lever on…
The Hidden Segment: What 165,000 Earnings Calls Show
Every earnings call has a moment where the numbers disappoint but the story stays rosy — and some companies live there. Across 165,182 transcripts read by AI, 34,917 calls (21.1%) fit a pattern best described as results …
Founder-Led Companies: What 165,000 Earnings Calls Show
Founder-led companies — those still run by the person who started them — account for 33,092 of the 165,182 earnings calls in Artul.ai's library, about one in five. Because founders speak with unusual authority and often …
When the CFO Dominates: What 165,000 Earnings Calls Show
On about one in seven earnings calls, the CFO dominates the conversation: 23,336 of the 165,182 transcripts in our corpus (14.1%) show the finance chief, rather than the CEO, carrying most of the Q&A and prepared remarks…
Scale-Dependent Advantage Claims: What 165,000 Earnings Calls Show
When a management team leans on scale as its competitive edge — network effects, fixed-cost leverage, data advantages — what does that say about how the call goes? Across 165,182 earnings-call transcripts read by an AI, …
Deferred Revenue Growing: What 165,000 Earnings Calls Show
When a company's deferred revenue is growing, management usually has something concrete to say about future business already booked. An AI read 165,182 earnings-call transcripts and flagged this pattern in 14,649 calls—a…
The Finished-Story Tell: What 165,000 Earnings Calls Show
Some earnings calls sound less like a conversation and more like a finished story — polished, complete, with no loose ends. We call this the finished-story tell. An AI read 165,182 earnings-call transcripts and flagged 7…

Business Verdicts

When Guidance Just Holds: Evidence From the Earnings-Call Corpus
Roughly half of all earnings calls in our corpus of 165,182 transcripts feature a company that simply maintained its guidance rather than raising, lowering, or withdrawing it. That group — 80,591 calls, or 48.8% of the s…
When Margins Are Expanding: Evidence From the Earnings-Call Corpus
When a company's margins are expanding, the tone of its earnings call tends to shift in measurable ways. Artul.ai's AI reader processed 165,182 earnings-call transcripts and scored each on a battery of dimensions; 73,217…
When Capex Ramps: Evidence From the Earnings-Call Corpus
When a company ramps capital expenditure, the language of its earnings calls changes. We flagged 64,183 calls out of 165,182 in the Artul.ai corpus — about 39% — where management discussed a capex ramp, and compared how …
When Demand Is Accelerating: Evidence From the Earnings-Call Corpus
When management teams describe demand as accelerating — orders stepping up, backlogs building, pricing recovering — the language leaves fingerprints. We asked an AI to read 165,182 earnings-call transcripts and answer th…
When the Backlog Is Building: Evidence From the Earnings-Call Corpus
When management talks about a growing backlog or deferred revenue, analysts lean in — order books are the cleanest forward indicator a company can offer. This study isolates 57,249 calls out of 165,182 in the Artul.ai co…
When Margins Are Contracting: Evidence From the Earnings-Call Corpus
When a company's margins are contracting, its earnings calls change in measurable ways. We had an AI read 165,182 earnings-call transcripts and answer the same battery of questions about each; 55,774 calls — about 33.8% …
Companies Raising Prices: Evidence From the Earnings-Call Corpus
When a company announces it is raising prices, the earnings call changes character: management is making a bet that customers will absorb it, and investors listen closely for whether that bet is working. Among 165,182 ea…
Companies Hiring Aggressively: Evidence From the Earnings-Call Corpus
When a company talks about hiring aggressively on its earnings call, it is usually signaling confidence about demand. We asked an AI the same battery of questions about each of 165,182 earnings-call transcripts, and 44,0…
Pricing Under Pressure: Evidence From the Earnings-Call Corpus
When a company starts talking about pricing on its earnings call, it is usually under pressure — from costs, competition, or customers. This study isolates that pattern: an AI read 165,182 earnings-call transcripts and f…
After Guidance Is Raised: Evidence From the Earnings-Call Corpus
When a management team raises guidance on an earnings call, the tone of the rest of the call changes in measurable ways. Among 165,182 earnings-call transcripts read by an AI, 34,779 calls — about 21% — included a guidan…
Companies Cutting Headcount: Evidence From the Earnings-Call Corpus
When a company announces layoffs on an earnings call, the tone of the whole conversation changes. We had an AI read 165,182 earnings-call transcripts and answer the same battery of questions about each; in 28,971 calls (…
When Capex Gets Cut: Evidence From the Earnings-Call Corpus
When a company cuts capital spending, the earnings call usually tells you why — or carefully avoids doing so. In Artul.ai's corpus of 165,182 AI-read earnings-call transcripts, 23,150 calls (14%) show capex getting cut. …
After Guidance Is Cut: Evidence From the Earnings-Call Corpus
When a company cuts its guidance, the call that follows is one of the most scrutinized moments in the corporate calendar. This study isolates 19,087 calls out of 165,182 in the Artul.ai corpus — about 11.6% of all transc…
When Demand Is Slowing: Evidence From the Earnings-Call Corpus
When a management team sees demand slowing, it changes how the whole call sounds — and this is one of the most common patterns in the corpus. Out of 165,182 earnings-call transcripts read by an AI and scored against the …
When the Backlog Shrinks: Evidence From the Earnings-Call Corpus
When a company's backlog shrinks, the language of its earnings call changes. Among 165,182 transcripts read by an AI, 7,699 calls — about 4.7% — discussed a declining backlog. Backlog is a forward-looking metric that ana…
After Guidance Is Withdrawn: Evidence From the Earnings-Call Corpus
When a company withdraws its earnings guidance, the call that follows is unlike almost any other. Out of 165,182 earnings-call transcripts read by an AI model, 4,393 calls — about 2.7% — came from companies that had form…

