# Ankit Kedia — Full Site Content > Complete text of every article and client case study on ankitkedia.com. Generated 14 September 2026. Source: https://ankitkedia.com --- # Articles ## Answer Engine Optimization: Getting Your Business Cited by AI Search Category: SEO & AEO | Published: August 2026 | URL: https://ankitkedia.com/blog/aeo-guide Someone opens ChatGPT and asks which company supplies premium office furniture in Telangana. An answer comes back. It names two or three businesses, and it quietly leaves out everyone else. Getting into that answer is what answer engine optimisation means. ### Search has split into two jobs Ordinary SEO gets you into a list of links. AEO gets you inside the answer itself, in ChatGPT, Perplexity, Google's AI Overviews, and whatever assistant your customer happens to be speaking to. The two overlap heavily. But AEO rewards something quite specific, which is short, checkable, well-organised statements of fact. ### What actually works #### Answer the question people asked These systems look for passages that settle a query, so write passages that settle queries. A FAQ block asking "what does this company supply?" followed by two plain sentences naming every service will get picked up. Three paragraphs about your journey of excellence will not. #### Write your facts down somewhere crawlable This sounds obvious. Almost nobody does it. Somewhere on your site, in HTML that does not depend on JavaScript, there should be sentences stating what you sell, where you operate, who you serve, and what you have that competitors don't. "The only authorised dealer for two states" is a fact software can repeat. "Your trusted partner in excellence" is noise, and it is noise to human readers too. #### Give it structure - Real heading hierarchy, one idea per section - FAQ and How-To schema, where the markup honestly matches what is on the page - Lists and tables for anything countable - Titles, meta descriptions, Open Graph and locale tags filled in properly #### Make sure the content exists before the page finishes loading If your text only appears after a framework boots, some engines will never see it. Server-rendered HTML remains the safest way to carry a fact you want quoted. This is one of several reasons I keep clients away from heavy page builders. ### Measuring it Once a month, in a fresh session, ask the engines the questions your customers ask. Note three things. Whether you appear at all, whether you get a link, and what the engines get wrong about you. That third one matters most. Wrong answers almost always trace back to something missing or badly worded on your own site, which means you can fix them. There is no trick to any of this. You are writing down what is true about your business, clearly enough that software can repeat it without garbling it. The businesses that do this early will be the ones getting named while everyone else is still arguing about whether AI search matters. If you want to know what the engines currently say about your company, that is one of the checks in my audit. --- ## Core Web Vitals: What Actually Moves the Numbers Category: Performance | Published: July 2026 | URL: https://ankitkedia.com/blog/core-web-vitals Every audit I run opens with Core Web Vitals, because they sit in the rare overlap between what Google measures and what customers feel. The advice online is mostly recycled generalities, so here is what genuinely shifts the numbers. ### LCP is nearly always the hero image Largest Contentful Paint usually lands on the big image or headline block at the top of the page. Three fixes, in the order I try them. - Serve the right format at the right size. Run every image through a resize and WebP or AVIF conversion with explicit width, height and quality settings. The saving over an unoptimised JPEG is usually substantial, and on image-heavy catalogue sites it is the single biggest win available. - Preload the hero and never lazy-load it. Lazy-loading the first image on the page is the most common self-inflicted wound I see. It happens because a plugin applies lazy loading to everything by default and nobody exempts the top of the page. - Cut render-blocking CSS and JavaScript. One stylesheet, deferred scripts. A whole framework loaded so that a logo strip can scroll is a tax paid on every single visit. ### INP punishes JavaScript you forgot about Interaction to Next Paint measures how long the main thread makes people wait. The pattern I see in audit after audit is a site carrying three carousel libraries, two abandoned chat widgets, a tag manager stuffed with pixels from campaigns that ended years ago, and jQuery sitting alongside a modern framework doing the same job twice. Deleting the unused half is dull work. It is also free performance. ### CLS is about reserving space Layout shift comes from images with no dimensions, banners injected after load, and fonts that swap late. Set width and height attributes on everything. Reserve a slot for anything that arrives dynamically. Use font-display: swap with a fallback that has similar metrics so the reflow is small. ### Test on the network your customers actually use A score achieved on office fibre tells you very little. Most Indian traffic arrives on a mid-range Android phone