Practical essays on the problems AI is creating in marketing right now, answer engine optimization, brand voice, cold email, and the tools I build to solve them. No fluff, no gated PDFs. Each post ends somewhere you can act.
The data pages live separately: the question index (120 questions customers ask AI, mapped to the fix that answers each) and the AI Visibility Index (35 Indian clinic sites scanned, scores published). Both free, both verifiable.
The 15 questions people type into ChatGPT before booking a dietitian (fee, online consults, PCOS/thyroid/diabetes, vegetarian Indian plans, RD credentials, follow-ups), each with its on-page fix. Grounded in the 16-clinic scan: dietitians top every vertical for AI readability, yet only 3 of 16 carry the FAQ markup that answers these directly.
Read Case studyThe 15 questions a business owner types into an AI before shortlisting an agency (cost, retainer vs project, services, industries, results), each with the plain-text fix on the agency's own homepage. Grounded in the 173-agency scan: median 89, but only 31 of 173 carry the FAQ schema that answers these directly.
Read Case studyThe 15 questions a customer types into an AI before booking (haircut price, timings, walk-in vs appointment, home service, bridal package, nearest branch), each with the plain-text homepage fix. Grounded in the 14-salon scan: median 80.5, and not one site (0 of 14) has the FAQ schema these questions need.
Read Case study14 salon and spa chain homepages, live scan. Median 80.5, one of the weaker verticals (20th of 28): 4 of 14 below a B, structured data not universal (10 of 14), and not one site (0 of 14) has the FAQ schema the price/timing/appointment questions need.
Read Case studyThe 15 questions a parent types into an AI before shortlisting a preschool (fees, age, timings, daycare hours, CCTV, teacher ratio, meals, admissions), each with the plain-text homepage fix. Grounded in the 14-preschool scan: median 87, but only 3 of 14 carry the FAQ schema these questions need.
Read Case study14 preschool and daycare chain homepages, live scan. Median 87 (9 of 14 grade A) and structured data near-universal (13 of 14), but only 3 of 14 carry FAQ schema, the format the fee/age/timings/safety questions parents ask AI actually need.
Read Case studyThe biggest scan I have published: 173 digital agency homepages across 32 cities. Median 89 puts the firms who sell findability 5th of 28 verticals, but it is not uniform. One in five scored below a B (the range runs down to 23), and only 18 percent (31 of 173) publish FAQ schema, the vertical's clearest gap.
Read Case studyA live scan of 16 independent dietitian and nutrition homepages across 8 cities, and the strongest vertical I have tested yet: median 91.5, 14 of 16 grade A, structured data near-universal (15 of 16). It ties the top of my 28-vertical ranking. The one gap left is FAQ schema (3 of 16), on the profession that answers the most FAQ-shaped questions of all.
Read Case studyThe 15 questions parents type into ChatGPT before picking a clinic (timings, consultation fee, vaccine schedule, Sunday and emergency hours, teleconsult, the doctor's qualifications), each with its on-page fix. Grounded in the 17-clinic scan: the content is there, but only 3 of 17 carry the FAQ markup that answers these directly.
Read Case studyA live scan of 17 independent pediatric and child-specialist clinic homepages across 10 cities. Median 77, but the sample splits in half on one signal: the 9 with structured data score a median 92, the 8 without collapse to about 54. Only 3 of 17 carry FAQ schema, the format built for exactly the questions parents ask AI.
Read Case studyA live scan of 15 coworking and flexible-workspace homepages, and the first vertical where structured data is basically solved: 14 of 15 have it, median 87, no failing grades. So the usual structured-data-is-the-divide story flips here. The real gap is one level up: FAQ schema (2 of 15) and a single clean H1 (9 of 15).
Read Case studyThe 15 questions founders type into ChatGPT before choosing a space (hot-desk price, day pass, locations, meeting-room rates, GST or virtual office, lock-in), and the on-site fix for each. Grounded in the 15-brand scan: structured data is near-universal, but only 2 of 15 carry the FAQ markup that answers these.
Read Case studyA live scan of 18 homeopathy clinic homepages, and a two-tier field: median 86, but the mean is only 77.8. The 11 clinics with structured data sit at a median 92; the 7 without it collapse to a median 59, close to 40 percent of the vertical. And despite homeopathy being one of the most question-driven fields in medicine, only 2 of 18 carry FAQ schema.
Read Case studyThe 15 questions patients type into ChatGPT before choosing a homeopathy clinic (does it work for thyroid or hair loss, is it safe for kids, can I take it with allopathic medicine, cost, timeline, BHMS or MD qualification), and the on-site fix for each. Grounded in the 18-clinic scan: only 2 of 18 carry the FAQ markup that answers them.
