What's new
Changelog
Every change a customer could notice, dated. Internal work stays out. Methodology changes also bump the version stamped on every report, so two reports are only compared like for like.
7 October 2026Improved
Reports say only what they measured
- The Deep preview no longer shows example per-engine numbers; a paid report now contains no figure that was not measured for your brand.
- The sample PDF is now the same document a customer receives, built by the same renderer.
- White-label reports carry no LensCore pricing, unlock buttons or share prompts — only one 'Measurement by LensCore' line.
- The Discovery Index is labelled a weighted blend rather than experimental: the five inputs are measured, the weighting is our judgement, and that does not change with time.
- The home page now says the ± range on AI visibility is the spread across engines, and that run-to-run volatility is measured by monitoring re-runs rather than a one-off audit.
17 September 2026Honesty
An answer we cannot read is not a zero
- When an engine answers but in a form we cannot read — another script, a refusal, an essay with no list — the question is now recorded as not observed, with the reason, instead of counting as a miss.
- Every percentage in the report now divides by the questions actually measured, and the report says how many answers were excluded.
- Brand names written in Indian scripts are matched as written.
10 September 2026Improved
Screenshots move to a proper store; Ctrl+K; docs, resources and notices pages
- Report screenshots are now served from our asset store instead of being embedded in the page — report pages are a fraction of their old size on mobile data.
- Ctrl+K (⌘K on Mac) opens a jump box to any page.
- New pages: Docs (how every number is measured), Resources (guides, articles, glossary), Changelog, Open-source notices.
- Pricing now spells out what each tier includes and leaves out, and confirms Deep includes the full Audit report.
- Locked reports show three rivals instead of one; sample reports say plainly that the brands shown are public references we are not affiliated with.
9 September 2026New
Local 10-second check, public leaderboard, creator monitoring
- Local businesses: type your business, city and kind of place; get four city-specific AI answers in about ten seconds.
- Leaderboard: who AI names, grouped by category and market, from real scans. Brands appear by name only when their owner made the report public.
- Creators: brand-side vs audience-side counts, an entity-kit checklist, and ₹499/month monitoring.
- Shopify stores are detected automatically and get three Shopify-specific findings that name the exact admin setting to change.
- Agencies: a workspace with white-label reports and PDFs, teammate invites, one client dashboard and a monthly prospect-scan allowance.
9 September 2026Honesty
Google AI Mode joins paid reports; live-search honesty
- Paid reports (Audit, Deep, re-scans) now include a sixth engine, Google AI Mode. The free scan stays at five.
- Every answer records whether the engine actually searched the live web. The report shows “Answers from live search — N of M”.
- The fifth engine is now called the Lenscore Retrieval Engine everywhere.
- Monitoring copy corrected to what it does: weekly re-scans, alerts by email.
6 September 2026Improved
Adaptive questions, KPI split, security headers
- Each scan builds a 25-question benchmark: 10 fixed core questions plus 15 adapted to what your site actually sells and where.
- New measured KPIs shown above the composite: recommendation rate, recommendation share of voice, ordinal prominence, citation rate and share, entity resolution. Anything unmeasured shows as not measured, never as zero.
- The composite is labelled “LensCore Discovery Index · experimental” with a formula version.
- Free scans ask 8 of the 25 questions; the Audit unlock tops up to all 25.
4 September 2026Fixed
Reprice and fourteen live fixes
- Audit ₹2,499 / $49, Deep ₹5,999 / $129, Founder Bundle ₹7,999 / $69, Monitoring ₹1,999 / $39 per month.
- All five engines measuring all 25 questions on every paid report; Retrieval Engine moved to a stronger index.
- Fourteen fixes from a live-site review: broken footer routes, sample labelling, run counts, zero-vs-not-observed wording.
Want a change you don't see here? Tell us. Numbers behind the measurement live in the docs.