over a patchy 4G connection, and that is the condition worth designing for. I hold builds to a sub-two-second load on a throttled connection rather than a green score in a lab run. ### Use field data, not lab data PageSpeed Insights reports both. The field data comes from real Chrome users and is the one Google acts on. Fix whichever field metric is worst, then wait. The reporting window rolls over 28 days, so a fix you ship today will not show up tomorrow, and people give up on good changes because of that lag. Performance work pays out slowly and quietly, which is exactly why so few competitors bother with it. --- ## AI Product Photography: Catalogue Imagery Without the Studio Day Category: AI Media | Published: June 2026 | URL: https://ankitkedia.com/blog/ai-product-photography A clothing brand wants its new colourway shown in six styled rooms. The traditional route is a location, a stylist, a photographer, two days of shooting, and the whole exercise again next season. For the D2C brands I work with, most of that now happens through a generative pipeline instead, and the cost difference is not close. ### How the pipeline runs - Start from honest photography. AI does not replace the product shot. It works from accurate, well-lit photos of the real item. Whatever inaccuracy you feed in gets amplified, so this step is not optional. - Generate the scene. Models place the product into styled settings, controlled by the brand's palette and art direction. Every colourway can be given identical treatment, which is something a series of physical shoots almost never achieves because light and styling drift between sessions. - Grade and check. Each image gets compared against the physical product for colour and texture accuracy, then graded to the brand look. Anything that misrepresents what will arrive in the box is thrown away. For e-commerce this is not negotiable, and it is where most cheap AI imagery falls over. - Add motion. Selected stills get animated into short vertical video for reels, marketplace video slots and paid social. ### What changes for the brand - A campaign concept becomes finished assets in days rather than weeks - The whole catalogue sits in one consistent visual world - Creative variants cost almost nothing, so teams actually run the A/B tests they always meant to run ### Where a camera still wins Founder portraits. Real customers using the thing. Products where the selling point is a texture you need to see at macro distance. Anything with legal exposure, like ingredient panels or certification marks, where an approximation is a liability rather than a shortcut. For most catalogues the sensible answer is a hybrid. Photograph the truth once, properly, then build the world around it generatively. That way the customer sees the actual product, and you are not paying for a location scout every quarter. If this is worth exploring for your catalogue, I usually start with a pilot batch of a few products so you can put the output next to your existing photography and judge it yourself. --- ## SEO for Indian SMEs: Where to Spend When You Cannot Spend Everywhere Category: SEO & AEO | Published: May 2026 | URL: https://ankitkedia.com/blog/seo-for-smes Most business owners I meet have already been through an SEO package. Monthly reports arrived, backlinks were promised, the phone did not ring. Search does work for smaller businesses, but only when the effort goes in the right order. This is the order I use. ### 1. Win your own name first Search your company name. If your site is not the confident first result, with the right phone number and hours attached, nothing further up this list matters yet. Fix your titles, your Google Business Profile, and make sure your name, address and phone number match everywhere they appear online. ### 2. Take the local, high-intent searches "Office furniture dealer Kolkata." "Travel agency Surat." "Nutritionist HSR Layout." Someone typing that is shopping, not browsing. Each of those queries deserves a real page with substance on it, not a line in a services list. Condition-specific and service-specific pages do a lot of quiet work here. ### 3. Say things AI assistants can quote A growing share of buying research now happens through ChatGPT and Google's AI Overviews. Those systems cite businesses whose sites state facts plainly. FAQ blocks answering the questions customers genuinely ask are cheap to write and currently under-used by almost everyone in the mid-market. It is the closest thing to an unclaimed channel I have seen in years. ### 4. Fix speed before you write anything A slow site wastes every rupee spent bringing people to it. Sub-two-second loads on mobile, images in modern formats, no plugin sprawl. Speed helps rankings, but the reason to do it is that people leave. ### 5. Then content, and only what you can sustain Blogs work when they answer questions your customers actually ask, in your own voice, with your own examples. One decent article a month beats twelve outsourced ones assembled around keywords. The easiest place to start is to write down what you already explain on the phone five times a week. ### What to skip - Backlink packages and domain-authority schemes. Risk without an asset at the end. - Rank reports for keywords