Earnings call preparation that knows which question will hurt — and answers it only from the public record
Likely analyst questions answered from approved disclosures, cited.
See one case, screen by screen ↓Software that helps an investor relations team get ready for the analyst Q&A on a results call: predicting the questions, drafting the answers, tying out the numbers, clearing disclosure and rehearsing the executives. Earnings Call Preparation adds six agents that read the notes, calls and peer news, rank each question by likelihood and stakes, and draft each answer only from approved disclosures, every sentence cited.
How to choose AI solutions for life sciences →
Updated
Four times a year a listed pharma company takes about 45 minutes of analyst questions on its results call. In the three weeks before it, the IR team reads every sell-side note and preview, re-reads last quarter's transcript, guesses the questions in a spreadsheet, copies answers forward from the old Word Q&A book and pastes numbers from the draft release. The controller ties the numbers out by hand and legal reads the book late.
The ranking is gut feel, and the question that hurts is usually the one a peer's news made urgent in the last days — a competitor's Phase 3 readout, a pricing policy remark. Numbers drift between release drafts: an answer written against draft v6 still says 26% after draft v7 says 24.1%. And an answer one step more precise than the public record — a month for a topline readout, a 2027 figure before guidance, a hint of a deal — is new material information. Said to analysts without public disclosure, it is a Regulation FD problem.
Estimated by the solution's own step model for an IR team, per quarter — by hand and with the agents.
Six specialist agents read notes, calls and peer news, rank each question by likelihood and stakes, draft and check it; the IR lead approves.
Northwind Bio reports third-quarter 2026 results on Thursday 29 October — release at 7:00 am ET, call at 8:30. It is Wednesday 7 October, 22 days out, six days into the quiet period that began on 1 October. At 07:58 a peer publishes Phase 3 data. Here is what happens next, screen by screen, in the working solution.
Lena owns the Q&A book. Her home screen places the 16 questions ranked above 20% by how likely they are to be asked against what is at stake: 41 predicted, 4 disclosure flags open. The Talmirex 2027 floor sits in “prepare hardest” at 88%, and the panel shows why — Street notes add 44%, past calls 41%, investor feedback 15% — with Rachel Kim's note and her question on the Q2 call as evidence.
Harbor & Pike, Rachel Kim: “We expect the Q&A to focus on where Talmirex bottoms in 2027.”
Calder Pharma, one of the three peers Lena watches, reports oral IL-23 Phase 3 results in plaque psoriasis. The banner already says what it touches: matched to Veltrimab and to 2027 pricing, it may move Q-16 and Q-05. Lena clicks “Re-rank with this news”. The Research Reader reads the release in 2.4 s; the Question Forecaster matches it to all 41 questions in 5.2 s; the Answer Drafter re-drafts in 3.3 s; the Disclosure Guard checks in 1.9 s.
Calder named pricing policy as its main 2027 uncertainty — so Q-05, 2027 gross-to-net, moves too.
Q-16 slides across the matrix into “prepare hardest”, leaving “Q-16 was here · 44%” behind. It is re-phrased the way Victor Salas of Meridian will ask it — does a once-daily pill change your view of Veltrimab above $20 billion in 2030? — and its stakes go from medium to high, because the 2030 target has been on the record since investor day. Q-05 rises from 71% to 76%. Peer news now adds 56% of Q-16's probability.
“Re-ranked at 08:15 with Calder Pharma's results … The new Q-16 answer is drafted and checked — it needs your approval.”
The answer for the CEO is drafted only from the investor day presentation and the prepared remarks: the 2030 outlook already assumes oral IL-23 pathway competitors from 2028; Veltrimab remains the leader in new-to-brand prescriptions in psoriasis; sales above $20 billion in 2030 are still expected. Each sentence links to its passage, and the checks read “Every sentence cited”, “Approved sources only”, “Numbers tie to release v7”. Lena approves it for Rehearsal 1.
Rehearsal card, do not say: “Do not comment on Calder's data or compare trials.”
The Questions view lists all 16 ranked questions with probability, stakes, likely askers, the executive who answers and the status — 4 with a disclosure flag, 3 drafted, 3 in review, 6 approved. Q-16 and Q-05 carry a “moved” tag. Filters by theme, stakes and speaker; the 25 long-tail questions under 20% sit in the book's appendix.
Anika Bose of Bluewater will press for the VLT-302 topline date. The draft for Dr. Maya Chen, CSO, says “We now expect topline results in March 2027.” The Disclosure Guard flags it as new material information: the public timing, in the September press release furnished on Form 8-K, is “the first half of 2027”. It proposes that wording, with its citation, and Lena chooses “Use the public wording” or “Keep it — ask legal”.
The Disclosure Guard never clears its own flag — a person decides.
The Rehearsal Coach asks Q-03 the way Rachel Kim of Harbor & Pike would, times Omar's answer against a 30-second target and throws the follow-up: “So is it fair to model a 2027 decline similar to 2026?” The card gives the opening line, three numbers — $1.12B, −41.3%, about 45% US erosion — the bridge to Veltrimab's growth, and the line not to cross.
Do not say: “Do not give a 2027 Talmirex number or a floor.”
Lena submits the book for sign-off. Tomás runs the tie-out across 41 answers and 188 numbers: 187 tie, and Q-02 still says Veltrimab was “up 26%” — a figure from release draft v6, before the third-quarter rebate true-up. Draft v7 says 24.1%. He uses the v7 number and signs. Adjusted EPS, operating margin and R&D are checked against the reconciliation tables.