Read Case studyA live scan of 18 immigration and visa consultant homepages, and the most AI-readable vertical I have measured yet: median 91.5, 14 of 18 grade an A, two perfect 100s. The likely reason is the business itself (eligibility, points cutoffs, timelines and fees are exactly what an AI reads well). Structured data still sorts the bottom three, which scored 43, 53 and 58.
Read Case studyThe 15 questions applicants type into ChatGPT before hiring a visa consultant (Canada or Australia PR eligibility, CRS cutoff, fees, timeline, ICCRC or MARA registration, refund on refusal), and the on-site fix for each. Grounded in the 18-firm scan: only 7 of 18 carry FAQ markup, the highest adoption seen but still under half.
Read Case studyA live scan of 20 wealth management and financial advisory homepages, and one of the highest scoring verticals I have measured: median 88, 14 of 20 grade an A, not one fails. Structured data still sorts the bottom (the 3 firms without it scored 57, 57 and 61), and firm size does not predict readability: a solo fee-only planner scored a perfect 100, a large family office a 57.
Read Case studyThe 15 questions investors type into ChatGPT before hiring an advisor (SEBI registration, fee-only vs commission, minimum portfolio, fiduciary duty, custody, NRI service), and the on-site fix for each. Grounded in the 20-firm scan: only 4 of 20 carry the FAQ markup that answers them.
Read Case studyA live scan of 15 interior design homepages. Median 78, but bimodal: the tech-enabled platforms hit a median 92 (one perfect 100) while the design-led studios trail. Structured data splits the field 87 vs 23, a 64-point gap, the widest I have measured. The irony: visual-craft firms build the least machine-readable sites.
Read Case studyThe 15 questions homeowners type into ChatGPT before hiring a designer (2BHK cost, price per square foot, timeline, modular kitchens, warranty, EMI, turnkey scope), and the on-site fix for each. Grounded in the 15-firm scan: only 3 of 15 have FAQ markup.
Read Case studyA live scan of 16 CA and accounting homepages. Median 87, a top professional-services vertical (against law firms at 59.5), thanks to the digital-first compliance platforms. Structured data (11 of 16) is the divider; FAQ schema, the cheapest fix, sits at 1 of 16.
Read Case studyThe 15 questions founders type into ChatGPT before hiring an accountant (best CA for a startup, ITR and audit fees, GST registration, company registration, virtual CFO, tax notices), and the on-site fix for each. Grounded in the 16-firm scan: only 1 of 16 has FAQ markup.
Read Case studyA live scan of 20 consultant homepages. Median 90, one of the strongest verticals I have measured (level with IVF), and the first non-medical category to reach that level: 14 of 20 scored an A, two hit a perfect 100. Structured data is nearly universal (18 of 20); the gap left is FAQ schema (5 of 20).
Read Case studyThe 15 questions students and parents type into ChatGPT before shortlisting a consultant (best country for my profile, service fees, visa success rate, scholarships, IELTS prep, refund policy), and the on-site fix for each. Grounded in the 20-site scan: only 5 of 20 have FAQ markup.
Read Case studyA live scan of 18 law firm homepages. Median 59.5, one of the weakest verticals I have measured, split hard by firm size: national firms median 83, mid-size and boutique far below. Structured data (7 of 18) is the dividing line, and FAQ schema sits at zero of 18, the whole vertical.
Read Case studyThe 15 questions clients type into ChatGPT before shortlisting a firm (practice areas, M&A in Mumbai, SEBI matters, fees, jurisdiction, free consultation), and the on-site fix for each. Grounded in the 18-firm scan: zero of 18 have FAQ markup.
Read Case studyA live scan of 18 ayurveda and panchakarma homepages. Median 87, among the strongest verticals I have measured: 10 of 18 scored an A. Structured data is nearly universal (16 of 18), so the remaining gaps are FAQ markup (2 of 18) and a single clear H1 (4 of 18).
Read Case studyThe 15 questions patients type into ChatGPT before booking ayurveda treatment (which panchakarma, BAMS or MD doctors, authentic Kerala lineage, program length, cost, residential or outpatient), and the on-site fix for each. Grounded in the 18-site scan: only 2 of 18 have FAQ markup.
Read Case studyThe 15 questions people type into ChatGPT before joining a gym (membership cost, 24 hour access, personal trainer, trial pass, timings, classes), and the on-site fix for each. Grounded in the 16-gym scan, the lowest-scoring vertical: only 1 of 16 has FAQ markup and 4 have a clean H1.
Read Case studyA live scan of 16 gym and studio homepages. Median 53.5, the lowest-scoring vertical I have measured: 5 of 16 scored an outright F. Structured data splits the field 87 vs 39, and the boutique studios a "gym near me" search should surface are near-invisible.