that never convert. - Bulk AI-written articles that read like everyone else's bulk AI-written articles. Name, local intent, quotable facts, speed, then content. Kept in that order it builds on itself. Taken out of order it mostly burns budget, which is why so many small businesses conclude that SEO does not work for them. --- ## Seven Findings That Turn Up in Almost Every Website Audit Category: Audits | Published: April 2026 | URL: https://ankitkedia.com/blog/website-audit-checklist Every website is different. Their problems are not. These seven turn up in the large majority of audits I run, so it is worth checking your own site first. ### 1. The hero image is lazy-loaded Your most important image, deliberately delayed. It damages LCP and takes one attribute to fix. If your theme lazy-loads everything by default, exempt the first screen. ### 2. Titles and descriptions are duplicated or left on defaults "Home | Company Name" tells a search engine nothing and earns no clicks. Every page needs a title naming the actual offer and the actual place, and a description written the way you would write an ad, because that is effectively what it is. ### 3. JavaScript nobody uses any more Old chat widgets, A/B tools from a test that finished in 2022, a tag manager full of dead pixels, two animation libraries doing similar things. Each one costs main-thread time on every visit. Audit your tags twice a year and delete without sentiment. ### 4. Adjectives where facts should be Pages full of excellence and passion, and nowhere a plain sentence saying what you sell, to whom, where, and roughly at what price. Human visitors skim past this. AI assistants cannot quote it at all. Specifics convert and vagueness does not, which has been true since long before anyone worried about answer engines. ### 5. Images several times larger than they display A 4000px photograph rendered in a 400px card, as a JPEG, over mobile data. A resize-and-convert pipeline usually removes most of a page's weight in an afternoon. ### 6. Forms that ask far too much Eleven required fields before someone can request a quote. Every field costs you enquiries. Name, one contact method, one question. Earn the rest on the call. In India a WhatsApp link with the message pre-filled will often outperform the form entirely, and it costs nothing to add. ### 7. Proof kept away from the promise Claims live on the homepage, testimonials sit on a page nobody visits, case studies are three clicks deep. Proof belongs next to the claim it supports. Client logos on the service page, results beside pricing, reviews at the moment someone hesitates. ### How did you do Five or more of these and a structured audit will pay for itself quickly, whether you get it from me or someone else. The value is not the list of problems. It is knowing which two to fix first, because that ordering is where most of the return sits. --- ## E-commerce Conversion: Boring Repairs Beat Expensive Redesigns Category: E-commerce | Published: March 2026 | URL: https://ankitkedia.com/blog/ecommerce-conversion Store owners ask for redesigns. Stores usually need repairs. Conversion problems cluster in three places, and most of the fixes are cheap and boring, which is good news for everyone except the agency quoting for a rebuild. ### Product pages: answer the question that stops the sale A visitor leaves at the exact moment a question goes unanswered. The ones I see missing most often: - What am I actually getting? Dimensions, materials, what is in the box, how to care for it. Scannable specifications, not a paragraph. - Can I trust these photos? Multiple angles, texture close-ups, the product in a real setting. This is where good product imagery earns its cost, because doubt is expensive and photographs are not. - What if it is wrong for me? Put the returns policy near the buy button in one sentence. A link to a policy page is not an answer. - When does it arrive? A delivery estimate on the product page beats a shipping page every time. ### Cart and checkout: take things away - Guest checkout, always. Offer the account after payment if at all. - Every form field is a toll booth. Autofill-friendly inputs, one address block. - Show the full cost, delivery included, before the final step. Surprise charges at the end are the most reliable way to lose a sale anywhere in the world. - For Indian customers, put UPI and cash on delivery where people can see them. Card-first checkouts leak badly here. ### Speed is a conversion feature before it is an SEO one Someone shopping on 4G abandons a slow store before your design gets a chance to work on them. Compressed images, few scripts, product pages under two seconds. On the replatform projects I have worked on, the speed and checkout work is usually what moves the numbers, and it happens before anyone opens a design tool. ### Then test, in small boring increments One change, one metric, a fortnight. Button wording. Where the delivery promise sits. Whether reviews appear above or below the fold. Small tests build on each other. A full redesign resets everything you have learned and leaves you guessing again, which is why the stores that grow steadily tend to be the ones that look much the same as they did last year. --- # Case studies ## FinnUp