Priya cannot sign while answers carry wording that is not public. Three are left: a 2027 Talmirex decline and floor, “late-stage talks for an obesity asset”, and “March 2027” for VLT-302. For each the Disclosure Guard shows the public wording — “We will give 2027 guidance in January, as we always do” among them. She uses the public wording for all three and signs: no new material information, safe-harbour and non-GAAP references in place.
The original sentences stay on the audit trail.
The disclosure committee approves the 12 answers still waiting for IR approval — each already cited and past the Disclosure Guard, finance and legal — and locks the Q&A book as one record: 41 questions, 41 approved, 0 disclosure flags open, version 1.0. IR, finance and legal each signed with the meaning “reviewed and approved”. The rehearsal cards go to the CEO, CFO and CSO, and any change from here needs the committee.
“Locked for the call · changes need the disclosure committee.”
The same results call, seen by the five people who carry it — what the run-up looked like, and what it looks like now.
Six specialist agents read notes, calls and peer news, rank each question by likelihood and stakes, draft and check it; the IR lead approves.
Reads each analyst note and preview from the research aggregator export, and every peer release, and lists the questions the analyst says they will ask, their concerns, estimates, rating and price target — with the page.
Splits our last eight call transcripts and peers' calls into who asked what and how it was answered, records which topics keep coming back, and marks answers that set a public position, such as a timing or an erosion curve.
Clusters raw questions into one well-phrased question each, scores probability from four signals — Street notes, past calls, peer news and investor feedback — and stakes from what is exposed, and names the likely askers.
Drafts each spoken answer for the named executive only from the approved disclosure library, the current release draft and prepared remarks, cites every sentence, and writes the rehearsal card.
Checks every sentence before people see it: new material information, numbers that differ from the latest release draft, outlook statements without the safe-harbour reference and adjusted measures without the reconciliation — and proposes the public wording with its citation.
Plays the analysts in rehearsal: asks each question in the analyst's voice, times the answer against its target and adds a realistic follow-up.
Reviews; finance and legal sign. The agents propose; a named person decides.
13 screens from the working solution, on its sample data. Pick one to see it large.
Who signs the book, the rules every answer must pass and how late-breaking news is handled are settings.
| Setting | Default | Choose from |
|---|---|---|
| Book owner | Lena Ortiz | Lena Ortiz · Sam Patel |
| Finance tie-out | Tomás Reyes | Tomás Reyes · Omar Haddad |
| Disclosure committee | Thu Oct 22 · chaired by Omar Haddad | Thu Oct 22 · Fri Oct 23 |
| Approved sources only · cite every sentence · flag new material information | On | Part of the disclosure policy — cannot be switched off |
| Tie every number to the latest release draft · flag forward-looking and non-GAAP wording | On | On · Off |
| Put a question in the book from this probability | 20% | 10% · 20% · 30% |
| When peer news or a new note arrives | Ask me before re-ranking | Ask me before re-ranking · Re-rank automatically |
| Watch these peers | Calder Pharma · Ostrava Therapeutics · Brightwell Biosciences | Their releases and calls feed the forecast |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
“demo” = seen in the working solution, on its sample quarter · “estimated” = the solution's own step model before deployment · Regulation FD (17 CFR 243), Regulation G (17 CFR 244) and the PSLRA safe harbour are named as context only. People, companies, analysts, products and studies named on this page are fictional — characters and sample data in the working solution.
Software that helps an investor relations team get ready for the analyst Q&A on a results call: predicting the questions, drafting the answers, tying out the numbers, clearing disclosure and rehearsing the executives. Earnings Call Preparation adds six agents that read the notes, calls and peer news, rank each question by likelihood and stakes, and draft each answer only from approved disclosures, every sentence cited.
The Question Forecaster clusters raw questions from sell-side notes, past call transcripts, peer calls and news, and investor feedback, then scores each one's probability from those four signals and its stakes from what is exposed — guidance, legal exposure, undisclosed timing or deals. Every probability shows what each signal added, and people can overrule it.
The news arrives as a banner matched to the questions it could move. One click re-ranks the matrix: in the working solution, a peer's oral IL-23 Phase 3 readout moved the oral-competition question from 44% to 82% and into “prepare hardest”, with its answer re-drafted and checked in 18.6 seconds. You can also set it to re-rank automatically.
Answers are drafted only from what is already public or becomes public on results day. The Disclosure Guard flags anything more precise or newer than the public record — a month for a readout, a figure for a year without guidance, talk of a transaction — and proposes the public wording with its source. Legal decides every flag.
Analyst notes are licensed research: they are read to predict questions and never quoted in an answer. Numbers come from the current release draft and are re-checked whenever a new draft arrives; the finance tie-out ties every number in the book to it.
Yes. The agents draft and check; the VP Investor Relations approves each answer, the corporate controller signs the tie-out, legal decides every disclosure flag and signs, and the disclosure committee locks the book. Every step is on the audit trail.
Yes. Who owns the book, who ties out and who chairs the disclosure committee, the switchable disclosure rules, the probability that puts a question in the book, how late-breaking news is handled and which peers are watched are all settings. The core rules — approved sources only, cite every sentence, flag new material information — cannot be switched off.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your disclosure policy, past Q&A books, transcripts and filings — and it goes live once every quality gate has passed. We will walk you through it on last quarter's call first.
Every answer drawn from one evidence library, cited and owner-approved.
1-min demoLegislative & Regulatory TrackingEvery relevant bill and rule read overnight and briefed with sources.
1-min demoPartnering Opportunity TriageEvery inbound pitch scored and answered within days, earlier decisions brought back.
1-min demoWe'll run Earnings Call Preparation on your last quarter's notes, transcripts and disclosures.