Management Behavior

Answers Full of Specifics: A Study of High Specificity on Earnings Calls
When an AI reader scored 165,182 earnings-call transcripts on a fixed battery of questions, nearly all of them — 161,829 calls, or about 98% — came back full of specifics: concrete numbers, named products, quantified dri…
Unusually Candid Management: A Study of High Candor on Earnings Calls
Most earnings calls follow a familiar script: rehearsed remarks, hedged answers, little that surprises. This study looks at the minority that break that mold — calls where the AI reader scored management unusually high o…
Maximum-Confidence Management: A Study of High Confidence on Earnings Calls
On earnings calls, confidence is usually the background hum: most executives sound reasonably sure of themselves. This study isolates the extreme end of that spectrum — calls where the AI reader scored management's expre…
Management That Doesn't Dodge: A Study of Low Evasion on Earnings Calls
Some management teams answer what they're asked. This study isolates the 86,188 calls (out of 165,182 AI-read transcripts) where the model scored evasion low — answers that engaged the question rather than deflecting it.…
Businesses Too Complex to Explain: A Study of High Complexity on Earnings Calls
Some companies are so sprawling—dozens of segments, unusual accounting, tangled geographies—that even a full earnings call struggles to explain them. An AI reader scored 165,182 transcripts for complexity, and 39% of cal…
Promotional, Hype-Heavy Calls: A Study of High Promotion on Earnings Calls
Some earnings calls read like marketing decks: heavy promotion, big confidence, sweeping market claims. An AI read 165,182 earnings-call transcripts and scored each one on a battery of dimensions; this study isolates the…
Management Admitting Uncertainty: A Study of High Uncertainty on Earnings Calls
Some management teams come to earnings calls and say, plainly, that they don't know what happens next. Across Artul.ai's library of 165,182 AI-read earnings-call transcripts, 18,704 calls — about 11% — show this pattern …
Highly Evasive Management: A Study of High Evasion on Earnings Calls
Some management teams answer almost nothing. Across 165,182 earnings-call transcripts read by an AI, 3,661 calls—about 2.2%—scored in the top band for evasion: deflecting questions, giving vague answers, and steering bac…
Management Under Visible Stress: A Study of High Stress on Earnings Calls
When management sounds visibly stressed on an earnings call — higher stress than roughly 98% of the 165,182 transcripts in our corpus — it tends to mark a genuinely rough quarter. Only 2,893 calls (about 1.8%) meet this …
Radically Simple Businesses: A Study of Low Complexity on Earnings Calls
Some companies describe their business in a way that reads as radically simple — few moving parts, easy to explain, low complexity. An AI reading 165,182 earnings-call transcripts flagged 861 calls (about 0.5%) matching …
Answers Without Numbers: A Study of Low Specificity on Earnings Calls
Some earnings calls answer almost every question without ever attaching a number to anything. Out of 165,182 transcripts read by AI, 431 calls — about 0.3% — scored this low on specificity, leaning on qualitative languag…
Guarded Management: A Study of Low Candor on Earnings Calls
Most earnings calls sit somewhere in the middle on candor, but a small group of management teams talk in a distinctly guarded register: low candor, high evasion, and low specificity all at once. Out of 165,182 transcript…

