BD search and evaluation where the data room redraws the landscape before the committee does
A first pass on a seller’s data room in hours, with cited memos for every function.
See one case, screen by screen ↓Software for the biopharma Search & Evaluation function: scouting external assets, screening them against your strategy and running diligence on the few that open a data room. BD Search & Evaluation puts every asset on one landscape of fit against differentiation and lets seven agents read each data room into cited function memos, with the red flags drawn on the asset itself.
How to choose AI solutions for life sciences →
Updated
Search & Evaluation screens thousands of external assets and takes a handful into diligence. When a seller's banker opens a data room, six functions — clinical, safety, CMC, IP, regulatory and commercial — read hundreds of files by hand against a bid deadline, and a first pass takes days of expert time before anyone can say proceed or pause.
The problems hide between documents: an investor deck that counts unconfirmed responses, a brochure on an older safety cut, a process report that moves drug substance to a new route after the toxicology lots were made, with comparability only planned. And the landscape slide never meets the diligence memos, so the asset that looks best on public data keeps looking best until the CMC lead speaks at committee.
Estimated split of one first pass, in hours, by hand and with the solution — about 60 hours by hand, as the solution's own dashboard shows.
Seven specialist agents scout, place each asset on the landscape, read the data room and raise red flags; function leads sign and the committee decides.
Northwind Bio watches 18,240 external assets and pursues those in its thesis zone: fit 82 or more, differentiation 65 or more. At 07:40 Hezhen Pharma's banker opens a 214-file data room on HZ-2301, an oral IL-23 receptor antagonist peptide for plaque psoriasis. Here is what happens next, screen by screen, in the working solution.
Priya's landscape plots 13 assets by strategic fit against differentiation, sized by deal value, coloured by therapy area. Tessivo Pharma's TSV-208, a bispecific antibody–drug conjugate in Phase 1b, sits top right at fit 92 and differentiation 84 — but its data room raised 8 red flags and 5 are still open, so the dot is dimmed and ringed in safety, CMC and IP: an ILD rate up from 5.2% to 6.9%, lot TSV-L005 with DAR 4.6 outside 3.4–4.2, and a change-of-control clause in the linker licence.
“In the thesis zone · 5 red flags dim it.”
Hezhen Pharma's room arrives with 214 files, 8 people wall-crossed. On public sources HZ-2301 sits outside the thesis zone at fit 87 and differentiation 52, a hollow dot: Phase 1 complete in healthy volunteers, the psoriasis cohort reported only in the data room. Priya selects it and clicks Read the data room.
“Public sources only — red flags appear once its data room is read.”
The Data Room Reader classifies every file and extracts 1,370 facts, keeping the page reference for each. Nine superseded drafts — among them a cohort summary draft v0.6 — are set aside so their values never reach a memo. The Clinical & Safety Analyst reads the psoriasis cohort: PASI 75 in 18 of 28 patients (64.3%) at week 12 on 50 mg once daily, against 1 of 9 on placebo.
The Data Room Reader reads one deal's data room only.
The cohort summary says HZ-2301 was taken with or without food, with a fed/fasted AUC ratio of 0.94. From 19 public sources the analyst finds two oral IL-23 pathway agents in Phase 3 or filed — both dosed on an empty stomach. Differentiation goes from 52 to 81 and fit from 87 to 89, and the dot crosses into the thesis zone, leaving a dashed ghost at its public view.
The monkey NOAEL is 200 mg/kg/day, about 30 times the clinical exposure — but only for pilot-plant lots HZ-DS-1901 and 1902. From lot HZ-DS-2104 the drug substance comes from a hybrid route at a contract manufacturer in Suzhou, with no comparability data (RF-41), and those lots carry a D-arginine epimer at up to 0.31% that the toxicology lots never had (RF-42). The US composition-of-matter patent runs to 2043 and rights outside Greater China are clear.
“In the thesis zone · 2 red flags dim it.”
Six function tiles show each memo's score and flags; CMC scores 2.5 with 2 flags. The memo, drafted by the CMC & IP Reviewer in 17 minutes, cites every sentence, and the process development report opens at section 6, page 44, with the route change and the planned Q1 2027 comparability assessment highlighted. Two questions for the seller are drafted for Lena to approve.
