Answer Engine Optimization

Target the intent. Claim the response.
LEGO · Phiture 2026 · GEO / AEO experience and what we found
cover
What we will cover

How LEGO earns the recommendation.

We ran the study on LEGO’s three apps before this meeting, so everything here is measured, not theoretical.
01
The shift in app discovery
02
What the AI recommends for LEGO today
03
The two surfaces that determine visibility
04
Our approach to Answer Engine Optimization
05
Measurement and what a pilot looks like
02
01 · the shift

Discovery is moving from search to the ask.

How it is changing
Before: “building instructions”
Now: “what can my 6 year old build with the bricks we already have?”
Parents ask ChatGPT and Gemini outside the store, and now, through Ask Play, inside it too. The job is no longer only to rank. It is to be the app the AI names.
Volumes are still small, so this is not yet a big traffic driver. But the surfaces and consumer adoption are changing fast, and the advantage goes to whoever is the named answer before the volume arrives.
The scale of the shift
20%
of informational queries now happen on chatbots1
34%
of Gen-Z use chatbots for research2
5.5×
faster search growth than Google3
+31%
conversion rate vs non-branded organic search4
11×
YoY growth in ChatGPT + Gemini app referrals, across some of our other clients
1Digital Applied   2Semrush   3Bond Cap   4Search Engine Land 2026. Referral growth from Phiture client data.
03
02 · what the AI recommends today

Named when asked by name. Absent on the job.

LEGO Builder
instructions and the brick ecosystem
Follow instructions
C100 G100
Kids help
C67 G100
Loose bricks
C33 G0
Catalogue
C0 G0
Identify a part
C0 G0
LEGO Play
creative play for kids
Creative for kids
C0 G0
Digital building
C0 G0
Learning
C0 G0
Open-ended play
C0 G0
Screen-time safe
C0 G0
LEGO Insiders
shop, deals and rewards
Shop sets
C100 G33
Deals
C67 G33
New releases
C50 G50
Track collection
C0 G0
Rewards
C0 G17
named 60%+   30–59%   under 30%    C = ChatGPT, G = Gemini Share of Answer; cell colour is the mean of the two. Prompts never contain the word “LEGO”.
LEGO owns the job only where the job is LEGO: following a set’s instructions, and buying sets. Everywhere the job is described generically, the brick ecosystem answers instead. LEGO Play is the sharpest case: zero across all five kids-creative jobs, on both engines.
40%
LEGO Builder
ChatGPT, 40% on Gemini
0%
LEGO Play
ChatGPT, 0% on Gemini
43%
LEGO Insiders
ChatGPT, 27% on Gemini
Three apps, 15 unbranded jobs, two engines.
180 grounded runs: 15 unbranded clusters × 2 phrasings × 3 reps × 2 engines (gpt-5 and gemini-3.5-flash, both with web search), US Android. Prompts describe the job without the brand, so this is a real contest rather than a branded tautology.
04
02 · where the answer goes elsewhere

The listings never claim the jobs LEGO loses.

GPT
Gem
Catalogue
0%
0%
vs Brickit, myBricks, Brickset
Identify a part
0%
0%
vs Brickit, Rebrickable, BrickLink
Loose bricks
33%
0%
vs Rebrickable, Brickit
Track collection
0%
0%
vs Brickset, myBricks, omgbricks
Rewards
0%
17%
vs —
The weakest jobs, and who is named instead. The brick ecosystem, not another toy brand.
Read the listings and the reason is plain. LEGO Builder never says catalogue, inventory, identify, loose bricks, sort or missing piece. Not once.
LEGO Play is starker. It says “creative” 11 times, but “ad-free”, “no ads”, “screen time”, “preschool”, “toddler” and “educational” appear zero times. Those are the words parents actually ask in, and it scores zero on every one of those jobs.
This is the cheapest fix available. The apps already do most of this. The listings, and the pages the engines cite, simply never claim it.
Listing text from the live US Google Play listings for com.lego.legobuildinginstructions and com.lego.common.legoplay, Sep 2026.
05
03 · the two surfaces

AI adds two layers on top of your listing.

Layer 1 · above the store
Win the answer
Be the app ChatGPT and Gemini name, via the store listing, LEGO-owned pages and the sources they cite. Answer Engine Optimization.
Layer 2 · inside the store
Win the store
Hold the Play AI-collection slots and be the app Ask Play recommends, by strengthening the keywords, listings and product signals that feed them.
Measure: Phiture’s rig tracks Share of Answer in ChatGPT and Gemini above the store, and probes Ask Play and the AI collections on a real device inside it.
06
03 · inside the store

The AI shelf is real, and LEGO is not on it.

