Each project says what it answers and replaces, how it's built, what arrives, and shows an example. Colours follow the pathway stages: Plan, Make, Sell, Grow and Defend. All examples are illustrative.
Plan · 01
Choose which markets to grow
A clear call on which product fits which market, what to keep, what to adapt, and where the portfolio has room to grow.
Human Intelligence leadsProduct Intelligence supports
What it is
Your portfolio read market by market through what local buyers say and what local products claim. The same product is rated, praised and complained about differently in every country. This shows where the fit is real and where the adaptation is.
Questions it answers
Which of our products travels best, and which never will?
What do buyers in this market expect that buyers elsewhere do not?
Which local brand sets the standard we have to beat?
What has to change on the claim, the format or the message?
Supports and replaces
SupportsPortfolio planning, market entry sequencing, local marketing briefs, regional business reviews
ReplacesLocal market reports commissioned one country at a time, and multi-country U&A studies that take a quarter to field
Market fit: what travels and what must changeIllustrative
Product
Best fit
Fit
What local buyers say
Adapt
Baby wipes, fragrance-free
Vietnam, Thailand
Doesn't leave the skin sticky in the heat
Lead with heat and humidity, not fragrance-free
SPF50 face sunscreen
Korea, Singapore
No white cast, works under makeup
Reword "lightweight" to "no white cast"
Oat milk barista
Australia
Froths like dairy, no split in hot coffee
Keep; add froth claim to pack front
Dog dental chews
Japan
Too big for small breeds
Needs a small-breed format before entry
Each cell opens to the reviews behind it, in the local language, and to the competitor products that already answer the need.
01 · DiagnosePortfolio by market
How each product performs in each market, and where competitors are rated higher.
02 · UnderstandLocal expectations
Needs, claims and expectations compared across countries, read in each local language.
03 · ActMarket picks
A call per product per market, with adaptation notes on claim, format and message.
04 · TestLocal personas
Persona panels tuned to each market to sense-check the adapted version.
You receiveMarket fit matrixLocal expectation map per marketAdaptation notes per productFindings deck and readoutMarket personas
Start with: Deep dive for one portfolio across up to five markets. Ongoing inside the Intelligence programme, and the core of a Multi-market rollout.
Plan · 02
Enter a new category
The adjacent categories where your buyers' needs are not yet met, where incumbents fall short, and what it would take to be chosen there.
Human Intelligence leadsProduct Intelligence supports
What it is
An adjacency map built from 1,000+ categories. It finds the categories whose buyers talk about the same needs your brand already serves, then reads the incumbents there: what buyers praise, what they complain about, and which claims carry weight.
Questions it answers
Which categories share our buyers' needs?
Where are the incumbents weakest, in buyers' own words?
Would our brand be believed here, and on which claim?
ReplacesAdjacency studies commissioned for one category at a time, and desk research built from category reports
Adjacency map for a baby skincare brandIllustrative
Adjacent category
Shared need
Overlap
Incumbent gap, in buyers' words
Entry case
Adult sensitive-skin body wash
No irritation, no fragrance
Says gentle, still stings after shaving
Strong
Baby laundry detergent
Safe on skin, rinses clean
Leaves residue on cloth nappies
Strong
Kids' sunscreen
Gentle, no stinging eyes
Burns their eyes at the pool
Medium
Pet shampoo
Gentle on skin
Incumbents already own "vet-approved"
Weak
Overlap is the share of your buyers' top needs that also lead the adjacent category. Each row carries the incumbents' claims and the reviews that undercut them.
01 · DiagnoseAdjacency map
Categories whose buyers voice the same needs your brand already answers.
02 · UnderstandIncumbent read
Who leads each category, what buyers praise and complain about, which claims hold.
03 · ActEntry case per category
The need to answer, the claims to make, the brands to position against.
04 · TestEntry concepts
Concepts shown to personas from the target category before anything is built.
You receiveAdjacency mapIncumbent scorecardsEntry case per categoryFindings deck and readoutTarget-category personas
Start with: Deep dive, one brand and up to eight adjacent categories. Refreshed on request inside the Intelligence programme.
Plan · 03
Decide who to target
Buyer personas built from real reviews: who chooses you, who you are losing, and the words and triggers that move each one.
Human Intelligence leadsMachine Intelligence supports
What it is
Segments drawn from how buyers actually describe their situation, their need and their reason to choose. Every persona comes with its share of the category, its language, its triggers and the brands it leans to. You can talk to each one.
Questions it answers
Who are our most loyal buyers, and what do they share?
Which buyer type do we keep losing, and to whom?
What words does each persona use, and which of ours do they repeat?
Which persona do AI assistants assume when they recommend?
Supports and replaces
SupportsTargeting and media planning, creative briefs, CRM segments, range and pack decisions
ReplacesSegmentation decks that sit unchanged for years, and persona workshops built on assumption
Three personas in coffee podsIllustrative
The morning rusher34% of reviews · you winOne button, no mess, out the door.
