Start with a free Shopify product-data check. Fix and recheck, then explore an optional AI shopping test to see whether your product gets recommended.
Check my product for freeProduct candidate-set diagnostic
Find out why your product is missing from AI shopping shortlists.
For one Shopify product that deserves to be understood correctly. We test it against natural buyer questions, then trace the gap to product positioning, commerce data or outside sources.
Captured evidence, expert diagnosis and a focused fix-and-retest plan.
Ask us to test one product
MC LAB RESEARCH NOTE · 01 SEP 2026
AI shopping is already measurable—not just forecast.
Consumer adoption, Shopify commerce data and controlled agent tests point in the same direction: AI is becoming part of how products are researched, compared and shortlisted.
01 · CONSUMER USE
What shoppers already ask AI to do
NRF × IBM, 2026 Consumer Research Study ↗Global consumer survey · n=18,000 · published 7 Jan 2026
02 · SHOPIFY COMMERCE DATA
AI-referred shoppers arrive closer to purchase
≈+50%conversion for AI-referred product-page sessions vs. organic search
AI search55%
Organic20%
Share starting on a product page>8×AI referral sessions YoY
≈13×AI-referred orders YoY
+14%average order value
Shopify Q1 2026 commerce data ↗Shopify storefronts · product-page conversion comparison · published 11 May 2026Early channel signal; Shopify reports organic search still drives more sessions overall.
03 · CONTROLLED AGENT TESTING
What the agent sees can change what it buys
≈26Kshopping-agent tests
6frontier models
+99pplargest reported shift from one prior source
PRODUCT GRID+PRIOR SOURCE→DIFFERENT CHOICE
Wharton Generative AI Labs, Prompting Science Report 6 ↗Fitness-watch shopping simulator · 200 runs per condition/model · published 27 Aug 2026
WHAT THIS MEANS FOR ONE SHOPIFY PRODUCT
See how we test, diagnose, fix and retest →
AI compresses the journey.
Shopify reports that 55% of tracked AI-referred sessions begin on a product page, versus about 20% for organic search.
The hard problem moves upstream.
Our inference: feeds make products accessible; clear facts, category associations and surrounding sources help determine whether they enter the right shortlist.
How to read these signals
These studies measure different populations and behaviors: reported consumer use, attributed Shopify sessions and controlled agent choices. Read them as complementary evidence that AI is shaping product discovery—not as one combined trend line.
Two products. Two different candidate-set decisions.
Each case begins with a real buyer need and shows where the product leaves the visible shopping decision.
CASE 01 · STORE-REVIEWEDSleep product
The right product. Missing from the shortlist.
A need-led test and material-aware control separated product fit from a weak category connection.
Read case 01 →CASE 02 · INDEPENDENT RESEARCHUSD 99 Snapshot
Readable by name. Missing by need.
The product passed direct retrieval, then remained absent through one four-turn unbranded shopping conversation.
Read anonymous case 02 →
ONE CASE, FULLY EXPLAINED
A product can be the right answer—and still miss the shortlist
We used two captured ChatGPT conversations to separate product fit from product discovery. The result is one clear diagnosis, not a wall of speculative examples.
01SHOPIFY / SLEEPSTORE-REVIEWED CASE
REAL DISCOVERY GAP
An independent organic sleep brand: visible by category, absent by need
The product matched the shopper's constraints. It was not recommended until the second prompt explicitly connected hemp-hull and buckwheat-hull pillows.
0mentions in the neutral discovery test
#1in one captured material-aware control
1category-connection gap isolated
Hot sleeper + natural + foam-free + adjustable + heavy
Five buckwheat products Brand absentSame constraints + hemp-vs-buckwheat comparison
The brand surfaced first Product retrieval verifiedTwo fresh ChatGPT conversations captured on 31 Aug 2026 under recorded conditions.
02HANDMADE / WEDDINGINDEPENDENT PUBLIC-SOURCE RESEARCH
USD 99 CANDIDATE-SET SNAPSHOT
Readable by name. Missing by need.
A handmade product passed the direct-product validity check, then remained absent through four turns of a natural, unbranded shopping conversation. The diagnosis isolated a candidate-set break before comparison—and one focused hypothesis worth testing.
Read anonymous case 02 →DIRECTPassProduct retrievable
T1–T4AbsentUnbranded conversation
PRIORITYFocused fixCategory bridge
Case 02 is independent public-source research, not commissioned by or affiliated with the merchant. Merchant identity is protected.
ABOUT MC LAB
Built where algorithms meet human decisions.
MC Lab is an independent four-person team of NYU and Royal College of Art graduates. Our careers span AI algorithm work, ecommerce practice, mathematics and service design—giving us the technical depth to interrogate AI systems and the commercial judgment to make the answer useful.
A compact team that stays close to every engagement.
Quantitative rigor paired with customer-centered design.
Long-term professional experience in AI algorithms and ecommerce.
BUILT FOR DECISIONS
Structured evidence. Clear next actions.
Our research workflow combines product-specific testing, source comparison and deeper AI analysis. You receive captured evidence, a source-ready artifact and prioritized actions your team can use.
