FOUNDING PILOT

Tell us what youmost want to learn.

PersonaSignal can scope structured shopper-context and AI recommendation tests around a whole market, named competitors, or your own brand and products.

If you came here from a PersonaSignal email, just reply to that email with your research question.

No existing email thread? You can also write to hello@personasignal.net.

RESEARCH DIRECTIONS

You can start with one of three research directions.

These aren't three separate services. They're three ways to frame the same persona-first AI recommendation research. A project can focus on one direction or combine several once the scope is clear.

MARKET / CATEGORY

Market / category recommendation landscape

Without naming a brand, which brands and products does AI usually recommend when different shoppers ask for buying advice?

WHAT CAN BE STUDIED

Which brands and products surface naturally across different shopper contexts
Which brands appear in recommendations repeatedly
Whether different shopper needs produce different recommendation patterns
How the competitive structure of a category shows up in AI recommendations

EXAMPLE QUESTIONS

"Across 100 different shopper contexts, which brands and products does AI typically recommend?"

"Which shopper needs produce clearly different recommendation results?"

Exact test volume is confirmed during scope definition.

COMPETITORS

Competitor recommendation performance

When shoppers have specific needs, which of several named brands or products does AI select—and in which contexts?

WHAT CAN BE STUDIED

How Brand A / B / C perform across different shopper contexts
Which brands replace each other in certain shopper needs
How AI describes and differentiates competitor positioning
Which public sources appear in answers when Search / Grounding is enabled

EXAMPLE QUESTIONS

"In which shopper scenarios are Brand A and Brand B more likely to be recommended?"

"When AI does not recommend Brand A, which alternatives does it usually choose?"

YOUR BRAND

Own brand / product recommendation diagnosis

Once AI knows your brand, does it actually recommend it in realistic purchase contexts?

NATURAL / BLIND

Run the test without naming your brand

We start from shopper needs without mentioning your brand name, then observe whether AI naturally considers and recommends your product.

"Without telling AI about my brand, does it naturally enter recommendations in contexts that fit us?"

BRAND-INFORMED / PRODUCT-FIT

Supply brand / product information

We give AI your public brand or product information, then observe how it describes the product, which shoppers it considers suitable, and whether it recommends it in concrete purchase contexts.

"Even when it knows our product, in which shopper contexts does it still prefer competitors?"

Research is based on observable answers and recommendation behavior. It does not claim to read AI's hidden reasoning.

COMBINE AFTER SCOPE CONFIRMATION

MARKET+OWN BRAND/OWN BRAND+COMPETITORS/ALL THREE

A project can focus on one direction or combine several. Not every project includes all three.

WHAT TO SEND

Here's what to send when you reply.

No polished brief is needed. Send what you already know, and say which question matters most to you.

Public website or product information plus a clear research question is usually enough for the first discussion.

What's useful depends on the research direction. Anything you're not sure about can wait until we confirm the scope.

Brand / company name

For own-brand research, tell us the brand name. For competitor or market research, a company or project name is enough for context.

Brand website

e.g. https://example.com

Competitor websites can be sent later together with competitor details.

Main products / category

e.g. facial sunscreen / anti-aging cream / travel backpacks / outdoor shoes / pet care products

Market research can state the category only; product research should name the key products.

Key product links

One to three product page links you want studied most. If you are researching a market or category rather than specific products, you can skip this.

Primary market

United States / Germany / European Union / Southeast Asia

This means the market you want us to study—not where the brand is headquartered or registered.

What should this project focus on?

Market / category landscape
Named competitors
Your own brand / product
A combination
Not sure yet—help us frame it

Competitors to watch

If you have clear competitors, include brand names, websites, and product links. If not, that is fine. Structured tests can surface naturally appearing alternatives, but we cannot guarantee a specific competitor will show up.

MODEL & GROUNDING

Model and search choices can go in the email too.

Not sure? We confirm these with you against the research goal. No technical choices are needed up front.

AI MODELS

Which AI should we test?

If you already know, tell us. If not, the model scope can be decided after the research goal is confirmed.

ChatGPT / GPT
Gemini
GPT + Gemini
Not sure—let PersonaSignal recommend

SEARCH / GROUNDING

Should the test include Search / Grounding?

Search / Grounding means the AI is allowed to consult public web information while answering. PersonaSignal can record public sources that appear or are cited—websites, retailers, media—and compare results with and without search enabled.

Yes, include it
No
Not sure—let PersonaSignal recommend

A source appearing or being cited in an answer does not mean it caused the recommendation.

YOUR MAIN QUESTION

What question should this research answer?

This is the most important part. No technical language is needed. Just tell us what you actually want to understand.

The examples below help you phrase a request. They are not a promise of results.

MARKET

"Across 100 different shopper contexts, which brands and products does AI usually recommend?"

"Under which shopper needs do clearly different brands get recommended?"

COMPETITORS

"In which shopper scenarios are Brand A and Brand B more likely to be recommended?"

"When Brand A is not recommended for certain needs, which brands become the alternatives?"

YOUR BRAND · NATURAL

"Without telling AI about my brand, does it enter recommendations naturally?"

"In which shopper contexts is my brand more or less likely to be naturally recommended?"

YOUR BRAND · INFORMED

"Once my brand and product information is provided, how does AI describe and understand the product?"

"Which shoppers does AI consider a fit for the product?"

"Even when it knows the product, in which scenarios does it still prefer competitors?"

SOURCES / GROUNDING

"With search enabled, does AI cite our website?"

"Which retailers, media, or other public sources appear in the answers?"

HOW TO START

If you came here from our email, just reply to it.

That keeps the existing context and means you do not need to repeat anything. Send what is described above. No polished brief is required.

THE FOUR THINGS THAT MATTER MOST

01What you want to research
02Which market
03Which brands / products
04The question you most want answered

Everything else can be added during scope confirmation.

We keep research requests email-first for now. Send your question from your own inbox, and we'll continue from there.

WHAT HAPPENS NEXT

What happens after you get in touch?

A human process: we confirm your question, scope, and deliverables, then run structured research. It is not automated onboarding.

Scope and pricing for the current Founding Pilot are confirmed during scope definition.

01
RESEARCH REQUEST
Send your question
Reply to the original outreach email, or write to PersonaSignal directly.
02
MANUAL REVIEW
We review it
A person checks whether the question fits the current research approach.
03
SCOPE CONFIRMATION
We confirm the scope
Direction, market, personas / scenarios, AI models, test volume, Search / Grounding, and deliverables.
04
PILOT
The pilot begins
Once scope and pricing are confirmed in writing, the project enters the pilot.
05
RESEARCH
The research runs
Structured tests are executed and reviewed by a person.
06
WRITTEN DELIVERY
Written delivery
You receive a written diagnostic and the agreed raw results.