This presentation is built for a bigger screen. Please open it on a desktop or laptop to explore the full experience.
AI SEARCH VISIBILITYSobeys
Where{{ clientShort }} showsupin the_answer▌
A research-grade read on how five AI models answer Canada's grocery questions, presented by Bold AI.
Canadians are increasingly asking AI engines "how to save on groceries."
When money's tight, the answer usually isn't {{ clientShort }}.
SCROLL↓
JUL 21, 2026
INTRODUCTIONS
Your team.
Alan Macdonald
Co-Founder, Bold AI
Doctoral researcher in AI and best-selling co-author of The AI Revolution.
Alexandra Kolar
Advisory & GTM Lead
Enterprise marketer, performance strategist and GTM lead for podium iq.
TODAY
The agenda.
01
The shift
How discovery moved from links you click to answers you trust.
02
The platform
Research-grade measurement across five models, traceable to source.
03
The findings
What five models actually say when Canadians ask about groceries. Live run data.
04
The partnership
From insight to action: how we work together and what it costs.
PREAMBLE
This is more than a proposal. It's a live readout of an actual podium run, completed July 21st, with real insights about how AI sees Sobeys. Pay close attention to the findings.
01
SECTION 01 OF 04
The shift.
How discovery moved from links you click to answers you trust. Four changes that decide who gets found.
01
SHIFT 01
AI search is overtaking traditional search.
2028
AI-driven responses overtake traditional search.
Model traffic will soon be the single source of truth.
SHIFT 02 · LIVE
If you're not in the first 200 tokens, you're effectively invisible.
PROMPT
“{{ sampleQuery }}”
Model: Claude Opus 4.7Web search: ON
TOKEN000
Scroll. You're driving the model's output.
↓KEEP SCROLLING TO REVEAL THE ANSWER
LIVE ANSWER
Won inside the windowLost after it
03
SHIFT 03
The journey now ends inside the answer. Not on your site.
Which grocery store should I shop at in Canada?
Canada's grocery market is led by a few national chains competing on price, quality and loyalty. Sobeys, owned by Empire, runs a full-service national network and the Scene+ rewards program, and is one of the most-cited names when shoppers ask where to buy groceries…
sobeys.comreddit.comwalmart.ca
Decision made.The shopper never leaves.
Zero-click is real.
The question is asked, weighed and answered in one place. If the model doesn't cite you, the shopper never meets you.
SHIFT 04 · A SIDEWAYS JOURNEY
The evolution of discovery.
Four eras. Each one moves the decision further from your website, and deeper into the model.
KEEP SCROLLING →
sobeys.com
FIG. 01 · THE URLVISIBILITY: YOURS
1 Direct load
You type the URL. The decision was already made.
best grocery store in Canada
FIG. 02 · THE QUERYVISIBILITY: RANKED
2 Search
You type the question. The algorithm picks the answers you choose from.
I'm feeding a family of four in the GTA and grocery bills keep climbing. I want good produce, real savings and a loyalty program that's actually worth it. Which Canadian grocery chain should I switch to?
+Tools
FIG. 03 · THE CONVERSATIONWE ARE HERE
3 Answer engines
You share your whole situation, not just a keyword. It understands. It answers.
FIG. 04 · THE GOALVISIBILITY: MACHINE-READ
4 Agentic
You set the goal. AI handles the rest: research, comparison, booking.
BEFORE
The user did the work.
Several open tabs. Mixed results. Users stitched the answer together themselves.
NOW
The model does the work.
AI reads, decides, and cites credible sources. One confident answer.
In search you compete for rank. In AI you compete for inclusion.
02
SECTION 02 OF 04
The platform.
podium iq: research-grade measurement of how five models see your brand, with every insight traceable to its source.
Evidence-based visibility intelligence that shows exactly where your content falls out of the model's line of sight, and how to fix it.
01 · MEASURE
Continuous synthetic testing across 5 LLMs and thousands of queries.
02 · DIAGNOSE
Identifies your specific visibility gaps, and why they exist.
03 · ACT
A prioritized action engine that closes the gap in days, not months.
Tested acrossAnthropicOpenAIPerplexityGrokGoogle
03
SECTION 03 OF 04
The findings.