“Has the D-arginine epimer been qualified in any toxicology study?”
RF-41 is a covenant: a comparability study as a condition to signing. RF-42 is price-adjusting: a bridging toxicology study of about six months, priced into the offer. Both are open and owned by Lena Ortiz, with their sources one click away, and the register shows no deal-breaker.
The data room view lists all 214 files by folder — 01 Corporate to 08 Contracts — with function, version and facts extracted, and marks the files cited in a memo. The agent-built competitive landscape is labelled as such, kept apart from the seller's documents.
Scored with the agreed weights, HZ-2301 comes to 3.5 of 5 — rounded up from just under the proceed line — so the draft says proceed with conditions. After diligence PTRS falls from 16% to 13%, rNPV from $0.74B to $0.55B and first launch from 2031 to H1 2032; bid guidance is a licence outside Greater China at $180–220M upfront. The two conditions are the comparability study and a bridging toxicology study that qualifies the epimer.
“6 function memos not signed yet … the committee sees which memos are open.”
The recommendation reaches the committee only after Priya signs it and Omar Haddad countersigns as deal lead; the committee decides. Dr. Vale chairs the Oct 15 meeting, where Quillara's QLR-310 is on the agenda at 4.0, proceed. Tessivo's TSV-208 follows at a special meeting on Oct 22: 3.3, proceed with conditions, with four conditions such as licensor consent to change of control and a root-cause report for lot TSV-L005, at $620–680M upfront, down from $720M.
The same data room, seen by the four people who decide what it means — what their week looked like, and what it looks like now.
Seven specialist agents scout, place each asset on the landscape, read the data room and raise red flags; function leads sign and the committee decides.
Sweeps registries, patents, congress abstracts and company news in nine languages against your criteria; merges code names across languages and ranks assets by fit, with citations.
Classifies every data-room file by function, document type, version and data cut-off, and extracts its facts.
Drafts the clinical efficacy and safety memos: confirmed responses, adverse events of special interest, class risks, data cut-offs, NOAELs and margins.
Reads lots, specifications, suppliers, patents and licences; finds out-of-spec lots, single sources, change-of-control clauses and royalty stacks.
Reads agency minutes and correspondence for what the agency asked and what it costs in time; builds the competitive landscape from public sources in nine languages and scores each asset's differentiation.
Compares the same fact across the investor deck, study reports and brochure, and explains why values differ — an older data cut, unconfirmed responses.
Scores the deal with the agreed weights, carries red flags into conditions and valuation inputs, and builds the committee pack.
Memos signed, then decides. The agents propose; a named person decides.
Ask about any deal, memo or red flag — answers cite the data-room page. Or say what to change.
What could stop the Tessivo deal?
One deal-breaker and four price items on TSV-208. The deal-breaker, RF-05: a change of control needs the linker licensor's consent, or the licence can be terminated — make it a condition to signing, or walk away. Price: lung disease up to 6.9%; dose optimisation adds about 9 months; the deck overstates response rate; three competitors in Phase 3.
Does the investor deck match the study report for TSV-208?
Not quite — the Contradiction Checker found two differences. The June deck says 45% response (18/40, confirmed and unconfirmed); the interim study report confirms 38.1% (16/42). The deck says 3 cases of lung disease, 5.2%, all low grade; the report has 4 cases, 6.9%, one Grade 3 — the deck uses the older February cut. The valuation uses the study report.
Why does Hezhen look dimmer now?
HZ-2301 moved into the thesis zone when its data room was read — PASI 75 in 64% at week 12 and dosing with food raised differentiation from 52 to 81. Two CMC red flags dim it: the route changed after toxicology and nothing bridges the old and new lots (RF-41), and the new lots carry a D-arginine epimer at 0.31% that the animals never saw (RF-42). Lena Ortiz decides whether these become conditions or price.
Who can see this deal?
Each deal sits behind its own information barrier. Tessivo: 9 wall-crossed people (S&E, BD, six function leads, legal); Quillara: 11; Brightmoor: 7. The agents serving a deal read only that deal's data room — no cross-deal search or memory. Every page view is watermarked and logged; exports need the deal lead's approval.