- Product-specific testing
- Relevant AI surfaces
- Source comparison
- Professional output
Independent studio. University names describe team members' educational backgrounds and do not imply institutional affiliation.
02 / YOUR ARTIFACT
Six modules. One decision-ready product diagnosis.
You receive captured evidence, an explanation of the failure point and the smallest useful next action—not a generic GEO checklist or opaque visibility score.
Candidate-set evidence matrix
Natural buyer questions, relevant AI systems, the observed shortlist and the original test conditions in one inspectable view.
Discovery failure classification
We identify whether the SKU was missed, misclassified, shortlisted but rejected, contradicted by its sources or weakened at the product page.
Category bridge map
A clear path from buyer need to product category, deciding attributes, supporting proof and the priority SKU.
Competitor evidence gap
We compare the evidence behind products that entered the answer with the facts or associations your product is missing.
PDP last-mile brief
A focused review of whether the product page can independently explain fit, difference, proof, trade-offs and purchase conditions.
Prioritized change-and-retest plan
Each proposed change names the diagnosed problem, implementation location, controllability and controlled retest condition.
03 / PRICING
Start with one product, one clear scope
Start with a $99 candidate-set snapshot. Continue to the $100 fix-and-retest stage when the evidence identifies a clear, addressable opportunity.
ONE SKU · TWO CLEAR STAGES
$199USD
$99 candidate-set snapshot + $100 fix-and-retest stage · 5 business days of active production
Start with the $99 snapshot- $99 · evidence snapshot delivered in 3 business days
- $100 · Focused Fix Brief delivered in 2 additional business days
- Included live retest once the approved change is live and available to the AI shopping systems being tested
- The snapshot is yours; continue to the next stage when ready
- One walkthrough + one in-scope revision
Implement it yourself
Use the artifact, source-ready brief and retest plan with your own team.
Ask MC Lab to implement
We can scope Shopify content and technical implementation separately after the diagnostic.
Measure the approved change
The corresponding live retest is included once the change is live and available to the AI shopping systems being tested. Additional scheduled or multi-SKU retesting is separately scoped.
04 / FAQ
Everything to know before we begin
Fit, process, access, ownership and payment—clear before either side commits.
Is this service right for my Shopify store?
It is designed for founder-led independent Shopify brands—from solo makers to compact teams—with one priority SKU, a public product page and a real buyer-discovery question. The strongest fit is a differentiated product whose value depends on explanation, comparison or category fit. We confirm fit before you pay.
What do you need from me to begin?
Your Shopify store URL, one priority SKU, your target market and any product claims or evidence that are not already public. A priority buyer or use case is helpful, but we can clarify that with you during scoping.
What is included in the $99 candidate-set snapshot?
You receive a tailored buyer-intent test set, captured AI answers, the observed shortlist, an initial failure classification and prioritized blockers. The snapshot is yours to keep even if you choose not to continue.
What happens during the $100 fix-and-retest stage?
Within two additional business days, we turn the evidenced failure point into a source-ready Focused Fix Brief with a category bridge, product changes, PDP last-mile guidance and a controlled retest plan, plus one walkthrough and one in-scope revision. The corresponding live retest is included once your approved change is live and available to the AI shopping systems being tested. Additional scheduled or multi-SKU retesting is separately scoped.
How do you decide what to change?
We do not ask AI to freely rewrite your product page. We trace an observed shopping failure to a missing, unclear or conflicting product fact, recommend the smallest change you can control, and retest the same buyer question under recorded conditions.
Which AI platforms do you evaluate?
Testing can include ChatGPT, Gemini, Perplexity and other major AI discovery surfaces relevant to your target market. The exact test set is tailored to your product rather than using the same generic prompts for every store.
Do you need access to the Shopify admin?
No access is needed for the diagnostic; the public storefront, product context and priority market are usually enough. If you ask us to implement changes, we request only the minimum access required after that separate scope is approved.
Can I implement the recommendations myself?
Yes. The solution is designed to be usable by you or your existing Shopify team. We can also quote implementation separately or provide ongoing support under a custom subscription.
How do you handle confidential store information?
We use supplied information only to perform the agreed work, request the minimum information needed and do not publish private store data or client deliverables without permission. A separate confidentiality agreement can be arranged when required.
Do you guarantee rankings or recommendations?
No. We deliver captured evidence, a diagnosis of the failure point and a controlled retest. Final retrieval, candidate selection, ranking and presentation remain controlled by each AI platform.
How does payment work?
For the professional diagnostic service on this site, the first $99 covers the candidate-set snapshot and the optional $100 stage covers the Focused Fix + Retest. We normally use PayPal Business Invoice; US clients may choose Zelle, which has no purchase protection. Instructions come only from hello@mclab.party. The automated Self-Check is a separate product: its $9/$19/$39 credit packs are purchased through Paddle.
Who owns the final deliverables?
After the relevant stage is paid, you may use and adapt the delivered diagnostic or solution for your own store. MC Lab retains its pre-existing intellectual property and reusable know-how.