What we do Packages How it works About FAQ Contact Get started
What we do

Three deliverables. One outcome: you appear in AI search.

Answer Intent delivers three deployment-ready assets that close the gap between what your website says and what AI engines can extract: validated JSON-LD schema markup, a structured LLMS.txt file, and AI-extractable content. All three are produced as files and handed to your existing developer with a one-page implementation guide. No CMS access. No logins. No agency owning your stack.

See the packages Start with an audit
Deliverable 01 Technical AEO

JSON-LD schema markup

JSON-LD schema is structured data added to your website's page code. It tells AI engines what type of entity each page represents: a business, a person, a service, a frequently asked question. When schema is absent, AI engines cannot confidently identify the entity. When schema is present and correct, the business becomes a named, citable entity in AI-generated answers. Answer Intent produces validated JSON-LD for every schema type your site requires.

Why it matters for AI citation

ChatGPT, Perplexity, and Google AI Mode all rely on entity recognition to produce confident citations. When an AI engine encounters a business webpage, it attempts to classify the entity: what type of thing is this, what does it do, who runs it, where is it located. Without schema, that classification is probabilistic - inferred from page text that was written for humans, not machines.

With validated JSON-LD schema, the classification is explicit. The Organization, Person, Service, and FAQPage types each map to known entity classes inside the AI's knowledge graph. Correct schema increases the probability that the AI engine assigns the business a stable, nameable identity - which is the prerequisite for citation. A business with no schema is not invisible to AI engines; it is ambiguous to them. Ambiguous entities are not cited.

What's included in the deliverable

The schema deliverable covers every type relevant to your site and business structure. Standard inclusions are: Organization (legal name, address, contact, areas served, known topics), Person (for named founders or practitioners), Service (one block per core offering), FAQPage (for any FAQ content your site carries), BreadcrumbList (for site structure), and HowTo where your content describes a process.

For businesses with a local physical presence, LocalBusiness replaces or extends Organization. Every block is validated against the schema.org specification and tested against Google's Rich Results Test before delivery. The deliverable is a single HTML snippet per page, structured as a @graph with @id cross-references between entities - so the AI engine sees a connected knowledge graph, not isolated fragments.

What it looks like once deployed

The deliverable is an HTML code block containing a single script tag with type "application/ld+json". Your developer pastes it inside the head element of your page template. It renders invisibly to users - there is no change to what visitors see or how the page looks.

AI crawlers and Googlebot read it during their next crawl cycle. The @graph structure means every entity on the page references every other entity by a stable @id, building a machine-readable picture of your business from a single paste.

Who deploys it

Your existing developer pastes the code block into your page template or CMS custom code field. The implementation guide that ships with every engagement specifies exactly where in your HTML structure it goes and which pages receive which schema blocks.

Answer Intent does not hold login credentials, access your live environment, or touch your CMS at any point. The handoff is a file. Deployment is your developer's single action.

Schema markup is included in every tier from Tier 2 upward.

See what's in each tier Next: LLMS.txt
Deliverable 02 Technical AEO

LLMS.txt file

LLMS.txt is a structured plain-text file at your website's root domain (yourdomain.com/llms.txt). It tells AI language models exactly what your business does, who it serves, what your key pages are, and what you want cited. Think of it as a direct communication channel between your website and AI systems, bypassing the crawler's inference layer. Without it, AI engines have to guess what your business does from unstructured page content.

Why it matters

AI language models are trained on enormous corpora of unstructured web content. When they encounter your website, they infer meaning from text written for human readers: marketing copy, navigation labels, and blog posts formatted for engagement, not machine extraction. The inference is imprecise. A well-written homepage can still fail to communicate the specific entity type, geographic scope, named services, or key facts that AI engines use to build citations.

LLMS.txt bypasses this inference layer entirely. It is a structured file read directly by AI crawlers in addition to your main page content. It communicates in structured plain text: what the business is, what it does, who it serves, and which URLs contain the most citable content. Direct communication is more reliable than inference, and more reliable inference means more consistent citation.

What's included

The LLMS.txt deliverable is a single plain-text file structured according to the 2026 LLMS.txt specification. It includes: a title block (business name, legal entity, type of business), a description section (two to three sentences, AI-extractable), a key pages section (each major URL with a one-line summary of what AI engines should cite from that page), a FAQ links section (pointing AI crawlers to structured FAQ content), and a contact and social section for entity disambiguation.

The file is written to pass AI crawler interpretation tests and reviewed against the Answer Intent AEO Framework criteria before delivery. Length is typically 40 to 80 lines - structured for machine readability, not human browsing.

What it looks like deployed

The LLMS.txt file is a plain-text file uploaded to the root of your domain so it is accessible at yourdomain.com/llms.txt. It has no extension beyond .txt and requires no server configuration on most hosting platforms.

Visiting the URL in a browser shows the raw text content. AI crawlers that support the LLMS.txt protocol request the file directly when they index your domain. It does not affect page load, browser experience, Core Web Vitals, or any other site metric.

