Artul.ai › Lab rankings
The lab rankings · 174,996 calls counted

What earnings calls actually say.

Literal counts from every transcript: how many times management said AI, how often analysts heard "great question", how many buzzwords per thousand words, who stonewalled, who talked longest. No opinions, just counting.

AI mentionsTimes management said AI, generative, LLM or agentic.
  1. NTSK165
  2. ASAN127
  3. ZM123
typical 0–0most, calls in the last 120 days →
"Great question"Management praising a question before answering it.
  1. INVH10
  2. FANG9
  3. GO8
typical 0–1most, calls in the last 120 days →
Thank-yousEvery thanks on the call, in both directions.
  1. GMAB87
  2. AMBA83
  3. KURA74
typical 13–31most, calls in the last 120 days →
BuzzwordsSynergies, headwinds, flywheels, best-in-class and friends.
  1. JOYY8.5
  2. ECX8.4
  3. JD8.3
typical 0.6–1.8most, calls in the last 120 days →
HedgesSort of, kind of, roughly, I think, probably.
  1. PR34.7
  2. PR31.3
  3. UTZ30.6
typical 5.5–10.5most, calls in the last 120 days →
SuperlativesRecord, unprecedented, tremendous, incredible.
  1. AAPL11.3
  2. USIO8.1
  3. AME7.8
typical 0.3–1.2most, calls in the last 120 days →
Outside forcesMacro, weather, currency, tariffs, supply chain.
  1. DE73
  2. BIRK68
  3. CRI67
typical 1–9most, calls in the last 120 days →
Stonewalls"Not going to comment", "we don't break that out".
  1. LEU9
  2. HUM9
  3. ASML6
typical 0–0most, calls in the last 120 days →
NumbersFigures per thousand words of management speech. Low means vague.
  1. FLO2.8
  2. DASH4.8
  3. GPOR4.9
typical 21.2–35.3least, calls in the last 120 days →
"Excited"Excited, exciting, thrilled.
  1. ROIV37
  2. RPRX32
  3. HIVE30
typical 0–3most, calls in the last 120 days →
Script shareShare of the call spent on prepared remarks rather than questions.
  1. PSKY92%
  2. HOV90%
  3. NIU89%
typical 35%–55%most, calls in the last 120 days →
CEO shareHow much of the talking the CEO did.
  1. FTLF100%
  2. SLE100%
  3. DGXX100%
typical 29%–68%most, calls in the last 120 days →
Longest answerThe longest single management answer in the Q&A.
  1. RH3,299
  2. PAR2,133
  3. VSAT2,096
typical 249–461most, calls in the last 120 days →
Reading levelFlesch–Kincaid grade level of management answers.
  1. YB15.6
  2. FLX15.4
  3. FLX15.1
typical 7.6–9.6most, calls in the last 120 days →
PromisesWill, expect, on track, next year.
  1. BSX30.0
  2. CMI29.7
  3. HESM28.7
typical 8.9–13.6most, calls in the last 120 days →
Typical answerAverage length of a management answer.
  1. HIMX712
  2. WB692
  3. AEVA681
typical 80–150most, calls in the last 120 days →
Call lengthTotal words spoken on the call.
  1. LYG28,052
  2. PAG19,784
  3. VSAT18,660
typical 4,868–8,679most, calls in the last 120 days →
FlatteryCompliments, thank-yous and first names between analysts and management.
  1. ANGO44
  2. AVPT43
  3. GMED43
typical 17–30most, calls in the last 120 days →
DodgingQuestions declined, deferred or answered with a different question.
  1. KURA14
  2. ALC14
  3. VFF14
typical 5–9most, calls in the last 120 days →
BlameResults attributed to weather, macro, currency or other outside forces.
  1. GFL26
  2. EPC26
  3. FOUR25
typical 3–12most, calls in the last 120 days →
HypeSuperlatives and excitement in management language.
  1. LYTS11
  2. ETON11
  3. GLOB11
typical 2–8most, calls in the last 120 days →
Analyst pushbackHow hard the analysts pressed: follow-ups, two-part questions, challenges.
  1. STLA20
  2. BG18
  3. INGR17
typical 5–10most, calls in the last 120 days →
ConfessionAdmissions, apologies and self-corrections.
  1. BRSL10
  2. KLAR9
  3. EOG9
typical 1–5most, calls in the last 120 days →
Call theaterStage management: operator mechanics, scripted cues, "color".
  1. CAVA28
  2. EPAM28
  3. TROX28
typical 12–19most, calls in the last 120 days →
Capital returns talkAttention paid to buybacks, dividends, M&A and the balance sheet.
  1. HRL7
  2. LOW7
  3. GFI7
typical 1–5most, calls in the last 120 days →

How to read this. Every number is a literal count from the transcript, computed the same way for every company since 2016. "Typical" is the middle half of all calls. Management language is measured on the executives' words; thank-yous count both sides. Calls shorter than 2,500 words are excluded.