13 screens from the working solution, on its sample data. Pick one to see it large.
Criteria, scoring, reviewers and the information barrier are settings, and each change is recorded.
| Setting | Default | Choose from |
|---|---|---|
| Thesis zone | Fit at least 82 · differentiation at least 65 | Step each threshold up or down |
| Therapy areas | Oncology · Immunology · Neuroscience · Cardio-metabolic | Assets outside these areas are not shortlisted |
| Modalities | Antibody–drug conjugates · bispecifics and multispecifics · small molecules · peptides · in vivo cell therapy · RNA medicines | Click a modality to remove it |
| Royalty stack limit | 10% | Step up or down; assets above the limit are flagged |
| Sweep schedule | Weekly · Monday 06:00 | Weekly · Monday 06:00 · Daily · 06:00 |
| Recommendation weights | Clinical 30% · safety 20% · CMC 15% · IP 15% · regulatory 10% · commercial 10% | Step each share by 5 points; the total must be 100% |
| Score bands | 3.5 and above: proceed · 3.0–3.4: proceed with conditions · 2.5–2.9: pause · below 2.5: pass | Fixed bands; the screen shows where the deal in diligence now falls as weights change |
| Function reviewers | Dr. Maya Chen · Dr. Rafael Moreno · Lena Ortiz · Sam Patel · Dana Okafor · Omar Haddad | Each function memo needs its reviewer's signature |
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 deal data · “target” = the design goal, measured in the live solution · “estimated” = our estimate · “typical” = published figures (AI-agent study at a biotech investment fund — first competitive landscape 2.5 days by hand to about 3 hours; arXiv 2602.15019, “Hunt Globally: Deep Research AI Agents for Drug Asset Scouting” — F1 79.7% against 46–59% for general research tools). People, companies and assets named on this page are fictional — characters and sample data in the working solution.
Software for the biopharma Search & Evaluation function: scouting external assets, screening them against your strategy and running diligence on the few that open a data room. BD Search & Evaluation puts every asset on one landscape of fit against differentiation and lets seven agents read each data room into cited function memos, with the red flags drawn on the asset itself.
When a seller opens a data room, the Data Room Reader classifies every file by function, version and data cut-off, and the function agents draft cited memos for clinical, safety, CMC, IP, regulatory and commercial. In the working solution's dashboard, the median first pass from data room open to drafted memos is 6.6 hours, against about 60 hours by hand. Function leads then review the cited pages and sign.
The Asset Scout sweeps registries, patents, congress abstracts and company news in nine languages, merges Chinese, Korean and Japanese code names into one asset, checks rights and royalty signals, and ranks the shortlist by fit with every claim cited. Assets above your royalty limit are flagged, not dropped.
Each diligenced asset carries a red-flag counter and a ring for clinical, safety, CMC and IP. Every open flag fades the dot, and an open deal-breaker outlines it in red, so an asset that looks best on public data visibly dims when its data room shows a problem. A switch shows the landscape on fit only.
Yes. Each deal's files are kept apart, and the agents serving a deal read only that deal's data room, with no search or memory across deals. Only wall-crossed people open a deal, pages are watermarked with the viewer's name, every view is logged and exports need the deal lead's approval.
Yes. The agents draft memos, flags, seller questions and the recommendation; they never sign, send or decide. Function leads sign their memos, questions go to the seller only after a function lead approves them, the head of Search & Evaluation and the deal lead sign the recommendation, and the committee decides.
Yes. Therapy areas, modalities, the thesis-zone thresholds, the royalty stack limit, the sweep schedule, recommendation weights by function and function reviewers are all settings. Rules such as a deal-breaker blocking the signature can be switched or added.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your search criteria, scoring weights and a past data room — and it goes live once every quality gate has passed. We will walk you through it on one of your own data rooms first.
Sourced answers from papers, patents and reports in seconds, our own disagreements first.
1-min demoCompetitive Pipeline TrackingEvery rival trial, readout and approval in a sourced brief each Monday
1-min demoPreclinical Study Report AutomationA draft GLP report in hours, every number traced to raw data.
1-min demoWe'll run BD Search & Evaluation on your criteria and a past data room of your own.