Type: Fintech Platform Website | Industry: SME Lending / Fintech | Location: Mumbai & Chennai | Live site: finnup.in | URL: https://ankitkedia.com/work/finnup Marketing site for an SME debt marketplace, built to serve borrowers, lenders and channel partners without any of them getting in the way of the others. ### The challenge FinnUp connects Indian MSMEs looking for capital with a large panel of lenders. A single marketing site had to work for three audiences whose questions barely overlap. A borrower wants to know how quickly money arrives. A lender is assessing whether the deal flow is worth the integration effort. A channel partner is working out what they earn. Written as one general page, that site says nothing useful to anybody. ### What was built I gave each audience its own route through the site, with its own landing page, its own argument and its own signup path into the web application, then made the navigation honest about which route you were on. Around that sits a product section covering eight lending products from working capital through to trade finance, a blog and media centre for organic reach, and trust material that does the work no copy can do on its own, meaning lender logos, named partner testimonials and real integrations. Demo booking runs through Calendly, enquiries can start on WhatsApp, and the mobile app is promoted with QR codes and Play Store links for people who arrive on a phone. ### Outcomes - Separate journeys for borrowers, lenders and partners - Eight lending products presented in plain language - Blog and media centre for organic reach - Calendly demo booking and WhatsApp enquiry paths --- ## Affillo Type: Affiliate Marketing SaaS Platform | Industry: Affiliate Marketing / SaaS | Location: Essex, United Kingdom | Live site: affillo.com | URL: https://ankitkedia.com/work/affillo An affiliate marketing and performance tracking platform covering merchants, publishers and agencies, with attribution, fraud control and automated billing. ### The challenge An affiliate platform has to keep three groups happy whose interests only partly align. Merchants do not want to pay for traffic that never converted. Publishers want to see their clicks tracked and their commission paid on time. Agencies need to work across many merchant accounts without ever seeing data they should not. Underneath all of it sits real money moving between parties, in the form of deposits, commission balances, payout thresholds and recurring subscriptions, where a missed webhook is not a cosmetic bug. ### What was built The platform is built as four distinct experiences rather than one application with permission checks bolted on afterwards. Merchants, publishers, agencies and administrators each see only the accounts and records their role allows, and agency access to a merchant is granted, reviewed and revoked individually. Programme management covers commission models, referral periods, creatives, coupons and publisher settings. Attribution is deliberately multi-touch, because no single method survives contact with the real internet. Click tracking carries SubIDs and session logging, server-to-server postbacks handle the cases where pixels fail, and coupon attribution catches the sales where no click was ever recorded. Fraud checks run before money moves rather than after, flagging duplicate order IDs, blocked coupon usage and transaction patterns worth a human look. The finance layer handles merchant deposits, publisher balances and payout thresholds, with subscription billing wired to GoCardless direct debit. That includes webhook handling and reconciliation, so a billing event arriving late, twice or out of order still leaves the ledger correct. I also built the public site, with separate signup funnels for each account type and tiered pricing. ### Outcomes - Four role-based experiences: merchant, publisher, agency, admin - Multi-touch attribution across click, SubID, S2S and coupon - Fraud checks applied before payout rather than after - GoCardless direct debit with webhook reconciliation - Separate signup funnels for each account type --- ## TBT Group Type: Corporate Website + SEO & AEO | Industry: Office Interiors / B2B | Location: Kolkata | Live site: tbtgroup.in | URL: https://ankitkedia.com/work/tbtgroup Corporate site for a long-established premium interiors firm, structured so that both search engines and AI assistants can find and quote what it does. ### The challenge Tirupathi Balaji Traders has spent more than two decades in premium workplace interiors, holds exclusive dealerships including MillerKnoll for Telangana and Andhra Pradesh, and has delivered for clients of the size of Amazon, HSBC and Infosys. Very little of that authority was reaching anyone who had not already been told about it. The old site did not make the credentials legible to search engines, and it certainly did not make them quotable by the AI assistants their buyers were starting to ask. ### What was built The site was rebuilt around how buyers actually shop, with solution categories for acoustics, data centres, commercial spaces and experience centres, and a dedicated page for each partner brand the company carries. That includes MillerKnoll, Knauf, Unitile, Dormakaba, Muuto, Roff and UniAcoustic. A