Who deploys it

Your developer uploads the file to the root directory of your hosting environment - the same location as your robots.txt file. No CMS configuration is required. The file needs to be publicly accessible and return a 200 status code when requested.

The implementation guide confirms the exact upload path for your hosting platform (Netlify, cPanel, Vercel, Render, and others). In most cases this is a single file upload with no configuration steps beyond that.

LLMS.txt is included in Foundation Sprint and higher tiers.

See what's in each tier Next: AEO content
Deliverable 03 Content production

AEO content

AEO content is editorial content rewritten or newly written so AI engines can extract it cleanly. The structural rules are specific: definition layers, answer-first paragraphs, named entities, structured FAQ clusters, and extractable numerical claims. Answer Intent produces pillar pages, FAQ sets, and definition layers as content drafts that your developer or content team deploys into your CMS.

Why it matters

AI engines extract answers from web content in specific ways. They prefer content that opens with a clear definition or direct answer, names the entity being described, includes structured FAQ pairs with explicit questions and complete-sentence answers, and cites numerical data in extractable formats.

Content written for human readers typically does the opposite: it hooks with a question, builds to a conclusion, and uses marketing language that AI engines cannot resolve into citable facts. AEO content rewrites reverse this structure without making the content unreadable to humans. A page that opens with "X is..." followed by a three-sentence definition performs better in AI citation than a page that opens with "Are you looking for..." regardless of word count or domain authority.

What's included

The content deliverable varies by engagement tier. Standard inclusions are: one pillar page per core service or product (1,200 to 1,800 words, definition-first structure, named entity throughout), a FAQ cluster per pillar page (8 to 12 question-and-answer pairs, each answer 40 to 80 words, answer-first, no marketing language), and a definition layer (50 to 80 words) for each service or product page that will carry schema.

Where you have proprietary data - client count, years in operation, pricing anchors, geographic scope - the content incorporates extractable stat blocks formatted for AI citation. All content is delivered as markdown drafts reviewed by Cameron Duncan before handoff. Readability for human visitors is maintained throughout.

What it looks like deployed

Content is delivered as markdown (.md) files. Your developer or content team pastes the content into your CMS's rich-text editor or markdown field. No special formatting or plugin is required beyond standard CMS editing access.

The heading structure (H1, H2, H3) in the draft maps directly to your CMS heading levels. FAQ clusters are formatted as individual question-and-answer blocks that your developer can paste into a dedicated FAQ component or accordion. The implementation guide covers the specific paste workflow for your CMS platform.

Who deploys it

Your content team or developer pastes the markdown into your CMS. No technical deployment is required beyond basic CMS editing access. If your team is not comfortable with markdown, the implementation guide includes a formatted plain-text version.

Answer Intent does not require CMS login access to produce the content, and the handoff is asynchronous - you receive the drafts as files and deploy on your own schedule.

AEO content production begins at Foundation Sprint (Tier 2).

See what's in each tier
Developer questions

What developers ask

5 questions developers and ops leads ask before signing off.

No. Every deliverable is produced as a file outside your stack. Your developer pastes the schema blocks into your page templates and uploads the LLMS.txt to your root domain. Content drafts are delivered as markdown your team copies into the CMS. Answer Intent never holds login credentials or touches your live environment.

Most CMS platforms (Webflow, Framer, WordPress, Shopify, Squarespace) have built-in custom code fields where the schema and LLMS.txt content can be pasted directly by a non-developer. The implementation guide that ships with every engagement covers each platform's specific paste-points.

If you genuinely have no technical resource, the Deployment Check Call add-on (R1,800) covers the entire deployment for you.

AI crawlers re-index at different rates. ChatGPT and Perplexity often pick up structured changes within 7 to 14 days. Google AI Mode follows Googlebot's normal re-crawl cycle (typically 14 to 30 days for established sites).

The 30-day retest built into Foundation Sprint and higher tiers measures the first visible movement after deployment.

No. JSON-LD schema is small (typically under 5KB total per page) and loads inside the head element, not as a render-blocking resource. The LLMS.txt file is text-only and is not loaded by browsers - only by AI crawlers when they request it directly. Page-speed scores are unaffected.

The deliverable format is platform-neutral. Schema blocks are HTML you paste into the head element. LLMS.txt is a plain text file uploaded to your root. Content drafts are markdown. If your platform supports any of those three things (almost all do), the deliverables work.

Edge cases are flagged at the audit stage so we know before scoping begins.

The implementation guide is platform-specific and step-by-step, but mistakes happen. Two safeguards apply. First, the Deployment Check Call add-on (R1,800) reviews your developer's deployment async before measurement starts. Second, the 30-day retest catches deployment issues - if citations have not moved by Day 30, the first diagnostic step is verifying the deployment is live and parseable.

Yes. Every schema block is delivered for review before deployment. You approve, then your developer deploys. There is no schema you have not seen.

See all FAQs
Next step

Start with a Visibility Audit.

We run 30 queries across ChatGPT, Perplexity, and Google AI Mode and show you exactly where your business is invisible - and why. R4,900. 5 business days. Delivered to your inbox, not a dashboard.

Request your audit Compare all packages