Three conversations. Same month. Same problem.

A sales VP at a global CPG brand: "We're paying our agency £2M a year, but I still can't see what's happening in my top 500 stores."

An agency CEO in London: "We've grown revenue 40% in three years. Headcount is up 38%. Margins are down."

A commercial director at a major beverage company: "My field team spends their entire Monday compiling last week's data. By the time we see a problem, it's already cost us the weekend."

Different titles. Different P&Ls. Same trap.

The Economics Don't Work Anymore

Here's what nobody says out loud: the traditional retail execution model is broken.

Not because people aren't working hard. They are.

Not because the strategy is wrong. It isn't.

The math just doesn't work.

A field rep visits a store. Walks the aisles. Checks shelf placement, counts facings, photographs displays, notes out-of-stocks. Maybe records some competitor activity if there's time.

Twenty minutes later, they're done. One store.

Forty stores a week. That's 800 minutes—over 13 hours—just documenting what they see. Not capturing: full category data. Competitor SKU details. The complete shelf picture.

Not fixing anything. Not coaching store managers. Not building displays or negotiating better placement.

Documenting a fraction of what's actually happening.

The Agency Trap

If you run an agency, you know this trap intimately.

Client wants more stores covered? Hire more reps.

Client wants more frequent audits? Hire more reps.

Client wants competitor intelligence tracked? Good luck—there's no time.

Revenue grows linearly. Costs grow linearly. Sometimes faster.

The pitch from six years ago—"We'll give you eyes and ears in every store"—made sense when data was scarce. Now your clients are drowning in incomplete data. What they want is action and complete category visibility.

But that requires time. And your reps are spending 65% of their time on documentation.

We've watched agency heads try everything:

Better training. (Helps, but doesn't scale.)

Tighter processes. (Makes things consistent, not faster.)

Traditional image recognition. (Takes 7 months to implement, breaks when packaging changes.)

The trap tightens. Your best clients demand more. Your margins compress. Your competitors undercut on price.

The CPG Dilemma

If you're the brand, you face a different version of hell.

Your agency gives you a report. Friday afternoon. Data from Monday through Wednesday.

The promotional display that was supposed to drive your Q1 numbers? Built in 40% of stores. You find out nine days after launch.

Your competitor just took 15% more shelf space in the top chain. They've had it for eight days. And now their share of shelf over-indexes their market share. 

The new SKU you spent £3M launching? Out of stock in 30% of locations during peak weekend trading. You get the data Tuesday morning.

And here's the kicker: your agency can't tell you what your competitors are doing at SKU level. There's no time to capture that level of detail.

You want real-time visibility. The agency wants to give it to you. But their field team is already maxed out. 

Your budget is flat. Actually, it's down 8% this year.

So you make a choice: fewer stores, less frequent audits, or lower depth of data.

All three options lose.

What Changed

A commercial director at a major tobacco company showed us their field data. They had 47 reps covering 3,000 stores. Each rep visited about 64 stores a month.

They were tracking five KPIs. Visibility, availability, pricing, promotional compliance, competitor activity.

We asked: "What percentage of stores are you not seeing each month?"

She paused. "We cover about 12% of our universe monthly. The rest we extrapolate."

Twelve percent.

This wasn't a small brand. This was a top-three player in their category. Well-funded. Smart team. Good agency partner.

Still flying blind in 88% of their stores.

The problem isn't effort. It's physics. There are only so many hours in a day. Only so many stores a human can visit. Only so much time they can spend per store.

Until recently, that was just the constraint everyone lived with.

The Zero-Shot Breakthrough

Image recognition has been seen like the saving grace to drive productivity. In reality, things have not been so smooth.

Until now. Enters zero-shot learning.

Not easy to explain this without getting technical.

But here's what matters:

Traditional image recognition requires training. You upload thousands of photos of your products. Someone must manually annotate those pictures. They will also make mistakes. Then someone must train the model so that it learns what a Heineken can looks like versus a Carlsberg can. Takes months. When your packaging changes, you start over.

Zero-shot learning changes the game. The AI doesn't need massive training data. Zero manual annotation required. It recognizes objects it's never seen before. New SKU launches? New packaging? Competitor products? Works immediately.

This isn't theory. It's working right now in 21 countries across thousands of stores visited daily by thousands of reps taking millions of shelves pictures every year.

The same audit that took a field rep 20 minutes now takes five.

Five minutes.

Same accuracy. Better, actually—95% or higher, verified against human audits. But 75% faster. And 10 x the amount of data because, yes, the full category is captured at SKU level. Not just your products. Every product on that shelf. Every competitor. Every emerging brand. Every packaging variant.

What Five Minutes Means

Let's stay with that tobacco company.

Their 47 reps were spending 13 hours a week on data capture. Per rep. That's 611 hours a week across the team.

At five minutes per store instead of 20, that drops to 163 hours a week.

That's 448 hours freed up. Every week.

What can you do with 448 hours?

Fix issues in real-time. Actually build the displays that aren't built. Coach the store managers. Negotiate better placement. Focus on the stores that matter most.

Or—and this is what most agencies choose—cover more stores. A lot more.

Same team. Same cost. Four times the coverage.

Suddenly that 12% visibility becomes 48%. You're not extrapolating anymore.

Three Perspectives

From the Agency CEO's desk:

Your client retention just got easier. When you can show a brand director live dashboards with yesterday's data—not last week's PDF—the conversation changes.

Your new business pitch just got sharper. "AI-powered execution" isn't marketing speak. It's a genuine capability your traditional competitors can't match.