Trigger: speed and no clean-up. Loses on: nothing yet.
The café copier29% of reviews · you loseCrema is thin, tastes nothing like the shop.
Trigger: taste and crema. Loses to: Brand B's "barista" range.
The guilt reducer18% of reviews · openFinally compostable pods that actually work.
Each persona carries its language bank, its switching flows and a chat interface briefed on its reviews. Share figures are of reviews, not of sales.
01 · DiagnoseWho chooses whom
Which buyer types choose you and which choose competitors, by market.
02 · UnderstandPersona needs
What each persona needs, worries about and says, kept current as reviews arrive.
03 · ActAudience profiles
Profiles with the language, triggers and channels that move each one.
04 · TestTalk to them
Brief work and check it directly with each persona.
You receivePersona set with shares and triggersLanguage bank per personaWin / lose mapPersona chat accessFindings deck and readout
Start with: Deep dive, one category and market. Personas stay live and talkable inside the Intelligence programme.
Plan · 04
Decide what to launch next
A ranked list of unmet needs with the evidence behind each one: how strong the signal is, whether it is growing, and who is closest to solving it.
Human Intelligence leadsProduct Intelligence supports
What it is
Buyers say what they wish a product did, in their own words, every day. This reads those wishes across your category and its neighbours, counts them, tracks whether they are rising, and checks whether any product already answers them. What's left is your pipeline.
Questions it answers
What do buyers wish this product did?
Which needs are growing fastest, here and in nearby categories?
Has anyone already launched against this gap?
Which format, ingredient or benefit is gaining momentum?
Supports and replaces
SupportsInnovation pipeline, R&D briefs, range extension, concept prioritisation
ReplacesAnnual U&A studies and early concept screens, and trend reports that arrive after the trend
Ranked unmet needs across three categoriesIllustrative
Unmet need, in buyers' words
Category
Signal
Trend
Answered by
A dental chew small dogs can actually finish
Pet treats
Rising
No one
Deodorant that doesn't stain black shirts
Personal care
Rising
1 brand, poorly rated
Instant noodles with real vegetables, not flakes
Food
Steady
2 brands, Korea only
Nappy that holds overnight at size 5
Baby care
Steady
Category leader, premium tier
Signal is how often the need appears relative to all needs in the category. "Answered by" checks every product page and claim in the category and its neighbours.
01 · DiagnoseGaps
Where buyers complain and no product in the category answers them.
02 · UnderstandMomentum
Rising needs, ingredients, formats and benefits tracked across categories and markets.
03 · ActOpportunity briefs
The need, the evidence, how often it comes up, who is closest to solving it.
04 · TestConcept check
Concepts and pack copy shown to personas to see which they'd pick and why.
You receiveRanked unmet-need listOpportunity briefsMomentum trackerCompetitor answer checkFindings deck and readout
Start with: Deep dive, one category and its neighbours. The momentum tracker runs continuously inside the Intelligence programme.
Make · 05
Decide what to say
Messages and claims built from the words buyers already use: which benefits to lead with, which claims land, and how each buyer type talks about the category.
Human Intelligence leadsMachine Intelligence supportsProduct Intelligence supports
What it is
A message hierarchy with evidence attached. For every benefit in the category it shows how often buyers mention it, which brand owns it in their minds, how buyers phrase it, and whether AI assistants repeat it. The brief writes itself from the top rows.
Questions it answers
Which benefit do buyers mention most when they recommend us?
Which of our claims does nobody repeat back?
How do different buyer types describe the same need?
Which competitor owns the message we want?
Supports and replaces
SupportsBrand and campaign briefs, pack copy, retail content, claims planning
ReplacesMessage testing rounds and one-off social listening reports
Message hierarchy for a laundry detergentIllustrative
Benefit
Buyers mention
Owned by
How buyers say it
AI repeats
Smells fresh for days
Brand A
Still smells clean on day three
Yes
Gentle on sensitive skin
You
No itching on the baby's clothes
Yes
Works in cold water
Nobody
Cold wash and the grass stains still came out
Rarely
Plant-based formulayour lead claim
Brand C
Almost never mentioned unprompted
No
Row three is the open message. Row four is the claim you lead with today. The language bank behind each row holds 50 to 100 real phrasings per persona and market.
01 · DiagnoseClaim check
Which of your claims buyers notice, repeat or ignore, against each competitor.
02 · UnderstandReasons to choose
Why people choose and reject the category, by persona and market, kept current.
03 · ActMessage hierarchy
Benefits in order, with a language bank of real buyer phrases for briefs.
04 · TestLine check
Draft lines run past personas built from real buyers before they go to market.
You receiveMessage hierarchy with evidenceLanguage bank by persona and marketClaim ownership mapAI echo checkFindings deck and readout
Start with: Deep dive, one brand and category. The hierarchy and language bank refresh monthly inside the Intelligence programme.
