INDEPENDENT PORTFOLIO DEMONSTRATION · SYNTHETIC DATA17 NOVEMBER 2026

Portfolio decision workspace

Opportunity & Investment Pipeline

28 opportunities. One view of value, readiness and funding.

Status describes the current incremental decision—not deployment maturity. Approval is limited to the recorded stage. Click a status to filter the heat map.

Portfolio balance

See the concentration.

Full portfolio snapshot · charts remain unfiltered

Opportunity mix by business outcome

Each opportunity has one primary category. Click a segment to inspect its opportunities.

Investment by category

Approved = recorded incremental commitments. Requests = requested amounts for active opportunities, including already approved portions. Neither measure is actual spend.

Investment criteria

Compare the opportunities.

Click any column heading to sort; click again to reverse. Cells show 1–5 ratings; select an opportunity for its weighted points, decision and funding detail.

Criterion strength: 12345

How to read the scores and statuses
  • Ratings: Each criterion is scored 1–5. Higher is stronger.
  • Weighted score: Rating ÷ 5 × weight, added across six criteria. Maximum: 100. Missing ratings mean an incomplete score.
  • Status: Approved = funding authorized for the stated stage; On Hold = conditions unresolved; Not Approved = rejected or merged; In Review = awaiting a decision.
  • Decision: Scores guide priorities. Required reviews and funding approval remain separate. Economic ratings are provisional.

Architecture hypotheses

Choose the simplest fit.

Likely technology, not a build commitment

Overlapping technology candidatesRAG, predictive machine learning and agentic orchestration can overlap. Counts represent exact combinations. Select a region or use the adjacent technology buttons.RAGPredictive MLAgentic orchestration

Exact-combination counts · circle areas are not proportional · includes rejected concepts to show portfolio exposure

Explore the technology mix

RAG retrieves source content for generated answers. A controlled AI workflow can use RAG or predictive models; an agent selects steps and actions. These choices overlap.

    Technology tags are analyst hypotheses inferred from proposed approaches. Agentic candidates require human oversight; rejected proposals are not recommendations. Always compare with a non-AI alternative.

    Data lineage and scope

    All values, decisions and opportunity records are synthetic. Source: the Healthcare AI Portfolio Interview Demo workbook and its JSON export (16 November 2026), with OPP-002 revisions from this case study and four new proposals dated 17 November 2026. No records were pulled from Google Docs, Granola Notes, Slack, Zoom or FinOps systems.

    Business categories and technology tags are added presentation classifications. OPP-002’s $45,000 seed request replaces its original $63,000 request; other original requests retain their 2027-Q1 incremental-work basis. New requests are illustrative 2027-Q1 discovery/pilot envelopes, not validated estimates. These heterogeneous requests need scope reconciliation before budget allocation.

    Rejected OPP-016 and OPP-024 and merged OPP-022 are excluded from active request totals, but remain visible in the heat map and technology view. Actual spending and time-series decision history are unavailable, so the charts show current concentration and decision patterns—not historical trends.

    Architecture references: workflows and agents · retrieval-augmented generation.