LLMs Optimization Report • Non-Profit (Sample)

GlobalHealth Access — LLMs Visibility & Accuracy Readout by OneMind Strata

Integrated findings from interviews, AI engines, and benchmark data. Produced with OneMind Strata confidential frameworks (sample data for demonstration).

Overall AI Visibility

0%

+11 pts vs. prior 30d

Citation Quality (idx)

0

↑ authority & recency

Accuracy Score

0.00

−28% hallucination

Executive Summary

LLMs Performance Overview

GHA appears in a majority of health-program and donation-intent queries across ChatGPT, Claude, and Perplexity; Gemini and AI Overviews trail due to uneven schema and fewer proof points.

Strengths: credible program pages, annual impact report, board governance transparency.
Gaps: regional FAQs, program cost benchmarks, localization of success metrics.

Key Metrics Summary

Metric	Value	Note
Mention frequency	58%	Across 120 test prompts
Citation quality	76	↑ official docs, ↓ third-party blogs
Accuracy	0.71	Improving; errors mainly on legacy program names

Strategic Recommendations (Top 5)

Canonize program names/IDs across all properties.
Publish 3 proof packs (clinic outcomes, cost per beneficiary, governance score).
Ship FAQ + How-it-works schema on program, donation, and transparency pages.
Localize top pages (ES/FR) with hreflang and region-specific stats.
Add explicit “Not provided / Limitations” blocks to reduce AI gaps.

ROI & Impact Assessment

+12–18% visibility across donation & program queries (90d).
2.1× faster donor Q&A handling via accurate AI references.
−30–40% misinformation on program names and coverage areas.

Current State Analysis

AI Presence Audit

Presence (%) Accuracy

Query Coverage & Competitive Positioning

GHA Peer A Peer B

Citation Quality Review

76 index: majority of citations now point to /programs, /impact, /governance.
23% still route to media/blog summaries; add canonical links & updated dates.

Content Performance Evaluation

High-Performing Content

Asset	AI Citations /100	Why it works
Annual Impact Report 2024	2.8	Clear stats, executive quotes, timestamps
Community Clinics page	2.3	FAQ schema + outcomes per $100 donated
Governance & Board	1.7	Transparency signals; policy links

Content Gaps & Structured Data

Under-cited: Donation Fees, Program Cost Benchmarks, Volunteer Training.
Schema: add HowTo, FAQ, NGO attributes; ensure JSON-LD parity on ES/FR.
Authority: add peer-review links for health outcomes; cite WHO/UNICEF benchmarks.

Technical Infrastructure Review

Website & Semantic SEO

Optimization -friendly markup on /programs and /impact; extend to /donate and /volunteer.
Topic graph: strong “primary care” & “maternal health”; weak “health financing”.

Citation Network & Multi-Platform

New backlinks from DevPolicy & Global Health Now; add rel=canonical on mirrored posts.
Ensure consistent org profile (address, EIN, board) across website, Medium, LinkedIn, Charity Navigator.

AI Engine-Specific Findings

Platform Breakdown (sample)

Engine	Visibility	Accuracy	Notes
ChatGPT	63%	0.78	Strong citations to /impact
Claude	60%	0.75	Good exec quote pull-through
Gemini	52%	0.66	Over-weights media posts
Perplexity	61%	0.72	Diverse sources; keep recency fresh
AI Overviews	49%	0.63	Locale variability; add `hreflang`
Grok	55%	0.68	News-driven; lacks proof packs

Query Type Performance

Program Donate Impact Governance

Context & Positioning

Engines frame GHA as a community-health implementer with strong maternal-health outcomes and transparent governance; expand positioning to include health financing efficiency and training scale.

Competitive Intelligence

Market Share in AI Responses

GHA Peer A Peer B Others

Best Practices & Risks

Peers: concise program one-liners, dated stats, and PDF alt-HTML snippets for crawlability.
Opportunity: donor-intent FAQs (fees, impact per $) seldom cited—win quick trust.
Threat: brand aliasing in older media—add redirects + canonical to preferred names.

Implementation Roadmap

Priority Action Items

Horizon	Action	Owner	Success Metric
Now (0–30d)	Program ID canon + FAQ schema	Web / Comms	+8 pts visibility
Next (30–90d)	3 proof packs + ES/FR localization	Impact / Comms	+0.06 accuracy
Later (90–180d)	Health financing explainer series	Programs	2× citations to finance pages

Resources, Timeline & KPIs

Resources: 1 FTE content strategist, 0.5 FTE web dev, analyst support.
Timeline: phased sprints aligned to report sections.
KPIs: visibility, accuracy, citation authority, donor-intent CTR.

Appendices

Methodology Documentation

Interview protocol (executive, field ops, donor relations); coding rubric & inter-rater notes.
AI engine setup: prompt battery (120 queries), sampling cadence, validation workflow.
Benchmark sources: NGO databases, filings, third-party health outcome studies.

Templates & Monitoring

Content checklists: program pages, FAQ, proof packs, governance.
Monitoring: weekly AEO crawl, schema diff, citation drift alerts, accuracy spot checks.