Request a proof-of-conceptRequest a POC

The intelligence bridge between America and Asia.

We help US security and intelligence teams model their Asia footprint with AI, watch the adversaries targeting it upstream, and verify whether they’re exposed.

Read by your feedsMissedRead in local context, checked against your environment

Illustrative, not to scale.

Request a 4-week proof-of-concept

One live use case, run end to end.

Work email addresses only.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

Asia remains a key supply chain hub for the world.

Engineers in India, vendors in China, contact centers in the Philippines. Adversaries reach you through them, in places and cultures your tools don’t read.

Local service offices, engineering functions, manufacturing assets and suppliers. What’s your exposure in Asia?

We’re seeing a growing number of regional incidents that materially impact global businesses.

  • CoinbaseBribed outsourced agentsUS$307M in breach costs
  • QantasSocially engineered contact center5.7M customers’ data exposed
  • StellantisBreached vendor platformCustomer contact data exposed

Public incidents. None is a Protos Labs customer.

See what West-centric feeds miss.

The threat signals targeting your staff, vendors and software in Asia, interpreted with local cultural context and drawn from restricted-access datasets.

  1. Nation-state intelligence tradecraft

    We’re cyber operators with national-security backgrounds, proven on national missions with defense and homeland-security agencies since 2024.

  2. Asian cultural context

    Firsthand understanding of the region’s cultural nuance, so we see what US-based teams can’t.

  3. Multi-source data fusion

    We ingest and analyze datasets across cyber and influence operations to identify early adversarial signs. For instance, LinkedIn recruitment of remote workers can be an early sign of an attack. We also have first-party access to specific Asian datasets.

  4. Around the clock

    Our Asia-based operators work while the US sleeps, so findings are ready when your working day starts.

The people you’ll work with.

Joel Lee

Joel Lee

Co-founder & CEO, Singapore

Ex-Booz Allen Hamilton (Cyber Intelligence).

Simeon Tan

Simeon Tan

Co-founder & CTO, Singapore

Ex-Booz Allen Hamilton (OT Security).

Sam Goh

Sam Goh

Field CISO & GM, New York

Previously CISO at Siam Commercial Bank DataX (Data & AI CoE).

sam_goh@protoslabs.io
Sam Goh, Field CISO at Protos Labs, at the Billington CyberSecurity Summit
Sam Goh, Field CISO at Protos Labs, at the Billington CyberSecurity SummitBillington CyberSecurity Summit · Sep 8–10, 2026 · Washington, DC

All events →

Model. Collect. Analyze. Hunt. Track.

Protos AI models your environment, goes upstream to collect early-warning signals, and carries every finding into a hunt to verify whether you’re exposed. You stay in full control, setting the requirements and approving every conclusion before release.

  1. Understand your environment first.

    We model your assets, vendors, software, AI systems and workforce, so collection targets what actually puts you at risk, not generic data from a shared feed.

  2. Intelligence that executes.

    We carry every finding into a hunt in your environment, ending in one answer: Are you exposed or not?

Draws on

  • OSINT and web searchSearch, news, RSS, web pages
  • Commercial intelligence feedsPremium threat data providers
  • Enrichment and malware analysisSandboxing, reputation, metadata
  • Internal enterprise toolsEDR, SIEM, asset inventory, case management
  • Software packagesPackages and dependencies in your stack
  • Social mediaPlatforms, forums, messaging
  • Financial documentsCase files, entity records, adverse media
  • CryptoOn-chain data, wallets, exchanges

Agent stages

Track feeds Collect, so every investigation sharpens the next.

Produces

  • Verified findingsExposed or not, with the evidence
  • Actionable intel reportTactical, operational and strategic
  • Detection rulesFor example, for your SIEM
  • Intelligence-led vulnerability managementRanked by exploitability and asset criticality
  • Hunt packagesReady-to-run queries, IOCs and detection logic
  • Adversary watchlistProfiles kept current as modus operandi changes
  • Financial crime and fraud investigationTrace suspicious flows and patterns
  • Entity and network risk analysisOwnership, affiliations and transaction links
Select a stage, or point at an input or output, to trace how they connect.
04

Hunt

Agents actively hunt to verify whether you are exposed.