20-Year Trends

Are Calls Getting Easier to Criticize? A 20-Year Trend Across 165,000 Calls
When an AI read 165,182 earnings-call transcripts and scored each one on candor, evasion, specificity, and related traits, it flagged 145,051 calls — about 88% of the corpus — as containing at least one criticizable mome…
Are Skeptics Getting Harder to Reassure? A 20-Year Trend Across 165,000 Calls
When an analyst on an earnings call is openly skeptical, how does management respond? Our AI reader scored 165,182 transcripts and flagged 109,699 calls — about 66% — as containing a "hard-to-reassure" skeptic dynamic. I…
Are More Questions Going Unanswered? A 20-Year Trend Across 165,000 Calls
On roughly half of earnings calls, at least one analyst question goes without a real answer. Our AI reader flagged this "unanswered question" pattern in 79,206 of 165,182 transcripts — a 48% share. For anyone who listens…
Are Earnings Calls Getting More Rehearsed? A 20-Year Trend Across 165,000 Calls
Some earnings calls read like scripts: polished, evasive, heavy on promotion and light on candor. We flagged this "rehearsed" pattern in 63,830 of 165,182 calls — about 38.6% — using an AI that scored every transcript on…
How Often Is Volume Really About to Step Up? A 20-Year Trend Across 165,000 Calls
Every earnings call contains some version of the promise that volume is about to step up — new capacity coming online, a big contract ramping, demand finally catching up to supply. An AI reader scored 165,182 transcripts…
Is the Founder-Led Company Disappearing? A 20-Year Trend Across 165,000 Calls
Some public companies are still run by their founders, and their earnings calls sound different. Out of 165,182 transcripts read by an AI, 33,092 calls — about 20% — came from founder-led companies. Analysts and investor…
Is the CFO Taking Over the Earnings Call? A 20-Year Trend Across 165,000 Calls
On most earnings calls, the CEO still speaks more than the CFO, but the gap has been narrowing for a decade. We flagged 23,336 calls out of 165,182 in the corpus where the CFO's share of the conversation dominated, and c…
The Rise and Fall of Prepaid Growth A 20-Year Trend Across 165,000 Calls
When management teams on earnings calls talk up prepaid orders, early product traction, or how tiny their share of the market is, they are often describing a specific kind of growth story. We asked an AI to read 165,182 …

Signal Combinations

Small Base, Fast Growth, Huge Market: When Two Signals Appear Together
When a company pairs a small revenue base with fast growth in a huge addressable market, the combination leaves a distinctive fingerprint on the call. Among 165,182 earnings-call transcripts read by an AI, 35,112 calls —…
Raised Guidance That Even Skeptics Buy: When Two Signals Appear Together
Raising guidance is one of the most scrutinized moments on any earnings call, but not all raises sound alike. In a corpus of 165,182 earnings-call transcripts read by an AI, 31,112 calls (18.8%) paired a guidance raise w…
Founder-Led and Unusually Candid: When Two Signals Appear Together
Founder-led companies are often described as more candid, but pairing the two signals—founder in the CEO seat plus unusually direct language—yields a distinct group worth watching. In Artul.ai's library of 165,182 earnin…
Backlog Building Into a Volume Step-Up: When Two Signals Appear Together
When a company reports growing backlog and simultaneously describes an upcoming step-up in volume, management is essentially telling you that demand is arriving ahead of capacity. Among the 165,182 earnings-call transcri…
Founder-Led and Promotional: When Two Signals Appear Together
Among 165,182 earnings calls read by an AI, 18,659 calls — about 11.1% — combined two signals: a founder-led company and a promotional tone. Anyone who tracks earnings calls knows founders get a benefit of the doubt, and…
CFO Front and Center While Margins Fall: When Two Signals Appear Together
When a company's margins are falling and its CFO takes an unusually prominent role on the earnings call, that combination marks out a distinct slice of disclosure behavior. Among 165,182 transcripts read by an AI, 9,192 …
Dodged and Left Hanging: When Two Signals Appear Together
Some earnings calls combine two red flags: executives dodge a question, and the analyst who asked it never gets a real answer. Across 165,182 earnings-call transcripts read by an AI, 3,629 calls — about 2.2% — show this …
The C1 Pattern: Clean Calls With Growing Prepayments: When Two Signals Appear Together
When an AI read 165,182 earnings-call transcripts and scored each on a battery of questions, it found 2,343 calls — about 1.4% of the corpus — where two signals appeared together: clean, candid language and growing custo…
Evasive and Under Stress: When Two Signals Appear Together
When an AI read 165,182 earnings-call transcripts and scored each one, it found 657 calls—about 0.4% of the corpus—where high evasion and high stress appear together. That combination matters because each signal alone is…
Rehearsed and Vague: When Two Signals Appear Together
When an executive leans on rehearsed talking points while staying vague on substance, the combination is easy to miss in real time but stands out when measured at scale. Across 165,182 earnings-call transcripts read by a…

