Overview
The client context, challenges, audience, and objectives that shaped our approach and defined success.
Context
A Tufts University team (Food is Medicine Institute) contributed to the Google Health AI project, which includes nutrition-related content and functionality. We support the Tufts team as scientific advisor—leading project meetings, serving as primary content reviewer, and ensuring scientific validity, clinical relevance, and public health alignment across key phases of model development, testing, and rollout planning.
Objectives
Support the Tufts team in content development and oversight for the Google Health AI project: lead meetings, serve as primary content reviewer, escalate risks as needed, contribute to peer-reviewed dissemination through methodological drafting, and help shape the evidence backbone, evaluation approach, and governance needed for responsible scale.
Requirements
Specific needs, deliverables, constraints, and compliance considerations required to achieve objectives within timeline and budget.
- Lead and coordinate project meetings on behalf of the Tufts team; align stakeholders and escalate risks when needed.
- Provide high-level scientific input on validity, clinical relevance, and public health alignment of nutrition-related content and functionality.
- Support core build phases, including development of expert question/answer assets, curation of credible nutrition sources, and planning for testing, maintenance, and dissemination.
- Contribute to peer-review dissemination by drafting methodological sections for a manuscript.
- Maintain clearly non-clinical scope (no patient-facing or laboratory activities).
Our Process
A stepwise, evidence-based workflow from discovery to delivery, measurement, iteration, and scalable knowledge transfer.
- Alignment & risk escalation pathways — kickoff, roles, coordination, and risk handling.
- Framework design — scientific input into the overall approach and guardrails.
- Expert Q&A development — structured creation of high-priority questions and evidence-informed answers.
- Credible source curation — identification and evaluation of trusted nutrition sources to support grounded content.
- Scientific review & iterative validation — review cycles on content validity and scope; collaboration with Google AI/data teams to assess response quality and reliability.
- Testing + rollout planning — early-access testing approach, maintenance cadence, and dissemination plan for intended audiences.
- Dissemination support — methodological drafting to support peer-reviewed outputs.
Results
The measurable outcomes, qualitative feedback, and lasting assets that created value and informed next steps.
- Stronger scientific defensibility of nutrition-related content through expert review and structured iteration.
- Expert content backbone developed (high-priority question set + evidence-informed answers) aligned to real-world needs and appropriate scope.
- Credible source base curated to support grounded nutrition information, plus an agreed approach for future maintenance and updates.
- Improved output reliability through collaborative evaluation and iteration with Google AI/data teams.
- Clearer governanceand risk handling via escalation pathways and defined non-clinical scope.
- Manuscript methods drafting and coordination support for dissemination.
Other Case Studies
Quantifying Bread Intake Worldwide
Read More
Saudi NHANES: Scientific Committee Advisory
Read More
