Four ways to buy AI visibility. Why we built a fifth.

Most AI SEO, GEO and AEO offers fall into four approaches. Each gets something right. Here is what each one leaves out, and how JefurryAxis is built differently.

JefurryAxis is the AI visibility platform of Jeffery Asia. We describe approaches, not companies.

The four common approaches

Typical of each approach, not of any one vendor. Many providers mix two or more.

Approach 1

SEO agencies adding “AI SEO”

Gets right
Technical foundations: crawlable pages, structured data, clear facts, authority. AI cannot use what it cannot read.
Leaves out
The answer is built for a person, not a keyword. Work is still planned around queries and measured in rankings and traffic.
Approach 2

AI visibility monitoring dashboards

Gets right
Measurement at scale: prompts tracked across models, share of voice, citations and competitor comparison.
Leaves out
Prompts are often generic or name the brand, and the dashboard stops at the report — deciding and doing the work is left to you.
Approach 3

Persona-based prompt trackers

Gets right
AI answers change with who is asking, so prompts are run as simulated buyer personas. A real step forward from generic prompts.
Leaves out
The person without the situation. The same buyer asks differently under a deadline or a budget cut, and that situation keeps moving. Coverage of Chinese models and closed platforms is usually thin, and the fixes come back as recommendations.
Approach 4

Content and PR “GEO packages”

Gets right
AI cites what exists, so publishing volume across trusted sources matters.
Leaves out
Volume without aim. No baseline, one package for every AI environment, and no way to show which piece moved which answer.

What JefurryAxis does differently

01

Person and situation, not keywords

Built on B = f(P, E): who is asking and what they are going through. Personas hold still; situations are watched and updated as policy, prices and sentiment move.

How E is tracked
02

A peer-reviewed method

The approach is written up in Persona-Driven AI Engine Optimization in Zero-Click Search, accepted for an oral presentation at ANZMAC 2026 following peer review.

The study behind it
03

Both internets, and the closed platforms

International and Chinese models scored as separate tracks, never averaged, plus RedNote Ask and Tencent Yuanbao, where only in-platform content counts.

The RedNote case
04

We do the work, not just report it

Jeffery Asia’s team writes and places the content — persona-led articles, PR and RedNote posts through a creator network — with a live link for every item.

The delivery team
05

Content graded before it ships

HCI checks every piece for persona fit, scenario fit, solution maturity and citability before it is published, so fixes happen before the budget is spent.

The four layers
06

The same questions, every day

A frozen, brand-free question set and fixed scoring, re-run daily and reported monthly. When a month does not move, the report says so, and says why.

How the score works

Why optimise for a persona, not a prompt

Most AI SEO and GEO work picks a list of search intents or a few dozen prompts and writes until the brand appears under them. JefurryAxis optimises for the people behind the prompts: a buyer persona in a real decision situation. On top of the usual GEO groundwork, that aims every piece of content at the group you actually want to reach.

Optimising for a persona

  • Aimed at your buyers. Content carries the persona’s goal, constraints and decision stage, so AI surfaces it when that kind of person asks, in whatever words they use.
  • One piece reaches a whole group. A persona stands for many people and many phrasings, so one article answers questions no prompt list would contain.
  • Less budget spent on the wrong people. Money goes into content for buyers who convert, not into winning prompts that bring in people who were never going to buy.
  • Tested, not assumed. In the SMU study, persona-targeted articles lifted the AI Visibility Score from 69.1 to 77.2 in seven days, and were cited only under the target persona’s questions.

Optimising for intents and prompts

  • Guesswork on wording. Real buyers phrase one need hundreds of ways. A fixed prompt list covers a sliver of them, and the wins fade when the model or the phrasing shifts.
  • Thin, repetitive content. Writing to match prompts drifts towards near-identical pages stuffed with the target phrase, with little a reader or a model can use.
  • It pollutes the answers. Mass-produced prompt bait floods the sources AI reads, lowers answer quality for everyone, and invites platforms to discount it.
  • Reach without aim. A prompt says nothing about who asked, so visibility is bought from people outside your market.

The research: Persona-Driven AI Engine Optimization in Zero-Click Search, by Dr. Yun Jia, accepted for an oral presentation at ANZMAC 2026 after peer review. Read the SMU case.

Side by side

Typical capabilitySEO + “AI SEO”Monitoring dashboardPersona prompt trackerContent packageJefurryAxis
Brand-free questions–Sometimes✓–✓
Buyer personas–Sometimes✓–✓ from your own material
Situations that change over time––Rarely–✓ tracked by Insights
Chinese AI modelsRarelySometimesRarelySometimes✓ separate track
Closed platforms (RedNote Ask, Yuanbao)–––Rarely✓
Content written and published for youSometimes––✓✓ per persona
Content scored before publishing––––✓ HCI
Frozen baseline and daily re-test–✓✓–✓
Peer-reviewed method––––✓ ANZMAC 2026

“Sometimes” and “Rarely” describe the approach in general. Ask any provider, including us, to show each capability on your own brand.

Who we are, and are not, a good fit for

A good fit

  • Brands whose buyers ask AI in both English and Chinese
  • High-consideration decisions: education, finance, healthcare, government, cross-border services, travel
  • Teams that want the work done, not another dashboard
  • Anyone who needs to show a board what changed, measured the same way twice

Probably not a fit

  • You only want a self-serve prompt tracker
  • You want a guaranteed first-place recommendation — nobody can honestly promise that
  • Your buyers decide on a single keyword search and never ask AI

Questions about how we differ

Is JefurryAxis a tool or an agency?

Both. JefurryAxis is a measurement platform, and Jeffery Asia’s team runs the optimisation and publishing on top of it. Clients get the platform, the method and the people doing the work.

How is this different from persona-based GEO tools?

Persona tools simulate who is asking. JefurryAxis also models the situation they are in and updates it as the outside world changes, measures Chinese models and closed platforms as their own tracks, and writes and places the content rather than handing back a list of fixes.

Do you cover Western AI models as well as Chinese ones?

Yes. ChatGPT, Claude, Gemini, Perplexity and Grok on the international track; DeepSeek, Qwen, Doubao, Kimi and Hunyuan on the Chinese track; Tencent Yuanbao, RedNote Ask and Google AI Overviews through team-assisted runs.

Can we keep our SEO agency?

Yes. SEO makes your site readable to AI; JefurryAxis works on whether AI chooses you for a particular buyer. The two fit together, and our source analysis often gives an SEO team a clearer brief.

Be the AI
Recommendation

Every approach promises visibility. Ask each one to show it on your own brand, measured the same way twice — starting with us.