An Ask Play highlights collection on a category search, filled with two apps.
Every collection is tied to one keyword and filled by that keyword’s organic top-2.
Ask Play, asked for a free creative kids app, names Drawing Games. Not LEGO.
The mechanic is simple and winnable. A collection is tied to one keyword and filled by that keyword’s on-device organic top-2. Rank first or second on the linked keyword and you hold the slot.
Across every keyword we tracked, no LEGO collection has surfaced yet, and asked the kids-creative question directly, the in-store assistant names a rival. The same gap as above the store, on a second surface.
Recommendation: monitor the key queries so we catch the first collection the moment it forms, and defend the organic top-2 that fills it. Cheap to hold now, expensive to win back later.
On-device, live Play client, US Android. Ask Play answers vary between sessions.
07
03 · how Ask Play picks

What decides who Ask Play recommends.

We got the store assistant to explain its own selection. It filters first, then ranks whatever survives. That order matters more than the ranking itself.
First it cuts
Anything rated under 4.0 stars.
Anything that does not look actively maintained.
Anything whose listing does not explicitly support the job.
We have watched it break its own rule elsewhere, ranking a 3.0-star app first. Treat the shape as the signal, not the threshold.
Then it ranks on
User rating, the heaviest single factor.
Download volume, and editorial picks like Editors’ Choice.
How recently you updated.
LEGO clears quality and downloads comfortably. The lever is explicit job support: the listing has to state the job, in the words people ask it in. That is the same fix Layer 1 needs, which is why one listing rewrite pays on both surfaces.
Ask Play asked to explain its own selection, physical Android device, US. Self-reported, not verified internals, and it describes itself differently by category and session.
08
04 · our approach

We speak the language of both ASO and AI.

ASO optimises for the search
Your proven foundation
Metadata: titles, short and long descriptions.
Keyword strategy for high-intent terms.
Listing structure built for conversion.
Rank, conversion and install tracking.
AEO optimises for the ask
The performance edge
Answer formatting: turn long descriptions into clear answers.
Cluster strategy: align content with the jobs people ask.
Cited-page alignment: the off-listing sources the engines quote.
Share of Answer testing: measure what actually moves.
The store listing is the single largest citation source ChatGPT uses for app recommendations, 47.5% of citations. Restructuring it around the job is the one change that lands on ChatGPT, Gemini and Ask Play at once.
AppTweak analysis of 125K+ ChatGPT responses across 9,489 app prompts, US, May 2026.
09
04 · how we win

How we win a category job, step by step.

1 · Prioritize
Pick the unbranded jobs LEGO should own, by fit, demand and value.
2 · Fix the listing
Say the job in the words people ask it in, across title, subtitle, description and screenshots.
3 · Cited pages
LEGO.com first, then the third-party pages the AI already trusts.
4 · Measure
Did Share of Answer move? Next job.
Start with LEGO Play. It is at zero on every kids-creative job while its listing never uses a single word a parent would search with. That is a pure positioning and citation fix, not a product one, and it is the largest gap in the portfolio.
10
05 · measurement

How we measure AEO impact: one KPI funnel.

Read left to right, vanity to sanity. Each stage has its own live view, so one optimisation shows up as movement all the way to installs.
1 · Awareness
Share of Answer
0255075100W1W2W3W4W5W6
Cluster visibility across apps, over time. Illustrative; latest point = real study.
2 · Intent
AI referral clicks
W1W2W3W4W5W6
Clicks from ChatGPT & Gemini through to the store listing.
3 · Action
App installs
04008001200AprMayJunJulAugSep
ChatGPT + Gemini installs, from App Store Connect App Referrer.
Caveat: counts only users shown a direct listing citation who tap it, not those who then search the store by name, so it under-reports.
Illustrative dashboard views; the latest cluster reading is the real study. KPI-funnel framework (Precis × Phiture).
11
05 · what a pilot looks like

A 90-day pilot, on the jobs that matter most.

Two apps, iOS and Google Play, one market, two experiments per app per platform.
Phase 1 · weeks 1–4
Set up & audit
Share of Answer baselined at day zero, across ChatGPT and Gemini.
Identify which pages the engines actually cite.
Start monitoring the collection keywords.
Phase 2 · weeks 5–8
Implement
Rewrite the listing around the job, in the words parents and builders use.
Align the cited off-listing pages.
One listing experiment per app, per platform.
Phase 3 · weeks 9–12
Measure & expand
Report Share of Answer, recommendation rate and accuracy vs baseline.
Tie to AI referral clicks and installs.
Plan the next market and cluster.
Fastest first win: LEGO Play on the kids-creative jobs. Zero on both engines, and a listing that never says ad-free, screen time, preschool or educational. Positioning and citations, not product.
12

Let’s discuss.

close