  • Execute hunt hypotheses against your own telemetry
  • Identify exposed assets
  • Verify exposure; never block or remediate
  1. 01

    Model

    Agents build a model of your environment, so the intelligence you receive is tailored to you.

    • Assets, vendors, software packages, AI systems and workforce
    • Collection directed by your priority intelligence requirements, not a shared feed

    Draws on

    • Internal enterprise tools: EDR, SIEM, asset inventory, case management
    • Software packages

    Produces

    • Intelligence-led vulnerability management
  2. 02

    Collect

    Agents collect threat signals upstream, at the source.

    • Advisories, researcher write-ups, code repositories, forums, social media, vendor disclosures
    • Logs, PDFs, emails, screenshots, chats, images and video in one pass
    • Bespoke collection where a source is hard to reach

    Draws on

    • OSINT and web search
    • Commercial intelligence feeds
    • Social media
    • Financial documents
    • Crypto

    Produces

    Signals for Analyze

  3. 03

    Analyze

    Agents analyze what is targeting you and how.

    • Read with cultural nuance: Tagalog groups, Chinese-language channels, regional forums
    • Model how adversaries would target your stack
    • Hand Hunt hypotheses it can execute

    Draws on

    • Enrichment and malware analysis
    • Financial documents

    Produces

    • Actionable intel report
    • Detection rules
    • Intelligence-led vulnerability management
    • Financial crime and fraud investigation
    • Entity and network risk analysis
  4. 04

    Hunt

    Agents actively hunt to verify whether you are exposed.

    • Execute hunt hypotheses against your own telemetry
    • Identify exposed assets
    • Verify exposure; never block or remediate

    Draws on

    • Internal enterprise tools

    Produces

    • Verified findings
    • Hunt packages
    • Detection rules
  5. 05

    Track

    Every adversary found goes on a watchlist, monitored as their methods change.

    • Build on previous investigations to watchlist new adversaries and their techniques
    • Every investigation sharpens the next

    Draws on

    • Your previous investigations

    Produces

    • Adversary watchlist
    • Entity and network risk analysis

Video · 2:52

Start at the signal, not the ticket.

A developer can patch a flaw quietly, with no announcement and no CVE. In this video, Protos AI reads one month of commits on a real dependency, 200 diffs, and four stand out as likely silent fixes for an analyst to confirm.

Captions are on screen.

Results from live deployments.

  1. Coordinated inauthentic behavior

    17,000 comments narrowed to 11 malicious profiles within 8 hours of first prompt.

    Homeland-security agency. 6,000 profiles in Singlish, Mandarin and Chinese dialects, analyzed by one analyst with Protos AI, against an estimated 48 analyst-hours (3 analysts × 16 hours).

  2. Retrospective threat hunt

    One real incident found among ~1,000,000 log rows, confirmed by a sectoral CISO.

    Command-and-control traffic invisible to human analysts, detected by correlating logs across dozens of research and education institutions.

  3. Threat-intelligence analysis

    ~4 hours of analyst work per threat cut to ~15 minutes.

    Critical-infrastructure CTI team.

CISOResearch & Education agency
“In one deployment, Protos AI surfaced a single anomalous event buried in over a million rows of our security logs — and our team confirmed it matched a real incident. That is not something we could have found manually.”
DirectorHomeland security agency
“On national-security missions, Protos AI has changed what our analysts can get through. Investigations that used to take days now take hours, and the conclusions hold up to the scrutiny our work demands.”

In the press.

Coverage of the Protos AI launch at RSA Conference 2026 in San Francisco, and since.

The Business Wire release was also carried by Yahoo Finance.

Built by intelligence operators, proven on national security missions.

Anyone can build an AI agent. Few can build it on nation-state intelligence tradecraft.

Trust you can measure

  • Every conclusion traces to its sources and reasoningFull audit logs
  • Trust Fabric scores AI output on accuracy, reliability, consistency and speedAt POC and after go-live

Recognition

  • CyberCall, Singapore’s national cyber innovation callTwo-time winner
  • NVIDIA InceptionMember
  • Microsoft AI AccelerateMember

Catch the adversary upstream.

Start with a 4-week proof-of-concept on one live use case. We build every POC on your own workflows, playbooks and data sources, with success criteria agreed with you at kickoff.

Sam Goh, Field CISO & GM · sam_goh@protoslabs.io

Work email addresses only.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.