Skip to content
Join the beta
Kairos beta is open

See every path.Choose yours.

Sightegy builds Kairos, decision intelligence for biotech. It reads every kind of evidence, from genetics and omics to trials, filings and deals, lays out the paths ahead and shows what each one rests on. You choose.

Join the beta Partner or invest
  • Built in Zürich
  • ~700K source-traced facts
  • 43,555 trials
  • {{ dueLine }}

Every cue, from genome to term sheet, read into a dated fact that keeps its source.

Biotech runs on decisions worth years and millions. Which indication first. When to raise. Who to partner with. Most are still made on slide decks and instinct. Kairos lays out every path and what it rests on.

From genome to term sheet.

Kairos reads the biology as closely as the balance sheet: human genetics, expression down to single cells, pathways, potency and safety signals, next to trials, regulators, filings, deals and markets. Everything lands in one dated record, and every path it lays out shows which threads of evidence it rests on.

Illustrative, for the fictional venture AST-201: seven threads of evidence (biology, clinical, regulatory, corporate and financial, deals and IP, market, ventures) are woven into one dated record, which splits into three paths. Psoriasis first rests on 6 threads, 1 of them contested. Psoriatic arthritis first rests on 5 threads, 1 contested. Partner after Phase 2a rests on 6 threads, 2 contested. You choose.

The principle

Language models read. Kairos reasons.

01Read

Every fact keeps its source.

Language models do what they are good at: reading. Agents extract dated facts from registries, regulators, filings and deals, and every fact keeps a fingerprint of its source file. About 700,000 so far.

Reading a filingIllustrative
Form 10-Q · filed 12 Aug 2026p. 14

Clinical development

AST-201 is an oral, allosteric inhibitor of TYK2. In July 2026 we completed the single- and multiple-ascending dose cohorts of our Phase 1 trial. No dose-limiting toxicities were observed at any dose level. Exposure rose in proportion to dose and supports once-daily dosing. We plan to start a Phase 2 trial in plaque psoriasis, subject to financing.

fact
Phase 1 SAD/MAD complete, no dose-limiting toxicity
source
10-Q, p. 14
sha256
9f2c…e41a
observed
12 Aug 2026

02Reason Next

Numbers are not a language task.

A probability should start from a base rate that counts the failures and move only on dated, cited evidence, one visible step at a time. This is the reasoning engine we are building into every number Kairos shows.

AST-201 · Phase 2 success, psoriasisNextIllustrative

Each step is one dated, cited signal. No source, no step.

03Record

Then it keeps score.

Every forecast is logged before the outcome, with a due date, and scored when it falls due.

{{ dueChip }}

Forecast recordIllustrative
ForecastpLoggedDue
Phase 3 readout meets its primary endpointCatalyst · psoriasis0.6202 Sep 2026Q4 2026
Approval decision arrives on its due dateCatalyst · regulatory0.7418 Sep 2026Q4 2026
AST-201 reaches Phase 3From the reasoning above0.3906 Oct 202630 Jun 2028
Logged before the outcome. Scored when it falls due.

Kairos

One platform. Three studios.

Prepare a venture, brief a decision, benchmark against the ecosystem: three ways into the same source-traced record.

KAIROS
Illustrative
Next: Track Record Next: Discovery Studio Explore Kairos

Inside the engine. Every layer shows its work.

An evidence store, a semantic memory, a dated graph and two dozen specialist agents, built so that language models read and models reason.

  • Evidence store
  • Semantic memory
  • Dated graph
  • Specialist agents
Inside the engine

Most AI tools quote accuracy. Kairos is built to be checked.

Accuracy on past data is easy to quote and hard to check. Kairos is built the other way round: every citation opens its source passage, and every forecast is logged before the outcome.

RowThe usual AI toolKairos
01Quotes accuracy on history it has already seenLogs forecasts before the outcome, with a due date
02Counts the winnersCounts the failures: base rates include ventures that dissolved
03One score, no trailEvery citation opens its source passage
04Fills gaps with fluent guessesSays “Not answered” rather than invent
05The model states the probabilityA model computes it, signal by signal Next

What Kairos has read so far.

September 2026

about 700K facts extracted from company filings and reports, each traced to its source file
43,555clinical trials
27,807R&D programmes, every field cited
1,128Swiss & Nordic ventures, failures included
15+public sources, from registries to deal records
20–40minfor a typical brief

Pre-registered forecasts

{{ dueBig }}{{ dueUnit }}

{{ dueText }}

Built for the people who make the call.

Founders, investors and partners read the same evidence from different chairs. Kairos shows each of them what a decision rests on.

Biotech founders & boards

Pressure-test your venture from the investor’s chair before you send it.

Founder Studio reads your venture from four chairs and lists, at most three at a time, the blockers a reader will notice first.

Join the beta

Investors

Diligence where every claim opens its source.

Briefs that show what each conclusion rests on, flag where the specialist agents disagree, and say “Not answered” rather than invent.

Investor conversations

Pharma BD & search-and-evaluation

Landscapes and catalysts with the evidence attached.

Competitive landscapes and dated catalysts across 43,555 trials and 27,807 programmes, each field traced to its source.

Partner with us

The road ahead

From the record to the reasoning layer.

Solid: built and runningDashed: building next

  1. Jul 2025

    Sightegy founded in Zürich.

    A company with one aim: let biotech teams see every path before they choose.

  2. 2026

    Kairos built.

    ~700K source-traced facts, three studios and 150+ agent tools, built end to end.

  3. 2026

    Forecasts on the record.

    Logged before the outcome, each with its due date, and scored when it falls due.

    {{ dueShort }}

  4. Horizon

    See what was knowable on any date.

    The reasoning layer for drug-development decisions.

Portrait of Bogdan Rivoal, founder of Sightegy Designed and built Kairos end to end

The founder

“Kairos began as my own frustration. Building ventures, I had to make positioning and partnering calls without a business-development team, from evidence scattered across trial registries, filings and deal announcements. And I saw colleagues with excellent science face the same calls, with no access to the tools large pharma takes for granted. Sightegy exists so that early-stage teams don’t have to.”

Bogdan Rivoal

Founder · Strategic Advisor & AI Platform Architect

Molecular biologist with 15+ years in RNA biology, oncology and liquid biopsy. He advises biotech leaders and investment committees on indication, diligence, partnering and fundraising.

  • ETH Zürich RNA Biology group
  • NIH U19 Programme Director
  • Institut Pasteur / Curie
  • $3M+ R&D, go/no-go authority
  • Executive MBA HSG / ETH
  • Built Kairos end to end
Meet the founder LinkedIn

Build the reasoning layer with us.

The beta is open. Bring a real programme and a real question, a partnership idea, or an investor’s eye.

Beta test Kairos

Run Founder Studio and Strategy Studio on your own programmes, with a direct line to the founder and a say in what we build next.

Join the beta

Partner with Sightegy

Pharma BD, data and cloud partners: bring Kairos into search and evaluation, connect your own AI agents, and shape the Evidence API with us.

Partner with us

Invest in the vision

A Zürich company building the reasoning layer for drug-development decisions.

Investor conversations
Next · The platformKairosOne platform, three studios, built so that language models read and models reason.