AEO / GEO and AI marketing

Your buyers now ask an AI which brand to trust. But do you show up?

Most don't. I measure that, then build the content that changes the answer. Ask an answer engine below.

Sourish Singh Roy at the National University of Singapore
What is AEO/GEO? A demonstration of answer-engine optimisation
Who is Sourish Singh Roy?
You

Sourish Singh Roy is a Master of Communication graduate from the National University of Singapore, specialising in Strategic Communication and Data Analytics. He works where marketing meets applied AI: measuring how brands appear inside AI answer engines, then building the content that changes the answer.

SR
Sourish Singh Roy
AEO/GEO and AI marketing · sourishsinghroy.com
Cited source

That is answer engine optimisation. Most brands have no idea what the AI says about them. I do this for a living.

Run this search. Press send and see what the AI answers.

Work & skills

How do I get this done?

Each one makes its claim, then shows the project and the number that back it.

Choose:
Share of answerIllustrative
CitedAbsent
SegmentationIllustrative
Engagement → ↑ Intent The segment worth reaching
01

AI-search visibility (AEO/GEO)

Answer Engine Optimisation (AEO) and Generative Engine Optimisation (GEO): I measure how often a brand gets cited inside AI answer engines, run after run.

Structured query frameworkJSON-LD schemaCompetitor gap mapOptimisation dashboard

The proof: a 154-query audit for StrongKeep Cybersecurity across ChatGPT, Perplexity and Google AI Overviews, with 16 competitors benchmarked on five signals. The gap map it produced now directs what StrongKeep publishes next.

154
queries audited
3
AI answer engines
16
competitors benchmarked
5
signals in the gap map
The engine, end to end
Idea in Worth it? The plan Three formats Five checks A human posts Did it work? learns
Every check can refuse. The loop learns; the engine never sends.
02

AI content systems

I build content engines that produce fact-checked, on-brand content, verified before it publishes.

Agentic AI workflowsStructured schemaBrand-voice filtersCitation verification
How it works

Now that you know what AEO is: how do you create content that gets you into AI answers?

With the engine, you go from a content idea to ready-to-post assets in a click. How?

An illustrative walkthrough

Research points at a question your buyers ask that nobody answers well. You feed that idea into the engine.

F
A founder@buildinginB2B

Asked ChatGPT which companies like ours it would recommend. We were not in the answer. How do we change that?

It scores the idea against what matters for your business: can you win this answer, does it fit, are the facts to hand? Weak ideas wait.

Can we win this?4/5
Right for the business5/5
Facts to handYes
Worth doing

One researched idea becomes a blog post, a LinkedIn post and a newsletter, each written for its channel, built only on facts it can back up.

BlogHow AI assistants decide which companies to recommend, and how to earn the mention
LinkedInYour buyers now ask an AI who to shortlist. Here is what it checks before it names you.
NewsletterWhat an AI answer engine needs before it trusts you

Every draft is scored before it reaches you: facts true, sounds like you, better than what the AI shows today. Then you post it. The engine never publishes on its own.

  • Every fact traced to a source
  • Sounds like your brand
  • Better than what the AI shows today
  • You give the go-ahead

The engine asks your buyers' questions again on a schedule and measures whether the AI now names you. Those results feed back into the engine, so the next piece starts smarter.

Scan run 1absent
Scan run 2absent
Scan run 3cited
Scan run 4cited

Named in 2 of 4 answers, up from none. Illustrative numbers.

A human always posts Never invent a fact It can say no It learns after every post
The proof
StrongKeep Cybersecurity · 2026

Closed-loop AI content engine

After publishing · scheduled scansIllustrative
0h 24h 48h 72h cited absent published Cited · 48h
Scans repeat the same buyer questions on a schedule and log who gets cited.
The problem

Research surfaced the content gaps, but production could not close them quickly or on-brand at scale.

How I solved it

Built a content engine on structured schema, verified citations, and internal cross-linking, tuned to brand voice, with agentic workflows automating multi-format output through messaging filters before handoff.

Visibility lift within 48 hoursof publishing, with two reusable production systems left running.

Why it matters. Speed with governance: content an AI engine can trust and cite.

Contact

For AI-search visibility or AI content systems work.

Master of Communication, National University of Singapore. Based in India, open to relocation and remote.