What the models say when Canadians ask how to shop and save on groceries — especially before they ever name a store. Run 1, live data, nothing left to opinion.
best grocery store in canada · cheapest groceries in canada · sobeys vs walmart · grocery delivery & pickup · scene+ points value · affordable groceries · online grocery shopping · best grocery store in canada · cheapest groceries in canada · sobeys vs walmart · grocery delivery & pickup · scene+ points value · affordable groceries · online grocery shopping ·
RUN 1 · {{ clientNameUpper }} · JUL 14, 2026
LIVE RUN DATA · TRACEABLE TO SOURCE
What we analyzed.
Branded and unbranded grocery questions across the full journey: best store overall, affordability and value, product quality, loyalty and Scene+, online and delivery, and how you stack up against Walmart, Loblaws, No Frills and Costco. 47 branded, 86 unbranded — and the unbranded ones are where new shoppers are won.
{{ nQueries }}queries analyzed
{{ nCitations }}model responses traced
{{ nSources }}unique source domains
{{ nTopics }}grocery topics
01
FINDING 01 · BRANDED VS UNBRANDED
The {{ visCbs }}% headline hides two very different realities.
{{ visCbs }}% reported average
{{ brandedVis }}%Brandedwhen they name you · brand defence
{{ unbrandedVis }}%Unbrandedwhen they don't · real discovery
{{ unbrandedVis }}%of unbranded answers mention you — the shopper who hasn't named you yet
The reported {{ visCbs }}% blends near-total brand defence with weak open-category discovery. Strip the branded prompts and you appear in fewer than half the answers a new shopper sees. That number — not {{ visCbs }} — is the one that should run the scorecard.
02
FINDING 02 · THE VALUE GAP
When money is tight, the answer isn't you.
Groceries are the cost Canadians feel most right now. When shoppers ask AI how to eat well for less, the models answer — but they reach for the discounters first.
↓KEEP SCROLLING TO STEP THROUGH
AFFORDABILITY · IN THE SET72%
You appear in 72% of affordability answers — on price, you're at least in the conversation.
"CHEAPEST" · WHO OWNS IT89%
But No Frills owns "cheapest" at 89% and Walmart at 77%. On price you're compared, not chosen.
SAVING ADVICE · UNBRANDED9.5%
Ask how to actually save on groceries and your unbranded coverage falls to 9.5%. The topic "win" is carried entirely by branded prompts.
You appear in 72% of affordability answers.
03
FINDING 03 · MODEL VARIANCE
Which AI they open changes how you show up.
Grok
{{ modelGrok }}%
Perplexity
{{ modelPerplexity }}%
Google
{{ modelGemini }}%
Anthropic
{{ modelClaude }}%
OpenAI
{{ modelChatgpt }}%
58%coverage when the model searches the live web — down from 63% on trained memory alone
Google and Grok lean on what they already know and cite you often. Turn on live web search and more alternatives crowd in — your inclusion slips. That is an authority problem, not a recognition one: the open web rewards the brands it can corroborate.
04
FINDING 04 · THE QUESTIONS YOU'RE MISSING
One in three value questions never says your name.
32%of unbranded questions — 23 of 72 — return zero Sobeys mentions across nine model runs each
Another 46% surface you in fewer than half their answers. The gaps cluster exactly where a stretched shopper is deciding: price, quality, trust and how to save.
HIGH-VALUE QUERIES · ZERO SOBEYS MENTIONS
0 / 9"Where can I find affordable healthy groceries on a tight budget?"
0 / 9"Best ways to save money on groceries without sacrificing quality"
0 / 9"What is the most trusted grocery store for product quality?"
0 / 9"Which grocery stores rank highest for product-quality satisfaction?"
0 / 9"Where to buy premium-quality seafood at a grocery store?"
These are the questions Canadians actually type when money is tight. Answer them, and you enter the conversation you're missing.
05
FINDING 05 · YOUR OWN PAGES CAN'T BE READ
AI already cites your pages. It can barely read them.
Forty-five of your own URLs are cited by the models — but the pages doing that work are nearly invisible. 86.7% carry a blank or generic title. The fastest-payback lever isn't more content; it's making the pages AI already trusts legible.
YOUR 45 CITED PAGES · BY TITLE QUALITY
26blank title
13generic "sobeys.com"
6descriptive title
Unique titles, answer-first headings and structured data make these pages far easier for AI to interpret and cite.
WHERE AUTHORITY FORMS · TOP EXTERNAL DOMAINS
reddit.com
43
supermarketnews.com
41
narcity.com
31
dailyhive.com
25
en.wikipedia.org
21
Citation occurrences. These are the surfaces to earn, correct and corroborate on — especially when models search the live web.
Repair the pages AI already cites, then earn corroboration on the surfaces above. That is how unbranded discovery moves.