Make · 06
Back up a claim
Evidence on whether buyers and AI believe your claims: which ones buyers confirm in their own words, how competitors word the same claim, and whether AI repeats it or doubts it.
Every claim on your packs and pages, scored against what buyers say after using the product and what AI assistants say when asked. A claim that buyers confirm is an asset. A claim they contradict is a liability that shows up in ratings and in AI answers.
Questions it answers
Do buyers say our claim is true, in their own words?
How do competitors word the same claim, and which wording do buyers trust?
Does AI repeat our claim, ignore it, or dispute it?
Which claims need stronger proof before they go further?
Supports and replaces
SupportsClaims planning, pack and page copy, regulatory review, launch readiness
ReplacesClaims reviews that rely on the brand's own research alone, and claim tests fielded one claim at a time
Claim credibility across three categoriesIllustrative
Claim on pack
Category
Buyers confirm
Buyers contradict
AI
Visible results in 7 days
Skincare serum
Three weeks before I saw anything
Disputes
No fishy aftertaste
Omega-3 supplement
Rare
Repeats
High protein
Greek yoghurt
Less protein than the plain one
Ignores
Leak-proof lid
Travel mug
Leaked in my bag on day two
Disputes
"Buyers confirm" is the share of reviews that mention the claim's benefit and agree with it. Each row opens to the competitor wordings of the same claim and the reviews behind them.
01 · DiagnoseClaim scorecard
Each claim scored for how often buyers confirm, question or ignore it.
02 · UnderstandCategory wording
How the same claim is worded across the category, and which wording buyers trust.
03 · ActClaims plan
Which to keep, which to reword, which need stronger proof before going further.
04 · TestReworded check
New wordings checked with personas and AI assistants for how they are understood.
You receiveClaim credibility scorecardCompetitor wording mapAI echo check per claimClaims plan: keep, reword, proveFindings deck and readout
Start with: Deep dive, one brand's full claim set. Every new claim is scored as it appears inside the Intelligence programme.
Make · 07
Test before we spend
Feedback on packs, concepts, claims and briefs before money is committed, from personas built on real buyers, with reasons rather than just a score.
Human Intelligence leadsMachine Intelligence supports
What it is
Your brand personas are briefed on thousands of real reviews from their segment. Put a pack, a concept, a claim or a script in front of them and they react the way the reviews say that buyer would, with the reasons in their own language. Run as many rounds as you need, then confirm the winner with real buyers if required.
Questions it answers
Which of these packs would each persona pick, and why?
What would make this concept believable?
Which line confuses people or sounds like a competitor?
What's missing that would tip the decision?
Supports and replaces
SupportsPack design rounds, concept gates, claim selection, creative development
ReplacesEarly concept screens and quick-turn surveys, and the internal debate that fills the gap when neither is budgeted
Three snack-bar pack concepts, three personasIllustrative
Concept
Gym regular
Desk snacker
Lunchbox parent
Call
A · "20g protein" front and centre
That's the number I look for
Feels like a meal replacement, not a snack
Too much for a 7-year-old
Fix
B · "Real nuts, nothing hidden"
Fine, but where's the protein?
This is the one I'd keep in my drawer
Nut-free rule at school, can't
Fix
C · "Made with oats, no added sugar"
Sounds like a kids' bar
Would buy for the afternoon slump
Finally one I can send to school
Go
Each reaction links to the reviews it was drawn from. The call combines persona reach with the strength of the reasons, and says what to change on the two that didn't win.
01 · DiagnoseBaseline
How buyers judge the category today, so new work is tested against something real.
02 · UnderstandReasons
The reasoning behind each persona's reaction, in their own words.
03 · ActGo, fix or drop
A clear call on each option, with the wording or element to change.
04 · TestRounds
As many rounds as you need, then confirm the winner with real buyers if required.
You receiveReaction grid per personaGo / fix / drop call per optionWhat to change, per optionPersona chat access for re-runsShort readout
Start with: Deep dive, one round across up to five options. Unlimited rounds inside the Intelligence programme.
Sell · 08
Fix our product pages
Every page scored against what buyers ask about and what competitors say, with the exact fixes: titles, bullets, claims, image order and Q&A.
A page-by-page read of the gap between what your listing says, what your reviewers praise and complain about, and what AI assistants pick up when they read it. Shoppers and AI both decide from the page. This makes the page say what they are looking for.
Questions it answers
What do buyers ask that our page never answers?
Which competitor page is winning, and what does it say that ours doesn't?
ReplacesManual page audits and one-off content refreshes, and shelf tools that score the page without saying what to write
Page scorecard for an air fryer, one listingIllustrative
Buyers ask or praise
In reviews
On your page
On leader's page
Fix
Fits a whole chicken
Missing
Title
Add capacity in real terms to the title
Quiet enough for an open kitchen
Missing
Missing
Open message: add to bullet 2
Basket is dishwasher-safe
Bullet 6
Bullet 1
Move up to bullet 1
Non-stick coating peels
Unanswered
Q&A
Answer in Q&A with the care guidance
"In reviews" is how often the theme appears across the category. The full scorecard covers every listing, with the rewritten title and bullets ready to paste.
