AI IP-INTELLIGENCE

The TLO and patent examiner's best friend.

PHOSITA searches prior art and scholarly literature, maps the technology landscape, scans Thai patent formalities, and drafts claim amendments. It does the legwork. You keep the discretion, and you sign.

Published data only · Real patent number + verified legal status · Named human sign-off

Real patent number + verified legal status, or it reads unverified
A named human signs off. The model never self-approves
Published data only. No confidential disclosure ingested
PDPA-compliant. Security built to ISO/IEC 27001 controls

One workflow, four pillars

Search to landscape. Patentability to report. Formality to amendment.

Each pillar feeds the next. Prior art found in a patentability run is cached and reused when you reach review, so PHOSITA remembers what it found.

Illustrative technology-landscape dashboard

IP & Scholar search to a Technology Landscape

Unified patent and scholarly search rendered as an interactive landscape. Every hit shows a real number and verified legal status, or it reads unverified.

Illustrative patentability Score Board

Patentability search to a report

Semantic prior-art search to a structured report: Key Features, a Score Board, and Relevant / Related / Additional, each hit cited to its source and saved to your cache.

Illustrative Thai formality scan

Thai Patent Formality Scan

Scans an application against Thai DIP practice and the Patent Act, flagging formality defects with inline, cited comments and the 90-day clock.

Illustrative substantive amendment redline

Substantive amendment

Proposes claim amendments to overcome cited art, as human-authored redlines with spec-support checks. Reviewed and signed by you, never auto-applied.

The whole landscape, not a black box

See the inputs behind every verdict.

  • Every chart cross-filters and drills to the real patents behind it.
  • Legal status is a first-class object, shown with an "as of" date.
  • The search you ran is logged and replayable, so it is defensible.
  • Uncertainty is shown as a signal, never a bare verdict.
Request a walkthrough
PHOSITA technology-landscape dashboard on a laptop
Illustrative interface

The principle we build on

Trust is anchored to traceable inputs and a named human, never to the model.

PHOSITA never shows a bare AI verdict. It shows the patents, papers, and rules it consumed, each traceable to a real source, with a human expert as the visible author and sign-off. No affordance implies an autonomous clearance or legal decision.

Encrypted, isolated data store

Your searches stay yours

Your own database. Your own cache. Never pooled.

PHOSITA stores your prior art in a per-user cache that is isolated per institution, never pooled across tenants, and never used to train any model. Deletable on request.

  • PDPA-compliant, with SCCs for any offshore processing.
  • Security built to ISO/IEC 27001 controls; certification in progress.
  • Aligned toward ISO/IEC 42001 for responsible-AI governance.

Built for TLOs and patent examiners.

A copilot that does the legwork and leaves you the decision. Bring PHOSITA to your office.