An earnings desk that covers the whole portfolio
Hawkclaw Capital Advisors covers a portfolio of listed companies with a small team, and earnings season arrives for all of them in the same fortnight, which is long enough to read three or four calls closely before the rest become history. We built them an earnings desk that reads all of them. Upload a call and the complete breakdown comes back without anyone sitting through it: red flags ranked by severity, forward guidance with direction and impact, every Q&A exchange attributed, and speaker-level tone and delivery.
Apple Inc.
Q1 FY2025 Earnings
Revenue
$124.3B
+4.0% YoY
Gross Margin
46.9%
+120bps YoY
Free Cash Flow
$27.1B
+8.1% YoY
Red Flags
Forward Guidance
Revenue Trend (B)
CEO Tone Analysis
Tim Cook
Confident
Measured delivery
Forward-looking emphasis
Deflected on China specifics
Tailwinds
Headwinds
- Covered every quarter, where before it was the three or four that fit the week
- Whole portfolioCovered every quarter, where before it was the three or four that fit the week
- Companies held to the same criteria, across three consecutive quarters
- 22+Companies held to the same criteria, across three consecutive quarters
- From upload to a finished report
- 18 minFrom upload to a finished report
- Ahead of the public report on TSMC's capacity cut
- 7 daysAhead of the public report on TSMC's capacity cut
- Processing steps completed without intervention
- 100%Processing steps completed without intervention
An analyst can read one call closely or skim ten. Neither of those is coverage. Every call here was analysed to the same depth, against the same criteria, whether or not anyone had time for it that week.
- Red flags ranked Critical, Major or Minor, with the moments behind each one
- Forward guidance items, with direction, impact and the executive who gave them
- Every Q&A exchange, classified and attributed to its responder
- Tailwinds and headwinds, separated and summarised
- Business performance and balance sheet, read from transcript and press release together
- Evasiveness scored line by line, alongside 32 voice dimensions
Across the portfolio the desk ranks what gets flagged and how often, which management teams answer directly and which do not, and how a company's delivery moves from one quarter to the next.
18 minutes, median
The average call runs 58 minutes and nobody has to sit through it. The report is ready in about 19 minutes unattended, median 18 and longest 27. It lands while the call is still the news, not after the market has finished repricing it.
22+ companies, one desk
Earnings season arrives for the whole portfolio in the same fortnight. In practice a small team reads the three or four calls that look most important and skims the rest. Every one of them was analysed to the same depth instead.
212 moments behind 73 flags
No flag arrives as an opinion. Each of the 73 carries the moments that produced it, the speaker, the timestamp and a read on how the answer was delivered, so an analyst can go straight to the passage and disagree with it.
5,109 lines, one standard
Every line of every call scored on the same scale. An analyst on the fourth call of the day does not hold the fifth to the standard they held the first, and across a portfolio that drift is invisible until it costs something.
- Red flags: 3 Critical, 66 Major, 4 Minor
- 73Red flags: 3 Critical, 66 Major, 4 Minor
- Transcript lines, speaker-separated and scored
- 5,109Transcript lines, speaker-separated and scored
- Tailwinds and headwinds identified
- 437Tailwinds and headwinds identified
- Important statements extracted and scored
- 392Important statements extracted and scored
- Q&A exchanges captured, across 159 themes
- 380Q&A exchanges captured, across 159 themes
- Forward guidance items, with direction and impact
- 268Forward guidance items, with direction and impact
One answer in five is not a straight answer
Of 3,297 classified answers, 79.6% were direct, 19.7% partial and 0.7% fully evasive. That last group is small on purpose, because a classifier that finds evasion everywhere is measuring its own bias. These are the twenty-two moments worth replaying.
Severity is a ranking, not a rubber stamp
Three calls produced a Critical flag; most produced two or three Major ones and nothing above. The ranking exists so an analyst covering twenty-two companies knows which two calls to open first, not so every call comes back looking urgent.
A signal, seven days early
Volume is one argument; this is the other. The desk flagged this on the call itself, and independent reporting confirmed it a week later. Both dates are on the public record.
“The CEO shows cognitive strain and evasiveness regarding mature-node capacity reductions and the conversion of 12-inch capacity to advanced packaging.”
15 January 2026
Basetone.AI · TSMC Q4 2025 earnings call analysis
“TSMC to Cut Fab14 Mature-Node 12-inch Capacity by 15%–20% to Reflect Changing Demand Mix”
22 January 2026
Counterpoint ResearchShares declined between the 16th and the 21st, before the capacity cut was reported. What the desk added was a better reading of the human signals already in the call.
“We grew our portfolio monitoring from 50 to 80 companies.”
EducationCircuit Stream
Immersive technology training · 6 campuses, US and Canada
Instructor evaluation, on every class instead of a sample
Every class, from every instructor, scored on Circuit Stream's own rubric, instead of one sampled session per instructor.
See what we builtGlobal investment bank
The questions a founder steers around
A global investment bank, management meetings from a live deal process, and everything the room noticed but nobody wrote down.
See what we builtDefence, intelligence and law enforcement
Interviews reviewed line by line, including the ones nobody had time for
Contradictions across sessions, questions answered around, delivery against the speaker's own baseline, accent and dialect markers, and what the video shows at the same moment.
See what we builtTell us the decision you want to fix
One call, no deck. We will tell you whether this is something we can build for you, and if it is not, we will say that too.