Hypotheses Tested

Hypothesis Tested: Identified near-term contributors not yet in the reported numbers
When a management team names a growth driver that has not yet shown up in reported results, listeners are being asked to underwrite the future rather than the present. This study tests that pattern: calls where the AI id…
Hypothesis Tested: Uncontested runway
Some companies spend their earnings calls describing a growth runway that nobody on the Q&A line contests — analysts don't push back, and management doesn't hedge. This study, "Uncontested runway," tests that hypothesis …
Hypothesis Tested: Coming out of the tunnel
Some companies spend a stretch of calls describing a downturn, then at some point start talking as if they have come out the other side. This study tests that pattern: an AI read 165,000 earnings-call transcripts and fla…
Hypothesis Tested: Answers go deeper than the script
When an executive stops reading the script and actually answers the question, does the call get better? To find out, we had an AI read 144,500 earnings-call transcripts from the 165,000-call corpus and score each one on …
Hypothesis Tested: Repeat expansion inside the base while new logos still arrive
When a company keeps expanding within its existing customer base while still landing new logos, that dual growth story shows up on earnings calls. Among 165,000 transcripts read by an AI, 186 calls matched this pattern o…
Hypothesis Tested: Earned edge, pressed harder
Some companies claim an edge they've earned — a durable advantage they can point to — and this study asks what happens when management presses that claim harder on the call. Among 485 calls in our 165,000-transcript corp…
Hypothesis Tested: Second demand front open and funded
When a company announces that a second demand front has opened and been funded — new customers or products beyond the core, backed by real investment — how does that show up on the call? We tested this hypothesis against…
Hypothesis Tested: This quarter could not have been described last quarter
Some earnings-call disclosures are genuinely new: the company says something this quarter that could not have been said last quarter, because the facts did not exist yet. This study isolates 142 such calls out of 318 fla…
Hypothesis Tested: Q&A has flipped from problems to possibilities, and management answers
When analysts ask about new products, untapped markets, or upcoming volume steps, management often answers with possibility rather than problems. This study tests that pattern across Artul.ai's corpus of 165,000 AI-read …
Hypothesis Tested: Existing customers are pulling more from the company than they were or
When an AI analyst read 165,000 earnings-call transcripts and asked whether existing customers are pulling more from the company than before, 126 calls out of a 438-call base matched the pattern — about 28.8%. That quest…
Hypothesis Tested: Front, middle, and back of the business all improving at once, in mana
When a management team reports that the front, middle, and back of the business are all improving at once — demand, operations, and financials moving together — it sounds like the strongest kind of earnings-call story. T…
Hypothesis Tested: Paid to expand
When a CFO takes center stage on an earnings call, the message is usually "we're spending to expand" — and it shows in the numbers. Across a corpus of 165,000 AI-read transcripts, 123 calls fit this "paid to expand" patt…
Hypothesis Tested: External validators converging
When multiple outside references converge on the same claim in an earnings call — analysts citing the same third-party data, or management pointing to the same industry sources — that's a pattern worth testing. Among 497…
Hypothesis Tested: Tone of discovery
Some earnings calls are organized around a discovery: management announces something it just learned — pricing is recovering, an early product is scaling faster than expected — and the rest of the call is built on that r…
Hypothesis Tested: Still getting better as they speak
Some executives warm up as the call goes on: their answers get more confident and more specific in the second half than the first. This study isolates that pattern — 106 calls out of 487 flagged by Artul.ai's AI reader, …
Hypothesis Tested: Engine explained, runway named
When a company explains why a metric moved and names the runway ahead, calls read differently: more candor, more specificity, less stress. This study tests that pattern across 165,000 earnings-call transcripts read by an…
Hypothesis Tested: Strong facts, held-back story
Some executives load their calls with hard specifics while keeping the narrative quiet — strong facts, held-back story. We tested this pattern across 165,000 earnings-call transcripts read by an AI, isolating 101 calls t…