01 · DiagnosePage scorecard
Each page against buyer questions and competitor pages.
02 · UnderstandThe gaps
What the page claims, what reviewers say, and what AI picks up, side by side.
03 · ActPage fixes
Titles, bullets, claims, image order and Q&A answers, per page.
04 · TestPre-publish check
New copy checked with personas and AI assistants before it goes live.
You receiveScorecard per listingRewritten titles and bulletsQ&A answer bankImage order recommendationBefore / after re-check
Start with: Deep dive, up to 20 listings in one market. Every listing, every market, re-scored monthly inside the Intelligence programme.
Sell · 09
Get recommended by AI
How often ChatGPT, Gemini, Claude and the rest name you for the questions buyers really ask, why they pick who they pick, and the exact fix on the page, the pack or the message.
AI assistants build their shortlist from what reviews praise and what product pages prove. This puts 200+ real buyer questions to seven assistants every week, records who gets named and why, and traces each reason back to the review or page it came from. The fix follows from the trace.
Questions it answers
When someone asks for the best option for their need, are we named?
Why does AI pick a competitor over us?
Which of our claims does AI repeat, and which does it get wrong?
Did the answer move after we changed the page?
Supports and replaces
SupportsAI visibility, digital shelf content, claims, brand reputation, launch readiness
ReplacesAd hoc prompt checks done by hand, and visibility scores that come without a reason or a fix
One question, traced and fixedIllustrative
Shopper asks
Best baby monitor that works without wifi?
AI names
Brand ABrand BBrand CYou: not named, 5 of 7 assistants
Why
AI cites "no wifi needed" and "range reaches the garden" from reviews of Brand A. Your reviews say both, 140 times. Your page says "DECT 1.9GHz".
Also
Two assistants describe your battery life as 6 hours. Reviews say 12. The 6-hour figure comes from an old retailer listing.
Fix
Say "works without wifi" and "reaches the garden" in the title and first bullet. Correct the old listing. Re-check in 7 days.
The full report covers every question in your category, with mention rate by assistant, the reasons behind every recommendation, and the fix list in priority order.
01 · DiagnoseMention rate
How often AI recommends you for real buyer questions, against each competitor.
02 · UnderstandThe reasons
What AI says and which reviews, claims and pages it draws on.
03 · ActFix list
Changes across product pages, claims and content that AI reads, in priority order.
04 · TestRe-run
The same questions re-run after changes to see whether the answer moves.
You receiveAI visibility scorecard by assistantReason trace per recommendationAccuracy check on what AI says about youFix list in priority orderWeekly re-run
Start with: the free report, then a Deep dive for one category and market. Weekly tracking inside the Intelligence programme.
Sell · 10
Own the search
The words buyers use in reviews are the words they type into retail search and ask AI. Get them by persona and market, before they show up as volume in the keyword tools.
Human Intelligence leadsMachine Intelligence supports
What it is
Keyword tools report what has already been searched. Reviews show the language before it becomes a search, because a buyer writes "doesn't give my dog gas" months before enough people type it. This turns that language into search terms and AI prompts, grouped by persona and market, and checks whether your listings use them.
Questions it answers
What words do buyers use that we never do?
Which questions do buyers put to AI about this category?
Which terms are rising before they show in search volume?
Which of our listings miss the words that matter?
Supports and replaces
SupportsRetail media keywords, SEO, listing titles, content calendars, AI prompt coverage
ReplacesKeyword tools that only report what has already been searched, and keyword lists built from the brand's own vocabulary
Buyer language versus listing language, dog foodIllustrative
Buyers say
In reviews
Trend
Your listings use it
Persona
doesn't give him gas
Rising
No · you say "digestive health"
Sensitive-stomach owner
picky eater finally eats
Rising
No · you say "highly palatable"
Fussy-dog owner
small kibble for small mouths
Steady
Yes
Small-breed owner
shiny coat in a month
Rising
Bullet 5
Show-and-groom owner
Each phrase comes with its AI-prompt form ("best dog food for a dog that gets gas") and the markets where it is rising. The list exports straight to retail media and SEO tools.
01 · DiagnoseLanguage gap
Where your listings and content miss the words buyers use, against competitors.
02 · UnderstandBuyer language
Tracked by persona and market, plus the questions buyers put to AI assistants.
03 · ActTerm and prompt lists
Ready for retail media, SEO and content teams, grouped and prioritised.
04 · TestPersona check
Personas asked how they would search, to check the lists before campaigns run.
You receiveBuyer language library by persona and marketSearch-term list, prioritisedAI prompt listListing coverage checkMonthly rising-terms alert
Start with: Deep dive, one category and market. Rising terms tracked monthly inside the Intelligence programme.
Sell · 11
Make the case to a retailer
Buyer evidence the retailer does not have: why people choose, what is missing from the fixture, and which claims they trust. Built for the buyer meeting.