Read Case studyThe 15 questions patients type into ChatGPT before booking a physio (session cost, home visits, back pain, sciatica, sports injury, post-surgery rehab), and the on-site fix for each. Grounded in the 15-clinic scan: only 3 of 15 have FAQ markup, the clearest gap even in a strong field.
Read Case studyA live scan of 15 physio and rehab homepages. Median 87, one of the stronger verticals: structured data is near universal (12 of 15) and, unlike labs and vet hospitals, the standalone clinics score almost as well as the chains. The gap that remains is FAQ markup, on just 3 of 15.
Read Case studyThe 15 questions pet owners type into ChatGPT before picking a vet (emergency hours, vaccination cost, home visits, surgery, is the clinic open now), and the on-site fix for each. Grounded in the 15-hospital scan: only 1 of 15 has FAQ markup and only 7 have a clean H1.
Read Case studyA live scan of 15 vet and pet hospital homepages. The most polarized field yet: 7 score 87 or higher, the other 8 score 66 or below, zero B grades in between. Structured data splits them 90.5 vs 46, and the standalone hospitals a "vet near me" search should surface score lowest.
Read Case studyThe 15 questions patients type into ChatGPT before picking a lab (test price, home collection, NABL accreditation, report turnaround, timings), and the on-site fix for each. Grounded in the 18-lab scan: only 3 of 18 have FAQ markup and only 5 have a clean H1.
Read Case studyA live scan of 18 lab homepages (national chains, regional labs, at-home collection). Median 79 but a 23 to 96 spread: structured data splits the field 83 vs 46, only 5 of 18 have a clean H1, and the local labs a "near me" search should surface score lowest.
Read Case studyThe 15 questions students and parents type into ChatGPT before choosing a coaching institute (fees, batch dates, results, scholarships, refund policy), and the on-site fix for each. Grounded in the 21-site coaching scan: 0 of 21 have FAQ markup and only 6 have a clean H1.
Read Case studyA live scan of 21 coaching and edtech homepages. Median 83 but the widest spread yet (23 to 92): the biggest JEE/NEET brands score lowest, UPSC institutes highest, and 0 of 21 have FAQ markup.
Read Case studyThe 15 questions LASIK and cataract patients type into ChatGPT (cost, safety, recovery, lens types, financing), and the on-site fix for each. Grounded in the 18-clinic eye scan: schema is near universal but only 1 of 18 has FAQ markup.
Read Case studyA live scan of 18 real eye and LASIK clinic homepages. Median 87, among the strongest verticals: structured data is near universal (17 of 18), but only 1 of 18 answers patient questions in FAQ schema.
Read Case studyThe 15 questions cosmetic surgery patients type into ChatGPT (rhinoplasty cost, recovery, scars, financing), and the on-site fix for each. Grounded in a scan of 15 real clinic sites: schema is near universal (14 of 15) but only 1 of 15 has the FAQ markup that answers these.
Read Case studyThe 15 questions derm patients type into ChatGPT (acne cost, pigmentation, PRP, laser), and the on-site fix for each. Grounded in a scan of 17 real clinic sites: median 89, yet only 1 of 17 carries the FAQ markup that answers these.
Read Case studyEvery real scan in one ranked table. Immigration and visa consultants top it (median 91.5, ahead of study-abroad and IVF at 90); gyms, law firms and veterinary hospitals sit at the bottom. Structured data is 73% solved; FAQ markup, at 14%, is the gap everyone shares.
Read Case studyMedian 87, tied with cosmetic surgery, just below the top aesthetic trio (IVF 90, hair transplant and derm 89). Schema is near-universal (13 of 14); the blind spot is FAQ markup, which only 2 of 14 carry.
Read Case studyMedian 87, strong but the lowest of the four aesthetic verticals (IVF 90, hair transplant and derm 89). Schema is near-universal (14 of 15); the real gap is on-page basics: only 7 of 15 have a clean title tag.
Read Case studyMedian 90, the highest of any vertical I've scanned. All 14 have structured data (the first fully-solved vertical), so the story flips: the edge has moved to FAQ markup, which still only 6 of 14 carry.
Read Case studyMedian 89, tied with hair transplant for the highest vertical yet. But structured data still splits it (the 4 sites without it are the 4 lowest scorers), and only 1 of 17 has FAQ markup, the biggest opening left.
Read Case studyI ran 36 fast-growing startups through my AI readability engine. Of the 24 I could read: median 81, a third below a B, 88% with no FAQ schema and 42% with no structured data at all. Being funded does not make you readable to the AI your buyers ask first.