Hypothesis Tested: Proven engine, fuel already in the tank
Some companies spend their calls describing a proven engine with fuel already in the tank: an established product line plus a large, mostly untapped market. We asked an AI the same battery of questions across 165,000 ear…
Hypothesis Tested: Substance without an audience
Some earnings calls deliver real substance — specific numbers, candid discussion, concrete guidance — but almost nobody seems to be listening: low analyst engagement, thin coverage, no audience. This study isolates that …
Hypothesis Tested: Fresh change already executed, benefit mostly still ahead
Some executives announce a change they have already made — a restructuring completed, a new product shipped, a cost program executed — and then frame the payoff as mostly still to come. This study tests that pattern acro…
Hypothesis Tested: Winning business away from competitors, with a stated reason why
When a management team says it is taking business away from competitors and names a reason why, that is a specific, checkable claim — and it is rarer than you might think. Out of 165,000 earnings-call transcripts read by…
Hypothesis Tested: The feared thing keeps not happening
On earnings calls, management often names a fear — a churn risk, a cost headwind, a regulatory question — and then, quarter after quarter, the feared thing keeps not happening. We tested this pattern across 165,000 AI-re…
Hypothesis Tested: Secure base, several live doors
Some companies face pressure from multiple directions at once yet keep answering from a position of security — candid about problems, specific about numbers, and clearly working several live strategic doors rather than o…
Hypothesis Tested: Running the company for a size the numbers don't show yet
Some companies talk about a scale they haven't reached yet — describing operations, capacity, or ambitions sized for a much bigger business than current numbers show. Among 497 earnings calls flagged for this pattern, 88…
Hypothesis Tested: Two-sided intensity
Some earnings calls read like a seesaw: strong promotion and confidence on one side, with candor and specificity on the other. This study tests whether that two-sided intensity pattern means anything. Out of 165,000 tran…
Hypothesis Tested: More where that came from
Some earnings calls leave investors expecting a follow-up — a pattern our library tags "More where that came from." An AI read 165,000 earnings-call transcripts and scored each one on the same battery of questions, letti…
Hypothesis Tested: Leaning into the storm
Some management teams respond to a bad quarter by confronting it head-on: naming the problems, quantifying the damage, and taking questions squarely. We tested whether that "leaning into the storm" posture shows up measu…
Hypothesis Tested: Acted like it's already bigger
Some executives talk as if the company is already much larger than it is — describing capacity, addressable markets, and volume as if scale were in hand rather than ahead. We tested that hypothesis on 165,000 AI-read ear…
Hypothesis Tested: Compounding evidence
Some earnings calls don't just show one good sign — they stack several at once: fast-growing early products, tiny market share, founder leadership, and volume about to step up. This study tests whether that compounding o…
Hypothesis Tested: Senior hire recruited
When a company announces a senior hire recruited from outside during an earnings call, it often signals a strategic pivot — a new growth push, a turnaround, or a bet on a capability the firm lacks. This study asks what e…
Hypothesis Tested: Demand has broadened across independent fronts
When a company says demand has broadened across independent fronts, it is claiming growth is no longer riding on one product, region, or customer group. That claim is easy to make and hard to verify, so it is worth knowi…
Hypothesis Tested: Candid stewardship of one big open question
Some management teams handle earnings calls by leaving one big question visibly unresolved while staying candid about everything else — acknowledging the open issue, quantifying what they can, and refusing to paper over …
Hypothesis Tested: Demand off the charts
When management teams describe demand as "off the charts" — signaling surging orders, capacity constraints, or backlog growth — what does that actually look like on the call? Our AI read 165,000 earnings-call transcripts…
Not investment advice. Artul.ai publishes AI-generated earnings-call quality grades and expected-volatility estimates — never buy or sell recommendations. We tested over 1,600 predictive hypotheses against 165,000 transcripts; the honest result, including what failed, is documented in our methodology.