Human Intelligence leadsProduct Intelligence supports
What it is
A category story told from the shopper's side rather than the brand's. It reads the reviews of every product on the retailer's fixture, including the ones only that retailer sells, and shows the needs the range meets, the needs it misses, and the claims shoppers trust. The retailer's own data says what sold. This says why, and what didn't sell because it wasn't there.
Questions it answers
What do shoppers want that the fixture does not offer?
Which claims do shoppers trust in this category, and which do they ignore?
Where does the retailer's own-label win and lose in shoppers' words?
What would make a shopper switch to this range?
Supports and replaces
SupportsRange reviews, listing proposals, joint business planning, category captaincy
ReplacesCategory stories built from the brand's own sell-in deck, and shopper studies commissioned retailer by retailer
Hair care fixture, one retailer: what shoppers want versus what's stockedIllustrative
Shopper need
Share of reviews
Products on fixture that answer it
Shoppers say
Case
Scalp itch and flakes, fragrance-free
1 of 42
Had to order online, nothing in store
Gap to fill
Curl definition without crunch
3 of 42, all premium
Only the expensive ones work
Mid-tier gap
Colour protection
11 of 42
They all say it, half of them fade anyway
Over-served
Fast-drying for school mornings
0 of 42
Didn't know this existed until a friend said
Rising
Share of reviews is across the whole category, not just the fixture. The deck adds the retailer's own-label read and the claims shoppers trust most, so the proposal is about the shopper, not the brand.
01 · DiagnoseCategory from the shopper's side
Needs met, needs missed, brands trusted, across the fixture.
02 · UnderstandProduct by product
Buyer language and complaints for every product, including retailer exclusives.
03 · ActRetailer-facing story
The category story and the assortment gaps it points to, ready for the meeting.
04 · TestRange check
Persona reactions to the proposed range or fixture before you present it.
You receiveNeed-versus-fixture gap mapTrusted-claims readOwn-label strengths and weaknessesRetailer-ready category deckPersona check on the proposal
Start with: Deep dive, one category at one retailer. Every retailer, refreshed before each review cycle, inside the Intelligence programme.
Grow · 12
Win on occasions
When, why and for whom people buy. Which moments you already own, which belong to competitors, and which are open to win.
Human Intelligence leadsMachine Intelligence supports
What it is
Buyers say who they bought for and when. "For my dad's birthday." "For the kids' lunchboxes." "For the office." This reads those moments across every review in the category, sizes each occasion by how often it appears, maps which brands own it, and shows the message that works in each. Nobody has to be asked a survey question about an occasion they've forgotten.
Questions it answers
Who is the product really bought for, and when?
Which occasions do we own, and which does a competitor own?
What triggers a repeat purchase versus a one-off?
Which occasions are growing, and which are open?
Supports and replaces
SupportsSeasonal and campaign planning, gifting ranges, pack formats, demand-space strategy
ReplacesOne-off occasion and demand-space studies, and planning calendars built from last year's sales
Occasion map for premium chocolateIllustrative
Occasion, in buyers' words
Share of reviews
Owned by
What wins the moment
Status
Gift for someone who has everything
Brand A
Box looks expensive, arrives undamaged
Theirs
Friday night treat, just for me
You
Small format, resealable, "worth the calories"
Yours
Thank-you for the teacher / neighbour
Nobody
Mid-price, looks thoughtful, no allergens
Open, rising
Office sharing box
Brand C
Individually wrapped, 40+ pieces
Theirs
Each occasion carries its trigger, its persona and its seasonality curve. The plan picks the moments to target and the message for each.
01 · DiagnoseOccasion ownership
Which occasions your brand owns today and which belong to competitors.
02 · UnderstandMoments and triggers
The moments, triggers and people behind each purchase, from gifting to everyday.
03 · ActOccasion plan
Which moments to target, the message for each, and where to show up.
04 · TestIdea check
Occasion-led ideas checked with personas before they go to creative.
You receiveOccasion map with shares and ownersTrigger and persona per occasionSeasonality curvesOccasion plan and messagesFindings deck and readout
Start with: Deep dive, one category and market. Occasions tracked through the year inside the Intelligence programme.
Grow · 13
Win switchers and keep buyers
Buyers say outright who they switched from and why. See the flows in both directions, the reasons on each side, and the moments that trigger a switch.
Human Intelligence leadsMachine Intelligence supportsProduct Intelligence supports
What it is
"Switched from Brand X after the leaks." "Came back to this after trying the cheaper one." Reviews are full of switching stories with the brand named and the reason given. This reads every one of them, so you see who you win buyers from, who you lose them to, and the exact reason in each direction, as it happens rather than months later in a survey.
Questions it answers
Why do people leave us, and where do they go?
Which brand are we winning buyers from, and on what?
What moment makes someone switch: a leak, a price rise, a new baby, a new size?
Who comes back, and what brought them back?