client wall carries the names that close meetings. For conventional search, every page got targeted metadata, Open Graph tags and clean semantic markup. For answer engines I added conversational FAQ blocks that state, in plain sentences, what the company provides, which brands it is authorised to supply and which territories it covers. Those are the exact questions a facilities manager types into ChatGPT, and they now have a source to draw from. ### Outcomes - Dedicated pages for seven premium partner brands - FAQ blocks written to be quotable by AI assistants - Clean semantic structure with complete metadata - Blog platform for ongoing content --- ## Alexis Pvt. Ltd. Type: Corporate Website + SEO & AEO | Industry: Data Centers / Infrastructure | Location: Kolkata | Live site: alexis.net.in | URL: https://ankitkedia.com/work/alexis Corporate site for an infrastructure firm working across data centres, civil and electrical projects, with real project galleries as evidence. ### The challenge Alexis delivers data centre infrastructure, civil and electrical turnkey work and institutional projects across India. It is a credible business in a niche where buyers are cautious and procurement is slow, and it needed a digital presence that matched the scale of the work rather than undercutting it. Being findable for a small number of high-intent searches matters far more here than broad traffic. ### What was built I structured the site around its six service verticals, covering data centre infrastructure, civil construction, HT and LT electrical works, institutional projects, manpower outsourcing and telecom support. Each one gets a page with enough detail to survive a procurement conversation. Photographs from completed projects do the persuading, because in this sector nothing convinces like evidence of execution. Authorised partner sections set out the distributor relationships with Everest Industries, OBO Bettermann, Hindalco Totalis, Fosroc, Birla Opus and VOX India. Metadata, Open Graph and locale tags cover conventional search, while service-level anchors and plainly worded capability statements give answer engines something they can cite. ### Outcomes - Six service verticals given proper depth - Completed-project galleries as evidence of execution - Six authorised partner brand sections - Page structure written for both search and AI answers --- ## Prejag Holidays Type: Travel Agency Platform | Industry: Travel & Tourism | Location: Surat | Live site: prejagholidays.com | URL: https://ankitkedia.com/work/prejag A travel platform for a Surat agency running since 2008, carrying destinations, packages and hotels without losing the human, call-us-directly character. ### The challenge Prejag Holidays has planned trips out of Surat since 2008, mostly on referrals from families who travelled with them before. They needed a site able to hold hundreds of destinations, packages and hotels, which normally pushes a business towards looking like an online travel agency. That would have thrown away the thing they actually compete on, which is that you can ring them and speak to the person handling your trip. ### What was built The catalogue is organised the way people plan, with destination guides grouped by region, package pages carrying day-by-day detail, hotel listings at agency rates, and separate sections for visas, weddings and corporate travel. Every image runs through a resize and WebP pipeline, which matters more than it sounds when a large share of visitors arrive on mid-range phones over mobile data. Enquiry is kept deliberately human. The planning form asks for budget as a band rather than a number, WhatsApp links open with the message already written, and each department publishes its own phone number instead of hiding behind a contact form. FAQ sections answer the questions travellers actually ask, in the language they ask them, which serves readers and answer engines equally well. ### Outcomes - Destination, package and hotel catalogues at scale - WebP image pipeline built for mobile data - WhatsApp-first enquiry with department-level contacts - Traveller FAQs written in plain language --- ## WTiN Type: B2B Intelligence Platform | Industry: Textile Intelligence / Media | Location: UK (remote) | Live site: wtin.com | URL: https://ankitkedia.com/work/wtin Front-end development on a subscription intelligence platform serving the global textile industry. ### The challenge WTiN publishes innovation intelligence for the textile industry, spanning articles, reports, patent databases, podcasts, webinars and virtual events, with more than a century of publishing behind it. Content at that volume creates a specific problem. Readers have to be able to find the one thing they need, subscribers have to hit the paywall in the right place, and none of it can feel like wading through an archive. ### What was built I worked on front-end delivery for the content-heavy areas of the platform. That covered article and insight layouts, the content hubs organised around fibres, digital textiles, smart textiles and patents, event and webinar promotion modules, and membership touchpoints. The work was mostly about making a large editorial catalogue behave: filtering that stays usable as result counts