Evidence Echo — Overview
AI Coding Echo
- Tier A Qodo A/B test of a GEPA-optimized review prompt: cost per run $0.24 → $0.12, miss rate 7.83% → 4.46%
- Tier A Cursor harness changes, A/B tested: user token costs −7%; dynamic MCP tool loading −46.9% tokens
- Tier A Controlled user study: developers with an AI assistant wrote less secure code and were more confident it was secure
- Tier A Controlled Copilot experiment: developers finished an HTTP-server task 55.8% faster than the control group
- Tier B Developers ready to let AI write application logic: 18% in 2025 → 50% in 2026
- Tier B Survey of 15,000+ developers: ~47% of their code is fully written by AI agents; 90% use agents weekly
- Tier B Gemini 4 Argon on Sonar's Java benchmark: 85.66% pass rate; vulnerabilities are 1.9% of findings but 28% of blocker-level ones
- Tier B GPT-6.1 Sol vs GPT-6 Sol on Sonar's Java benchmark: pass rate 85.29% vs 83.09%, 39.4% more code, total findings +31.7%
- Tier B Qodo survey of 800 U.S. developers and leaders: 89% of organizations report an AI-related production incident
- Tier B Coding agents on SWE-Bench Pro: splitting each issue into 3 messages made runs 2.0–2.6× more expensive per task
- +19 more in the report
CRO Echo
- Tier A Personalization vs shipping the best single variant: +18% in one field study, +4% in another
- Tier A AI vs human interview moderators (randomized, n=20): participants spoke 45% less to AI
- Tier A Simplified homepage (checklist + shop button) for The Raw Honey Shop: +24% sales vs control
- Tier A Free included services listed in the checkout order table: +9% orders vs control
- Tier A Redesigned Breakfast Train page with live calendar and menus: +10% bookings vs original
- Tier A Tutor sign-up page reframed as a job posting with a real student request: +15% applications
- Tier A Gift add-ons moved to their own funnel step: +28% revenue per order, no conversion loss
- Tier A Promo codes shown only to AI-detected hesitant visitors at Orange: +11.6% conversion vs no promo
- Tier A Trial button 'Try it free for 14 days' vs 'Start free trial' at Exclaimer: +39% trial sign-ups
- Tier B Mass Merchant UX benchmark of 10 sites: only 1 rated good across 5,600+ UX scores
- +29 more in the report
Exercise Echo
- Tier A Moderate vs low-intensity cycling in stress-related exhaustion: lower post-exercise HRV at 55% VO2max, not in controls
- Tier A Creatine with 6 weeks of resistance training in perimenopausal women: +7.4 points sleep consistency vs placebo, no cognitive edge
- Tier A HIIT in D-galactose-aged rats: reversed liver SIRT1/P16 changes and oxidative stress markers
- Tier A Evening HIIT session: next-morning flow-mediated dilation 7.6% vs 4.7% without exercise
- Tier A 14 weeks of periodized max-strength vs strength-power training in 80 athletes: both raised squat 1RM, neither clearly superior
- Tier A Hyperoxic vs hypoxic Tabata HIIT for 3 weeks: VO2max 37.8 to 41.5 vs 38.3 to 39.8 mL/kg/min
- Tier A Adding blood flow restriction to on-field team-sport training: relative mean power -0.50 W/kg in favour of BFR, low certainty
- Tier A Exercise prehabilitation before ankle ligament surgery in 276 patients: 2-year FAOS 92.4 vs 80.0 with standard care
- Tier A Creatine-based multi-ingredient supplement over 4 weeks: no extra leg-press 1RM gain vs placebo
- Tier A Caffeine mouthwash before a Wingate test: only a 30-second rinse raised peak power vs placebo
- +57 more in the report
SEO Echo
- Tier B Ahrefs survey of 301 marketers: 69% publish a larger share of AI content than a year ago
- Tier B Insurance AI-search benchmark: Google AI Mode drives 43.9% of citations; AI Overviews on 40.7% of searches
- Tier B Nonprofit AI-search benchmark: Google AI Mode drives 50% of citations; AI Overviews on 44.2% of searches
- Tier B Search Signals Index, Aug 10-31 2026: index -2.4% across 1,842 companies; Amazon -16.9% to a record low
- Tier B ChatGPT query fan-out across 66.6M responses: only 14.4% search the web, ~1.6 queries each
- Tier B AI Overviews on branded Google searches: 61.3% in July to 73.3% in September 2026 (82.91% on Sept 29)
- Tier B AI Overviews on 400 commercial keywords: 45.75% to 88% in a year; 'review' queries now ~90%
- Tier B 1,000 transactional prompts: 52.5% of the most-cited domains in LLM answers are affiliate or sponsorship sites
- Tier B Hidden AI traffic in Direct: only 0.003% of Direct sessions across 12 sites; no reliable AI-visibility link in 632 tests
- Tier B LLMs.txt across ~300,000 domains: 10.13% adoption, no measurable effect on AI citations
- +11 more in the report