Read Case studyThe 15 questions hair-transplant patients type into ChatGPT (cost, grafts, timeline, pain), and the on-site fix for each. Grounded in a scan of 13 real clinic sites: median 89, the highest vertical, yet only 4 of 13 carry the FAQ markup that answers these.
Read Case studyI scanned a real dental clinic, found two fixable flags, fixed them, and rescanned. Two structural edits moved it from 81 to 92, B to A, in three days.
Read Case studyEvery real scan this month in one picture: 93 sites across 5 verticals. Structured data is the divider everywhere; FAQ markup is the blind spot only 7 of 77 have fixed.
Read Case studyMedian 89, the highest of any vertical I've scanned. Structured data is near-universal here; the open gap is FAQ markup. What a competitive paid-search vertical already gets right.
Read Case studyMedian 79, a 31.5-point structured-data gap (the widest of any vertical yet), and all 15 missing FAQ markup. Why AI sends diners to Zomato, not your site.
Read Case studyThe heavy questions IVF patients ask ChatGPT (success rates, cost, IVF vs IUI), and the on-site fix for each. Grounded in the clinic study, built on honest sourced data.
Read Case studyThe 15 prompts diners type into ChatGPT to pick a restaurant, and the exact on-site fix for each. Grounded in a fresh scan of 15 real restaurant sites (median 79, a 31.5-point structured-data gap).
Read Case studyThe 15 questions travelers type into ChatGPT to pick a hotel, and the exact on-site fix for each. Grounded in a scan of 25 real hotel sites.
Read Case studyMedian 76, a 27-point structured-data gap, and every one of the 25 missing FAQ markup. Why AI travel answers cite the OTA, not the hotel.
Read Case studyHalf scored an A, but the sites without structured data sat 30 points lower. The AI-visibility gap is legibility, not market maturity.
Read How-toThe single highest-impact AI-visibility fix, with a paste-ready JSON-LD example and the one rule that keeps it safe.
Read MarketingThe 15 questions dental patients actually ask ChatGPT, and the exact on-site fix that lets the answer be your clinic. Grounded in a scan of real clinic sites.
Read MarketingCompare free AEO checker tools and learn how to make pages easier for ChatGPT, Perplexity, Claude, and other answer engines to understand and cite.
Read MarketingLearn practical AEO fixes that make webpages easier for AI answer engines to read, quote, and cite.
Read MarketingDiagnose subject line, personalization, CTA, length, and copy problems that stop cold emails from getting replies.
Read MarketingUse a five-part cold email format: trigger, relevance, pain, proof, and ask. Includes examples and a scoring checklist.
Read MarketingCompare brand voice analyzer tools and learn what a useful tool should extract from website copy.
Read MarketingLearn how to analyze homepage, product, email, and founder copy to create usable brand voice guidelines.
Read MarketingCompare name-only generators with mini-brand generators that create name directions, logo marks, colors, fonts, and voice.
Read MarketingStart with one seed word and create a name direction, visual metaphor, palette, font pair, logo mark, and voice.
Read MarketingLearn what reverse-prompt tools can and cannot infer from AI-written text, with examples and reusable prompt patterns.
Read MarketingLearn what websites can see about your browser and compare free tools for privacy, fingerprinting, and browser exposure checks.
Read MarketingA practical AI marketing tool stack for Indian founders and small teams across AEO, brand voice, cold email, prompts, content, analytics, and CRM.
Read CareerA practical hiring guide for founders and marketing leads evaluating AI marketers, growth marketers, automation specialists, and fractional CMOs.
Read CareerLearn what to include in a marketing portfolio: campaigns, tools, teardowns, experiments, case studies, and before/after work.
Read CareerCompare fractional CMOs and growth marketing consultants by strategy, execution, team design, experiments, and startup stage.
Read CareerWhy deterministic marketing tools can be useful in an AI-heavy market, and what the vishkul.com tool suite says about practical marketing philosophy.
Read MarketingA practical, low-cost AI marketing stack by stage: positioning, analytics, content, cold outreach, AEO, CRM, and experiments.
Read MarketingRun a three-layer landing page audit: AI-search readability, brand voice, and conversion clarity. Includes checklist and example teardown.
Read MarketingSee cold email before-and-after examples for founder sales, job outreach, freelance pitches, partnerships, and LinkedIn follow-up.
Read CareerLearn how to write marketing resume bullets around revenue, pipeline, CAC, conversion rate, experiments, content, and systems built.
Read CareerLearn how marketers can build a stronger personal brand through tools, teardown libraries, case studies, public experiments, essays, and mini-products.
Read
Tools at vishkul.com/tools · portfolio at vishkul.com
Built by Vishesh Kulshrestha in Bengaluru · vishkul2905@gmail.com