Supports and replaces
SupportsRetention and win-back plans, CRM messaging, product fixes, competitive response
ReplacesLoyalty surveys that ask about switching months after it happened, and churn models that show the exit without the reason
Switching flows for a nappy brandIllustrative
From Brand A+ 41%Theirs leaked every night at size 4, yours didn't
From Brand B+ 27%Fragrance gave her a rash, switched and it cleared
From own-label+ 19%Tried to save money, back after a week
→
YouNet switching in your favour this quarter
LeavingWhere your buyers go
→
To Brand C− 38%Yours stopped fitting at size 6, theirs runs bigger
To Brand A− 24%Subscription price went up, theirs was cheaper that month
Percentages are shares of switching mentions, not of sales. Each flow opens to the reviews, the persona and the trigger. The size-6 exit is the fix; the price exit is the win-back.
01 · DiagnoseSwitching flows
Which brands you win buyers from, which you lose them to, and the net picture.
02 · UnderstandReasons each way
The reasons on each side, in buyers' words, tracked as new reviews arrive.
03 · ActWin and keep plan
The messages, fixes and moments that win switchers and close the exits.
04 · TestWin-back check
Retention and win-back ideas checked with the personas most likely to leave.
You receiveSwitching flow mapReason library, both directionsTrigger moments by personaWin and keep planMonthly flow update
Start with: Deep dive, one brand and market. Flows updated monthly inside the Intelligence programme.
Defend · 14
Answer competitor launches
Early warning on rival launches and claims, and a read on whether buyers believe them, so you decide how to respond on evidence rather than instinct.
New products, new claims and new pack messaging are captured the day they appear on the shelf, in every market you care about. Then the first reviews come in and tell you whether the launch delivers what it promises. Most launches don't need a response. The few that do, you'll know about in the first fortnight rather than the next quarter.
Questions it answers
What did competitors launch this month, and with which claims?
Are buyers backing the new claim, or finding it out?
Has AI changed its recommendation since the launch?
Which launches need an answer, and which can be left alone?
Supports and replaces
SupportsCompetitive response, claims defence, launch planning, sales team briefings
ReplacesStore checks and manual competitor tracking, and the quarterly competitor deck that arrives after the response window has closed
Launch alerts, protein powder, last 60 daysIllustrative
Day 1 · New launchBrand A · "Clear whey, tastes like juice"New format, 3 flavours, launched in 2 of your 5 markets. Pack claim: "zero chalky texture".
Day 14 · First reviewsBuyers back the taste, not the mixability"Tastes like juice" confirmed in 68% of reviews. "Foams and won't settle" in 41%. Your mixability is your most praised attribute.
Day 21 · AI shiftTwo assistants now name Brand A for "protein that doesn't taste like protein"Your brand still named for "mixes without a shaker". No change on your core questions.
Response callDefend, don't matchLead with mixability in the two launch markets. Watch the taste question. No format response needed yet.
Every launch in the category gets the same three-stage read. Only the ones that move buyers or AI reach your inbox.
01 · DiagnoseLaunch detection
New launches and pack claims across the category as they appear.
02 · UnderstandBuyer verdict
Whether buyers believe the new claims, what they praise, where the launch falls short.
03 · ActResponse plan
Which claims to defend, counter or match, and where.
04 · TestAfter the response
How AI now compares you, and how personas react to your answer.
You receiveLaunch alerts with pack claimsFirst-review verdict at day 14AI shift checkResponse call per launchMonthly launch digest
Start with: Deep dive, one category, the last 12 months of launches. Continuous alerts inside the Intelligence programme.
Defend · 15
Fix what buyers complain about
Problems show up in reviews weeks before they show in returns or ratings. See which complaints are growing, which fix matters most, and where the pack promises more than the product delivers.
Human Intelligence leadsProduct Intelligence supportsMachine Intelligence supports
What it is
Every complaint in your reviews, grouped, counted and tracked over time by product and market. Rising complaints trigger an alert. Each is checked against the claim on pack, so you can see whether it's a product problem or a promise problem. The list is ordered by what each fix would do for the rating, so engineering and marketing work on the right thing first.
Questions it answers
What is our most common complaint, and is it growing?
Is this a batch, a market, or every unit?
Where does the pack promise more than the product delivers?
Which fix would lift the rating most?
Supports and replaces
SupportsQuality and R&D priorities, pack claim review, customer care briefings, rating protection
ReplacesWaiting for returns data, ratings slides and customer-care tickets, and quarterly review summaries read too late
Complaint ranking for a baby bottle rangeIllustrative
Complaint, in buyers' words
Share of 1–3★
Trend
Pack says
Priority
Leaks from the collar when tipped
"Leak-proof"
Fix first
Measurement marks rub off
Nothing
Fix next
Hard to clean the vent
"Easy clean"
Reword or fix
Teat collapses mid-feed
"Anti-colic"
New, watch
Row one is rising and contradicts the lead claim, so it's first. Row four is small but new and climbing, flagged three weeks before it would show in the star rating. Each row shows which market and which size it comes from.
