AI visibility is whether a business is named when someone asks ChatGPT, Perplexity or Google's AI Overviews a question. Search & Content at Devodin builds it alongside traditional SEO: technical fixes, pages written to be cited and a managed Google Business Profile, reported monthly against terms, citations and inquiries.
Ready to scope a Search & Content project? Send the details for a reply within one business day, then a written scope within one business day of the call.
Get a QuoteA page either answers the question before an AI can cite it, or it never gets cited at all.
Because the clicks a ranking used to earn are being absorbed into the answer itself. Ahrefs found that an AI Overview correlated with a 34.5% lower average clickthrough rate for the top-ranking page, across 300,000 keywords, comparing March 2024 to March 2025.
Proportion matters too: SparkToro and Datos measured Google at roughly 373 times the search volume of ChatGPT in 2024. AI visibility is a smaller channel that sits next to Google rankings, not instead of them, and the work here covers both.
Because almost nobody searches by those acronyms. Buyers search "ai visibility," the term used consistently on this page and in every monthly report, sitting next to the traditional rankings it does not replace. Type "geo" into Google today and it returns a Geo Metro and a phone game before anything about generative engines.
"AEO" returns a stock ticker for American Eagle Outfitters. "AIO SEO" returns a WordPress plugin.
By solving whichever problem is stopping it, in the order that problem is hurting the business, not by running the same fixed checklist for every client. Getting found starts with working out which problem applies, using live keyword and competitor data instead of guesswork, then working down the list by impact.
A landscaping company might have a fast, well-built site nobody can find because a stray line in a configuration file is quietly blocking search engines from crawling it. A veterinary clinic might crawl just fine but never answer the exact questions a pet owner types, "is it safe to give a dog aspirin," "how long does a dental cleaning take." Both look like the same problem from the outside. They get fixed in the opposite order.
Whatever is losing the business the most visitors for the least amount of work. A page that takes six seconds to load on a phone loses a visitor before they read a word on it, and that usually outranks writing a new article as the first move.
A site with no structured data marking up its services stays invisible to the answer boxes a better-marked-up competitor already occupies. Neither fix is glamorous. Both move the needle faster than a new page of writing does. A plumbing company losing ground to its own slow gallery page gets that page fixed before anyone drafts a new blog post, because the gallery page is what is hurting rankings today.
Four things: page speed on a phone connection, measured against Core Web Vitals and Lighthouse; whether every important page can be indexed; whether structured data correctly describes the services and locations served; and whether internal links connect related pages the way a search engine expects.
Checked first, these four catch a large share of what looks like a content problem but is a plumbing problem underneath it. An electrician with good service pages that never show up in search, because one line in a configuration file is quietly blocking every crawler, is not going to fix that by writing more content. Fixing the block is a ten-minute change with more impact than a week of new writing.
Each section is written to stand on its own, with a direct answer in the first sentence or two, because a language model pulling out one passage to answer a question never sees the paragraph above it. A section that opens with "as mentioned above" fails the reader and the model.
A section that opens with "as mentioned above, our process is different" fails a reader who landed there from a search result, and fails the model too. A furniture repair shop that ranks on page one for "how to reupholster a couch" can still lose the buyer who asked an AI assistant the same question, if a competitor's page states the answer in one clean, quotable sentence and this one buries it in the fourth paragraph. Writing for citation is a learnable habit built into how a page is structured, not something bolted on after the writing is done.
It is the listing that shows up with a map, a phone number and hours attached, the one a buyer taps before they ever reach a website. For a local business it is often the highest-traffic result on the page, ahead of every blue link.
Keeping it accurate and complete, categories, service area, hours, the questions and posts it supports, is part of the monthly work, because a profile that drifts out of date loses the map pack position to a competitor whose profile didn't.
Ranking gets a page onto a results list a person still has to click. Citation gets the sentence cited or summarised inside the answer itself, sometimes with no click at all. Both matter, and they are won by overlapping but not identical work, so both get tracked and reported.
The click gap is measurable. Pew Research found that users clicked a traditional result on 8% of visits when an AI summary appeared, against 15% when it did not. A page can rank respectably and still never get cited, if the one passage a model would need to lift out was never written as a clean, self-contained answer. A business watching only its rankings is missing half of where its buyers now look.
The terms ranking, the citations picked up in ChatGPT, Perplexity and Google AI Overviews, and the inquiries that came in that month, named individually rather than folded into a single traffic graph. A number nobody can act on does not appear in the report as though it were a result.
That means the report says which search a call or a form came from, not just that a call or a form arrived. A business that knows "roof repair near me" brought in three inquiries this month can decide what to do with that fact. A traffic chart cannot tell it anything at all.
The early weeks are research: keyword and competitor data pulled from live sources such as Semrush or Ahrefs, not brainstormed on a whiteboard, alongside a technical check of the basics most search problems turn out to be. Fixes get prioritised by impact rather than worked in the order they happened to be found.
From there the work settles into a steady, repeating cycle: technical fixes and new content moving in parallel, the Google Business Profile kept current, and a report landing every month naming what ranked, what got cited and what inquiries came in. Google Search Console and Bing Webmaster Tools are connected at the start, and the report is built from what they show. Search engines and AI answers move on their own schedule, which is why this is ongoing work rather than a project with a finish line.
It fits a business that wants monthly evidence, terms, citations and inquiries, rather than a promise. It is not the right fit for a business wanting a guaranteed ranking or an overnight result: no honest agency can promise a specific position, because neither Google's algorithm nor an AI model's citation behaviour belongs to any agency.
What can be delivered is research, work prioritised by what moves a business forward, and a monthly report that says plainly what moved and what did not.
See how a project runs, from first email to handover and read the answers on timelines, reporting and payment.
The questions below cover how the work runs, how it shows up in search and in AI answers, and what a client needs to provide. Anything else can go to hello@devodin.com or +1 772-209-3586.
from $1,500/mo
Every project starts from these figures, in US dollars, last reviewed October 2026. The exact scope is agreed with you before work begins. See full pricing, or read the answers on payment and ownership.
Tell us what you need. A reply comes within one business day, the written scope with dates follows the call within one business day, and you own every file, login and domain from the day it ships.