THE TOPIC MAP · 9 TOPICS, 133 QUERIES
Where you win. Where you're contested.
WINNING
Sobeys vs competitorsGrocery loyalty & Scene+Best grocery store overallOnline & deliveryBrand defence (already saturated)
Comparisons, loyalty and delivery — where a shopper is already close to choosing — are decisively yours.
CONTESTED
Product quality · behind Whole Foods & CostcoAffordability · behind No Frills & WalmartGrocery saving strategiesPeople Also Asked · unbranded
The value and quality questions where new shoppers are formed. No Frills and Walmart own "affordable"; Whole Foods and Costco own "quality". This is where your next shopper is lost.
Every claim above is traceable to its source in the trace layer. Nothing is left to opinion.
06
THE PLAY · BUILD ONCE, DEPLOY EVERYWHERE
Modular content, deployed across the ecosystem.
The answer no longer lives on sobeys.com. Author one authoritative asset, atomize it once, then place the pieces on the third-party surfaces where the answer is actually formed. That is how you move beyond branded.com into the platforms and environments that accelerate exposure and awareness.
1ASSET
THE CANONICAL ASSETOne authoritative piece, authored once on sobeys.com.e.g. "Where to buy groceries in Canada: the complete value guide." Research, evidence, structure and schema done once — then fragmented into every format each channel can cite.
→ → →
OWNED3SURFACES
The surfaces you fully control — fix the metadata, then let them anchor the story.ATOMIZED INTO
· Value & weekly-flyer hub· Store locator & regional pages· Scene+ & rewards pages
sobeys.comsobeys.com/flyersceneplus.ca
FEEDS · Sobeys-owned surfaces
EARNED · PR5SURFACES
Where authority actually forms — the third-party surfaces models trust when they search the web.ATOMIZED INTO
· Food-media & trade-press pickup· Reddit & community answers· Wikipedia & retail-directory citations
One authored asset, every surface in the ecosystem. Beyond branded.com — into the platforms and environments where awareness compounds.
04
SECTION 04 OF 04
The partnership.
From insight to action: how we work together, what it costs, and who does the lifting.
INVESTMENT
Insight doesn't move the needle. Action does.
Every level of partnership moves you closer to decisions and outcomes.
DONE BY YOU
DONE FOR YOU
Starts at$25KONE-TIME REPORT
Podium Benchmark
A research-grade visibility report delivering competitive intelligence, funnel performance analysis, and prioritized content recommendations in a single readout.
GVI score across 3 AI models
Competitive benchmarking against 4 named competitors
Funnel stage breakdown showing where the brand wins and loses
Prioritized content recommendations ready to action
Delivered as PDF within 10 business days
INCLUDED3 models · 344 queries · 4 competitors
Starts at$36KANNUAL
Guided Intelligence
The intelligence, the strategy, and the expert in your corner. You execute. We make sure you never have to guess what to do next.
GVI score across 3 to 5 models
500 to 1,000 queries, 5 to 10 competitors
Always-on platform and visibility insights
AI prioritized recommendations for full-funnel visibility
Implementation battlecards, monthly strategic reviews, Slack support, system generated content
Speed. To operational, versus 4 to 6 months coordinating across departments.
Consistency. One unified narrative across ChatGPT, Claude, Perplexity and Google AI.
Measurable lift. Visibility, first-mention rates and citation authority tracked sprint by sprint.
One continuous sprint loop.
01Set up queries
02Analyze battlecards
03Content strategy
04Draft briefs
05Brief the pod
06Develop content
07Publish
08Re-run & measure
09Optimize ↺
Two dials drive pod investment: configuration (Core / Full / Custom pod) and sprint pace (6 or 8 sprints per year). Each sprint ships 6 to 12 battlecard-driven assets.
6-month minimum commitment. Final pricing determined through discussion of scope, goals, and capability requirements.
WHY PODIUM IQ
Results this quarter, not next year.
Built for operators, not presentations
Intelligence that spots the gaps, pods that ship the fixes: 10 to 30 pieces of content per month built for model dominance.
Validated by the best
Accenture's AI Search partner. TD Bank, Samsung, Rona, Raymond James, Bauer Hockey, Canada Goose, Hockey Canada.
Signal over noise
Data that drives decisions, not another dashboard that leaves your team drowning in metrics.
An extension of your team
Weekly working sessions, dedicated experts, mid-market speed without enterprise overhead.
Do you want to own the podium, or watch from the stands?