01 · DiagnoseComplaint ranking
Complaints by how often they appear and whether they are rising, by product and market.
02 · UnderstandEarly warning
Alerts on new issues, plus where the pack claim and the experience in use disagree.
03 · ActFix list
In order of impact: the issues that drag ratings down and the quick wins.
04 · TestAfter the fix
The fixed version checked with the personas who complained most.
You receiveComplaint ranking by product and marketRising-issue alertsClaim-versus-experience checkFix list in order of impactMonthly quality readout
Start with: Deep dive, one range and market. Alerts run continuously inside the Intelligence programme.
Defend · 16
Prove it worked
Before-and-after evidence across buyers, pages and AI. Did buyers start using the new language, did the pages change, did AI change its answer. Read at 30, 60 and 90 days.
Media metrics stop at the click. This measures what happened after: whether the message landed in buyers' own words, whether the shelf changed, and whether AI assistants now say something different about you. A baseline is taken before the work ships, then the same three reads are repeated against competitors over the same period, so what moved can be separated from what would have moved anyway.
Questions it answers
Did buyers start using our words?
Did AI change its recommendation, and on which questions?
Did the page and pack changes land, and did competitors move too?
What should the next round fix?
Supports and replaces
SupportsCampaign effectiveness, agency QBRs and renewals, launch reviews, board reporting
ReplacesMedia metrics that stop at the click, and brand trackers that report a quarter after the campaign ended
90-day read on a sunscreen relaunch, "no white cast" messageIllustrative
Buyers
9% → 31%+22 pts
Share of your reviews that mention "no white cast" unprompted. Category average moved 2 pts.
Pages
3 of 14 → 14 of 14
Listings carrying the message in the title or first bullet. Two competitors added similar wording in the same window.
AI
1 of 7 → 5 of 7+4 assistants
Assistants naming you for "sunscreen that doesn't leave a white cast". Reason cited: your reviews, not your ad.
Read
Day 0
Day 30
Day 60
Day 90
Next fix
"No white cast" in reviews
9%
14%
24%
31%
Hold
"Greasy feel" complaints
18%
19%
17%
17%
Unchanged, next round
AI accuracy on SPF rating
4 of 7 wrong
4 of 7
2 of 7
0 of 7
Fixed via listing
All three reads are run on competitors over the same 90 days, so the readout separates your effect from category drift.
01 · DiagnoseBaseline
Buyer language, page claims and AI recommendations as they stand before the work ships.
02 · Understand30 / 60 / 90
The same three reads repeated, against competitors over the same period.
03 · ActReadout
What moved, what did not, and what the next round should fix.
04 · TestPersona explain
Personas asked whether the new message landed, to explain the numbers.
You receiveBaseline across the three layers30 / 60 / 90-day readsCompetitor comparison over the same periodPlain readout of what movedNext-round fix list
Start with: Deep dive wrapped around one launch or campaign. Every campaign measured by default inside the Intelligence programme.
Agencies · 17
Win a pitch
Walk in with evidence on the prospect's buyers, shelf and AI visibility. A fast, evidence-led read of their category that becomes the backbone of the pitch.
Human Intelligence leadsProduct Intelligence supportsMachine Intelligence supports
What it is
The prospect's category read in days rather than weeks: how their brand shows up with buyers, on the shelf and in AI answers, where the open opportunity is, and the evidence to prove it. The client hears their own buyers' words in the room. Personas can be brought into the pitch so the client questions their buyers live.
Questions it answers
What does this brand's buyer really care about, in their words?
Where does the prospect fall behind on the shelf and in AI answers?
What is the one insight the incumbent agency hasn't shown them?
ReplacesDesk research squeezed into the week before a pitch, and the strategist's hunch presented as insight
Pitch pack spine for a dog food prospect, built in 5 daysIllustrative
Slide
The finding
Evidence
Layer
1 · The buyer
Three owner types; the "sensitive-stomach owner" is 38% of reviews and nobody talks to them
12,400 reviews, 3 markets
Human
2 · The shelf
Prospect leads on "grain-free"; buyers have moved on to "doesn't give him gas"
84 listings, every claim mapped
Product
3 · The machine
Named by 2 of 7 assistants; competitor named by 6, citing reviews the prospect also has
200 buyer questions, weekly
Machine
4 · The opportunity
Own "settles sensitive stomachs" across pack, page and message; AI follows within a quarter
Switching flows, claim gaps
All three
5 · Live
The client interviews the "sensitive-stomach owner" persona in the room
Persona chat
Human
Figures shown are the kind the pack carries; the categories and numbers here are illustrative. The pack is yours to present under your own brand.
01 · DiagnoseFast read
How the prospect's brand shows up with buyers, on the shelf and in AI answers.
02 · UnderstandCategory story
The category in buyers' own words, by market.
03 · ActPitch pack
The core insight, the opportunity and the evidence behind it, ready to present.
04 · TestPersonas in the room
The client questions their own buyers live.
