GEO and AEO Consulting in Nepal
Your customers now ask ChatGPT or Google's AI Overview which trekking company to book or which hospital in Kathmandu does knee replacement, and the answer names three businesses. If yours isn't one of them, nobody clicks through to find out. This is the work that makes you citable.
Animated answer engine: four source documents are retrieved one by one into a single generated answer, each line of the answer written as its source arrives and marked with a citation chip, showing GEO and AEO as becoming the source an AI answer cites.
I'm Shashikant Shah, a digital marketing expert in Kathmandu and a software engineer. GEO and AEO consulting is the part of my work about being cited by AI answers: ChatGPT, Gemini, Perplexity and Google's AI Overviews. It's for a business in Nepal, or one abroad, whose buyers ask an assistant before they search. It sits beside my technical SEO work, and the background is on my GEO expert and AEO expert pages.
Short answer: GEO and AEO consulting in Nepal means checking that AI crawlers can fetch your pages, making it unambiguous who you are and what you do, putting a clear answer at the top of each page that matters, adding schema and an llms.txt, and citing your sources. Then measuring with a fixed set of prompts. Nobody can promise a citation, and I don't.
When you need this
The pattern I see in Search Console now is impressions holding steady and clicks falling, because Google answers the question above the results. On ChatGPT and Perplexity the buyer never sees a results page at all. They ask, they get three names and a summary, and they contact one.
Most Nepali sites weren't written to be quoted. The home page says "welcome to our website", the service pages describe the business in general, and the one fact an assistant needs, what you do, where, for whom, at what price, is nowhere on the page in plain words. So the assistant quotes a directory or a competitor's blog instead.
There's a technical layer too. Some sites block the AI crawlers in robots.txt without knowing it. Some render every word in JavaScript, so a crawler that doesn't run scripts sees an empty page. Some have three different names for the business across Google, Facebook and the site, so no system is sure they're the same entity.
The problems I see most often:
- GPTBot or PerplexityBot blocked, or a firewall that returns a challenge page to anything that isn't a browser.
- Pages with no direct answer in the first paragraph, so the model has nothing it can lift as a quote.
- Business name, address and services that differ between the site, the Google Business Profile, LinkedIn and Facebook.
- No schema, or schema copied from a template with someone else's details still in it.
- No way to measure any of it, so the argument about whether GEO matters goes round in circles.
What I actually do
- Access audit: I fetch your pages the way the crawlers do and check the returned HTML, status codes, robots rules, canonicals and rendering, then read your server or Cloudflare logs for which AI crawlers actually visit. You get an evidence backed list of what's blocking you, fixed by me where I have access.
- Entity work: one clear page for who you are, what you do and where, with business, Person and Service schema that matches your Google Business Profile, LinkedIn and every other profile word for word. You get a business AI systems can identify without guessing.
- Answer first pages: for the questions that matter commercially, a plain answer in the first paragraph, the supporting detail below, the author named, the sources cited and the date of the last review visible. You get pages a model can quote in one sentence.
- FAQ and schema: real questions with real answers, marked up so Google and the assistants can read them, plus an llms.txt that tells a model what your site is and where the important pages are. You get the structured version of what your pages already say.
- Sources and mentions: the places assistants read from, industry directories, local news, your own case studies and the profiles that describe you, checked and corrected where the facts about you are wrong or missing. This connects to the link building work. You get consistent facts about you wherever a model looks.
- Measurement: a fixed set of prompts run on a schedule across ChatGPT, Gemini, Perplexity and Google, referral traffic from those assistants in GA4, and crawler visits from the logs. You get a dated baseline and a way to repeat the test without me. The technical half overlaps with technical SEO, and for software companies selling abroad the technology companies page has the wider picture.
How it works
First call and audit. The first thirty minutes are free. Send me your domain, your market and three questions a customer would ask an assistant before choosing you. On the call I'll say whether GEO is your problem or whether plain technical or local SEO is more urgent, because it often is. The audit then gives you a written 30 day plan: verified defects, opportunities, unknowns, and what's not included.
The work. The following weeks go into access fixes, the entity pages and schema, then the priority pages, rewritten with you so the facts are yours. I make the site changes myself where the stack allows, or your developer gets requirements they can check against. It happens in your own hosting, Search Console, repository and accounts.
Reporting, and what you keep. Every month you get a plain language report: the prompt set results against the baseline, AI referrals in GA4, crawler activity, what changed and what stayed uncertain. The prompt set, the schema, the llms.txt and the instructions for repeating the test stay with you, so your team can run the check without me.
Proof from GEO and AEO consulting work
Kyte Research and AirLyft, two technology companies outside Nepal, are on my case studies page with AEO and GEO in the scope I delivered, alongside their SEO and web development. I haven't published a citation result with numbers for either, and I won't invent one, because a prompt sample is directional and I'd rather say that than dress it up.
The other proof is this site. It ships an llms.txt, Person and FAQ schema on the pages that matter, a short answer near the top of every service page, a named author and a review date, and it passes 15 automated checks on every deploy. I built it, and the page you're reading is the method applied to myself.
What GEO and AEO consulting costs
My prices are published. An audit of your AI search eligibility, your entity and your priority pages with a 30 day action plan is USD 249, about NPR 34,000. A four week sprint is USD 899, about NPR 122,000. Ongoing work is USD 1,800 a month, about NPR 245,000. What each one leaves out is on the pricing page. The first thirty minute call is free. If the audit shows that ordinary technical or local SEO is the real problem, I'll say so and point you there instead of selling you GEO.
Questions about GEO and AEO consulting
What's the difference between GEO, AEO and SEO?
SEO is ranking in a list of results. AEO is being the answer Google shows above that list, in a featured snippet or an AI Overview. GEO is being cited inside a generated answer from ChatGPT, Gemini or Perplexity. The work overlaps a lot, and the longer version is in GEO versus SEO.
How do you measure GEO if there's no rank tracker?
Search Console has a report for AI Overviews and AI Mode impressions, but only on some properties, and it doesn't show prompts, citations or positions. For the assistants, I run a fixed set of prompts on a schedule, track referral traffic from them in GA4 and read crawler visits in the logs. It's directional, and I report it as directional.
How long does it take?
Access and entity fixes take a few weeks and show as soon as the crawlers come back. Being cited for a commercial question takes longer, usually three to six months, because the models need to see the same facts about you in several places. I give you review points, not a promise.
What does it cost, and do I need the retainer?
The audit is USD 249, about NPR 34,000. For many sites the audit plus one sprint at USD 899 (about NPR 122,000) covers the access fixes, the entity pages and the priority pages, and the monthly measurement is yours to run. Ongoing work at USD 1,800 a month (about NPR 245,000 a month) makes sense when you have many pages or a market abroad. The details are on the pricing page.
Do you work with businesses outside Nepal?
Yes. Kyte Research and AirLyft are outside Nepal, and the work is remote in any case. For a Nepali business selling abroad, the country the buyer asks from matters more than where you sit, and that goes into the prompt set.
When are you not the right person?
If you want fake review markup, mass generated pages or claims you can't support, no. That risks a manual action and it's dishonest. If your site is barely indexed or your Business Profile is broken, fix those first, and I'll tell you so on the call. If you need a PR programme to get mentioned in the press, that's a separate scope.
Pages related to GEO and AEO consulting
All services · GEO expert in Nepal · AEO expert in Nepal · Technical SEO · Technology companies · Pricing.
Tell me about your site
Send me your domain and three questions a customer would ask an assistant before choosing you. I'll tell you the first thing I'd fix, whether or not you hire me.