Your margins improve. Not because you're cutting corners. Because your cost per store just dropped. 

You can finally invest in growth. In training. In technology. In the smart people you've wanted to hire but couldn't justify.

From the CPG Commercial Director's desk:

Monday morning. You log in. Live data from Friday and Saturday—your highest volume days. Out-of-stocks by store, by SKU. Promotional compliance by location. Share of shelf versus target.

You see problems while they're still solvable, not post-mortems.

Your field team's time isn't wasted on administration. They're solving issues. Building relationships. Driving execution.

Your agency relationship improves. You're not constantly asking for more data at the same price. They're delivering more value at better economics.

From the Field Rep's perspective:

You walk into a store. Take photos of the shelf and displays. five minutes for the ones with several categories to cover.

The app immediately shows you: green lights where execution is good, red lights where there are issues. Out-of-stock on SKU #47. Promotional display missing. Competitor took 20% more space than last week.

And something you couldn't do before: you see exactly what competitors are doing. Every SKU. Every price point. Every new product they've launched.

You spend the next 15 minutes fixing things. Not documenting them. Fixing them.

You leave the store having actually improved execution. Not just reported on it.

Your job gets more satisfying. More strategic. More valuable.

The Implementation Question

I know what you're thinking. This sounds good. Too good.

What's the catch?

The catch is change. Not technology change—that's the easy part. It takes six weeks, not six months.

The hard part is mental.

For 20 years, field execution meant human observation plus manual reporting. That's how everyone was trained. That's what "good" looked like.

Now you're trusting an AI to see a shelf and tell you what's there.

The agencies who've made this transition did a few things right:

They piloted small. Fifty stores. Ran it parallel to traditional methods for three weeks. Verified the accuracy. Built trust.

They involved their field teams early. Showed them how the technology makes their jobs better, not obsolete. Trained them on what to do with the 15 minutes they get back.

They set clear KPIs. Not "use the new technology." But "improve promotional compliance by 15%" or "capture 100% competitor SKU data."

They chose technology partners who understood retail operations, not just AI. Because knowing why the shelf level  actually matters makes the technology much more than a productivity tool

The agencies who struggled? They treated it as a technology project instead of an operational transformation.

What We're Learning

We work with agencies across Europe serving over 300 CPG brands. Food and beverage, tobacco, pharma, consumer electronics.

The pattern is consistent:

Week 1-2: Deployment and training. Field teams are skeptical. "This will never work in my stores."

Week 3-4: Pilot phase. Accuracy checks. The skepticism starts cracking. "Wait, this actually caught things I missed."

Week 5-6: Scale-up. The field teams become advocates. "Can we get this in all my stores?"

Month 2-3: Results become visible. Out-of-stocks drop. Promotional compliance improves. Category intelligence appears for the first time.

Month 4+: Expansion requests. "Can we cover these additional channels?" "Can we roll this to our other markets?"

The average agency sees measurable ROI within 90 days. Not technology ROI. Business ROI.

The Uncomfortable Truth

Not every agency will make this shift.

Some will stick with the traditional model. Keep hiring reps linearly as they grow. Keep compressing margins. Keep competing on price.

Some will try to build this technology themselves and will inevitably fail. Not because they're not smart, but because AI development isn't their core business.

The ones who win will recognize this moment for what it is: a fundamental reset in the economics of retail execution.

Where the constraint was always human time and cost, and now it isn't.

Where the choice was always coverage or depth, and now you can have both.

Where real-time execution was aspirational, and now it's operational.

Why This Matters Now

The CPG industry is under pressure. Physical retail still drives FMCG sales. Winning at the shelf matters more than ever.

Retailer consolidation means fewer, larger customers with more power. They're demanding better execution as table stakes for distribution.

Brand portfolios are getting more complex. More SKUs, more limited editions, more localized variants. The old "survey 500 stores quarterly" approach can't keep up.

And budgets aren't growing. If anything, they're shrinking.

So you need better execution, more visibility, faster action—at the same or lower cost.

The agencies who figure this out will own the next decade. 

A Question Worth Asking

What is your field team actually doing with their time?

Not what you think they're doing. Not what they should be doing.

What they're actually doing.

Track it for a week. Audit time. Documentation time. Drive time. Administrative time. Problem-solving time. Relationship-building time.

If more than half their time is on documentation and administration, you have a problem worth solving. 

Actually, you have a burning platform.

What We Believe

Retail execution should be about execution, not documentation.

Field teams should spend their time fixing problems, not recording them.

Agencies should grow by serving clients better, not just by adding headcount.

CPG brands should see their stores in real-time, not in weekly reports.

Technology should make hard jobs easier, not replace the people doing them.

And complete category intelligence should be standard, not a luxury.

The Real Choice

This isn't about BeMyEye versus anyone else.

It's about two paths:

Path one: Keep doing what you're doing. Hire more reps. Compress margins. Compete on price. Capture incomplete data.

Path two: Fundamentally rethink field efficiency. Use technology to multiply human capability. Deliver more value at better economics. Build a competitive moat.

Both paths are viable. For now.

But the distance between them is growing fast.

The agencies choosing path two are seeing it in their numbers. Client retention up. New client win rates improving. Margin expansion for the first time in years.

The CPG brands working with these agencies are seeing it too. Execution improvements of 25-30%. Out-of-stocks down. Promotional compliance up. Share of shelf gains.

Not because anyone is working harder.

Because 15 minutes per store just became available to do actual work.