You receiveCategory read in daysThree-layer brand snapshotPitch-ready evidence packPersona access for the roomFirst-90-days plan
Start with: Pitch Pack, one prospect, one category, up to three markets. Every pitch covered inside a Network licence.
Agencies · 18
Grow an account
A reason to bring the client something new every month. Fresh evidence on their buyers, shelf and AI visibility, so the account team always has the next idea and proof the last one worked.
Human Intelligence leadsMachine Intelligence supportsProduct Intelligence supports
What it is
A standing read of the client's category, embedded in the account team. Every month it says what changed with buyers, on the shelf and in AI answers, and turns that into opportunity notes for the client, briefs for planning and creative, and measurement for QBRs. The agency becomes the team that always knows first.
Questions it answers
What changed in the client's category this month?
What should the next brief be about?
Did last quarter's work move anything we can show?
Which client team haven't we helped yet?
Supports and replaces
SupportsAccount planning, QBRs and renewals, proactive briefs, cross-selling into insights, shelf and innovation teams
ReplacesAd hoc research bought one brief at a time, and the monthly status meeting with nothing new in it
One month on a plant-based milk accountIllustrative
Signal this month
Layer
Turned into
For
"Splits in hot coffee" complaints up 40% in Australia after a formula change
Human
Early-warning note to the client's quality team
Client R&D
Competitor launched "barista" variant with "froths like dairy" claim
Product
Response brief: defend froth, don't match pack
Planning
Two assistants now name the competitor for "oat milk for lattes"
Machine
Page fix list for the top 6 listings
Client eCom
"Makes my kid actually drink milk" rising across 3 markets
Human
Proactive occasion brief: family breakfast
Creative
Last quarter's "no gums" campaign: mention in reviews 6% → 19%
All three
QBR proof slide
Account lead
Five signals, five pieces of work, four client teams reached. The same read runs every month without a new brief.
01 · DiagnoseStanding read
What changed in the client's category with buyers, on the shelf and in AI answers.
02 · UnderstandAlways-on data
Planning and strategy explore the data for the next brief.
03 · ActMonthly notes
Opportunity notes for the client, plus measurement for QBRs and renewals.
04 · TestPersonas for every team
Available to planning, creative and account for briefing and checking work.
You receiveMonthly category readOpportunity notes per client teamQBR measurementLive data for planningPersonas for the whole account
Start with: Client-embedded, one account. Every account across the network inside a Network licence, or under your own name as White-label.
Build · 19
Build our own AI on buyer evidence
Real buyer truth as the feed for your own tools and copilots. Structured reviews, claims and AI answers delivered as query-ready data and personas, so your internal AI is grounded in what buyers actually say.
Human Intelligence leadsProduct Intelligence supportsMachine Intelligence supports
What it is
Most internal copilots are trained on the company's own data: sales, CRM, past decks. They can tell you what you sold. They cannot tell you why buyers chose you, what they say about the competitor, or what AI assistants are telling shoppers today. This is that missing feed: structured, refreshed, in the shape your tools need, with personas ready to plug in.
Questions it answers
Can our copilot answer why buyers choose us, with evidence?
Does our AI use buyers' words or our own?
Is our data fresh enough to trust for a decision this week?
Can every team ask the same questions without a research brief?
Supports and replaces
SupportsInternal copilots and agents, data platforms, insight self-serve, AI transformation programmes
ReplacesInternal AI trained on first-party data alone, and one-off data exports that are stale by the time they're loaded
A beauty company's copilot, before and after the feedIllustrative
Team asks
Why is our vitamin C serum losing to Brand B in Korea?
Before
Sales down 8% in Q2. No further detail available.
After
Korean buyers switching to Brand B cite "no sticky finish" (412 reviews, rising). Your serum is praised for brightening but "tacky under sunscreen" appears in 23% of 3-star reviews. Brand B's page leads with "absorbs in 10 seconds". Four of seven AI assistants now name Brand B for "vitamin C serum for oily skin". Suggested first fix: texture claim on page and pack; persona "Morning-routine minimalist" available to test wording.
Feed
Structured reviews by topic and persona · claims by SKU and market · weekly AI answers by question · refreshed daily · delivered as API, warehouse tables or files.
The copilot is yours. The feed is ours. Each answer links back to the reviews, pages and AI responses it was built from.
01 · DiagnoseWhat your tools can't answer
The buyer questions first-party data alone cannot answer today.
02 · UnderstandThe feed
Structured review, claim and AI-recommendation data in the shape your tools need, refreshed.
03 · ActPlug in
Persona bots and language libraries ready for your assistants and workflows.
04 · TestSound like buyers
Your tools' outputs checked against the personas to confirm they sound like real buyers.
You receiveQuery-ready data feedTopic, persona and claim taxonomiesPersona bots for integrationLanguage libraries by marketDaily refresh and documentation
Start with: a Multi-market rollout with data access, scoped with your data team. Available as an add-on to the Intelligence programme.