grow, subscription gating placed so it converts rather than irritates, and reading experiences that hold up on a phone as well as on a desk. ### Outcomes - Content hubs across several data products - Filtering and subscription gating for a large catalogue - Event and webinar promotion modules - Editorial layouts built for daily publishing --- ## Diet Shaala Type: Coaching Website + Funnels | Industry: Nutrition & Wellness | Location: Bengaluru | Live site: dietshaala.com | URL: https://ankitkedia.com/work/dietshaala Coaching brand site with per-programme landing pages, WhatsApp-led enquiry and content built to capture high-intent health searches. ### The challenge Diet Shaala runs seasonal cohort programmes, each with its own dates, pricing and audience, and most visitors arrive from Instagram or a WhatsApp forward rather than from search. The site had to launch a new offer quickly, convert people who came in warm from social, and still do useful organic work in the background between launches. ### What was built The homepage carries a rotating hero for whichever programmes are live, and each programme has its own landing page with dates, pricing and what is actually included. Calls to action open WhatsApp with a message already composed, because that is where this audience is comfortable and asking them to fill a form loses people. Success stories appear as cards with real numbers and durations, backed by video testimonials for the ones who agreed to be filmed. Underneath the offers sit condition pages for PCOS, thyroid, diabetes, fatty liver, perimenopause and teenage obesity, which is where the high-intent search traffic lands. A courses area, a recipe library and a blog give returning visitors a reason to come back between programmes, and a client login supports delivery once someone enrols. ### Outcomes - Landing page per programme with its own offer mechanics - WhatsApp-led conversion paths with pre-filled messages - Condition pages targeting high-intent health searches - Courses, recipes and blog for returning visitors --- ## Benekleed Type: AI Product Photography & Video | Industry: Home Textiles / D2C | Location: India | Live site: benekleed.com | URL: https://ankitkedia.com/work/benekleed AI-generated product photography and video for a clothing brand, including model variations produced from a single garment shoot. ### The challenge For a clothing brand, photography is a recurring cost that scales badly. Every colourway, every season and every campaign concept implies another shoot, and showing a garment on more than one type of model multiplies that again. The brand needed lifestyle imagery and video across the catalogue at a volume that conventional shoots could not fund. ### What was built The starting point is the real garment, photographed properly. From there, generative models produce styled scenes and model variations while keeping the garment itself accurate, which is the part that matters and the part cheap AI imagery gets wrong. The comparison on this page shows what that means in practice. The original catalogue shoot is on one side, and on the other the same garment on a different model, with the pose matched across front, three-quarter and back views. No second shoot, no second fitting, and the fabric behaves the way it does in the photographs it came from. Selected stills are then animated into vertical video for reels and paid placements, and everything is graded to the brand look before delivery. ### Outcomes - Model variations produced from one garment shoot - Matched poses across front, three-quarter and back views - Vertical AI video for reels and paid social - Consistent treatment across colourways --- ## DoReMe Type: AI Product Photography & Video | Industry: Consumer Brand / D2C | Location: India | Live site: doreme.in | URL: https://ankitkedia.com/work/doreme AI-enhanced product imagery across a mixed catalogue of bags, apparel and accessories, prepared for marketplace listings and campaigns. ### The challenge DoReMe sells across a wide range of categories, from backpacks and sweatshirts to sunglasses and hair accessories. Each product needs several angles, marketplaces impose their own image rules, and campaign concepts change faster than a shoot schedule can keep up with. Doing all of that conventionally means either a permanent studio arrangement or a catalogue that always looks slightly behind. ### What was built Products are photographed and then enhanced and finished through an AI-assisted pipeline. Backgrounds are cleaned, lighting is normalised across the range and scenes are extended where a listing needs a different crop or context, with the product itself kept accurate throughout. The result is a catalogue where a hair clip photographed in one session sits comfortably beside a backpack photographed in another, which sounds minor until you see a marketplace listing page where it is not true. Campaign variants can be produced from existing shots rather than booking new ones, so testing a second creative direction costs a day rather than a fortnight. ### Outcomes - Consistent finish across a mixed product range - Multiple angles prepared per product - Campaign variants produced without new shoots - Assets sized for marketplaces, store and social ---