Global Tech UpliftLinking intelligence

Workflow P1 · From the Meta CTIS and Google Health studies

Respiratory and COVID-19 situational awareness

What leaders see: Daily, privacy-safe signals of where illness and behaviour are shifting, checked against official data.

Who runs it
Epidemiology, health officer, data and privacy teams, partners
Cadence
Daily signals; weekly briefing
Plugs into
Survey platformsState and national case dataVaccination registriesDashboards and APIs
Touchpoints timeline

The workflow you run today, with 4 AI touchpoints

  1. 1Survey and study dataOpt-in responses arrive daily
    Privacy

    Federated analysis: answers computed on phones; only private totals leave

  2. 2Weighting and checksCorrect sampling bias
    Agent

    Automated weighting and anomaly checks on every daily batch

  3. 3BenchmarkAgainst cases, tests and vaccination
  4. 4Hotspot watchWhere signals rise first
    Predict

    Forecast positive tests and rising areas from survey signals

  5. 5Partner briefingHealth officials and agencies
    LLM

    Draft the weekly brief with charts and caveats; epidemiologists approve

  6. 6Policy and outreachTesting, messaging, resources
  7. 7EvaluateDid it work, for whom

Highlighted steps are where AI plugs into the existing workflow. People keep every decision.

GTU AI workflow

Privacy-safe signals, validated daily

Inputs
  • Daily opt-in survey responses
  • Official cases, tests and vaccination
  • Opt-in mobile study (federated)
  • Mobility and behaviour indicators
Intelligence
  • Bayesian weighting for sampling bias
  • Benchmarks against official data
  • Forecasts and hotspot detection
  • Differential privacy and suppression thresholds
Decision
  • Health officials set measures and messaging
  • Partners agree which signals to trust
Action
  • Target testing and outreach early
  • Tailor messages to groups that lag
Learning
  • Indicators re-validated against outcomes
  • Question wording refined with partners
New outcomes

What changes

KPIs

How we measure it

KPITodayWith GTUType
Lead time over official reportingWait for case reportsDays earlierPlanning target
Indicators validated against outcomesUnvalidated survey numbersEvery published indicatorNew measure
Individual-level data shared centrallyOften requiredNone: federated, with differential privacyNew measure
Decisions citing the weekly briefAd hocTracked each weekPlanning target

Targets are set with your team and tested in a pilot before anything scales.

Proof from our work: The Meta/UMD COVID-19 Trends and Impact Survey covered 114 countries and was used by WHO, CDC and national agencies (PNAS); the Google Health federated study measured masking and distancing without central data (npj Digital Medicine, 2024). Global COVID-19 impact monitoring · Privacy-preserving behaviour study

Next step

Run this workflow on your data.

Start with a free 30-minute discovery call. Most teams pilot one touchpoint first.

Book the call More public health workflows