Our Meta Ads have stopped performing. Can you help us figure out what's wrong?

Absolutely. But we rarely start inside Ads Manager. When Meta Ads stop performing, it's often a symptom rather than the problem itself. Businesses begin optimising for cheaper clicks, higher ROAS or more traffic while losing sight of the customers they actually want to attract and grow. We start by understanding who your most valuable customers are, what they're trying to achieve and whether your advertising, website and customer experience are helping them make progress. Only then do we optimise campaigns. Better ads don't build better businesses. Better customer journeys do. Meta Ads don't stop working overnight. Businesses gradually lose alignment between who they're trying to attract, what customers need and the experience they deliver. That's what we look for first.

If we stop advertising, our sales will disappear. How do we reduce our dependence on paid media?

Paid media should be one part of a healthy growth system—not the system itself. We help businesses build stronger acquisition, customer experience, retention and referral engines so paid media can focus on what it does best: attracting and growing your most valuable customers. The result is a business that compounds growth instead of constantly buying it. The goal isn't to spend less on advertising. It's to build a business where every customer becomes more valuable over time.

We have traffic, but conversions are flat. What's stopping people from buying?

Most businesses treat all traffic as if it's the same. It isn't. Every visitor arrives with different intent, different context and a different level of readiness. Yet most websites deliver the same generic experience to everyone. The result is that businesses optimise for more clicks instead of more meaningful customer progress. Conversion isn't a single moment—it's a journey. We listen to the signals customers leave behind, understand what they're trying to achieve and design experiences that help the right customers progress towards meaningful action. The question isn't 'Why aren't people converting?' It's 'What are our best customers trying to achieve, and are we helping them get there?'

How do we move past plateaued growth?

Plateaued growth is rarely caused by one broken channel. More often, it's a sign that the systems which got you here weren't designed to take you further. We identify the constraint that's limiting progress, prioritise what matters next and build better systems that compound growth over time. Businesses don't outgrow markets. They outgrow their systems.

Everyone says we should be customer-centric. What does that actually mean?

Being customer-centric isn't about trying to serve everyone equally. It's about understanding which customers create the greatest long-term value, what they're trying to achieve and designing every interaction to help them make progress. The goal isn't simply to attract more customers—it's to attract, convert, retain and grow the customers you're best equipped to serve. Customer-centric businesses don't optimise for transactions. They optimise for customer progress—and profitable long-term relationships.

We have customer data everywhere. How do we make it useful?

Most businesses don't have a data problem. They have a listening problem. Customer signals exist across your website, CRM, support conversations, marketing platforms and purchase history. The challenge is connecting those signals to understand where customers are progressing, where they're getting stuck and what they need next. The most valuable data isn't what customers tell you. It's what their behaviour has been telling you all along.

How do we optimise our website for ChatGPT and AI search?

AI rewards businesses that communicate clearly, build trust and organise information around real customer needs. The businesses that will stand out aren't the ones producing the most content—they're the ones creating the clearest understanding of who they help, what they sell and why customers choose them. AI doesn't reward better content. It rewards better understanding.

We're growing, but everything feels harder than it used to. Is that normal?

Very. Growth creates complexity. Complexity exposes weak systems. The businesses that continue to grow aren't necessarily the ones spending the most on marketing—they're the ones building better operating systems that scale with their customers, teams and ambitions. Growth doesn't create chaos. It reveals it.

We know something isn't working, but we can't figure out what's holding us back.

That's where we start. Rather than jumping straight into solutions, we identify the constraint that's preventing customers and the business from progressing. Solving the right problem creates momentum everywhere else. The hardest problems aren't difficult because they're complex. They're difficult because they're usually the wrong problems.

How do we build a business that continues to grow?

Sustainable growth isn't built campaign by campaign. It's built by creating systems that consistently attract, convert, retain and grow your most valuable customers while learning from every interaction along the way. The businesses that grow the fastest aren't the ones with the biggest budgets. They're the ones that learn the fastest.

How to Build a Business AI Can Understand

🎧 Listen to this Insight

AI shopping is creating a strange new problem for consumer brands.

A customer can now ask an AI what to buy, compare a shortlist, leave the AI interface, search a brand name, check Reddit or reviews, visit Amazon, buy through a retailer, and never generate a clean referral trail back to the recommendation that started the journey.

For years, digital commerce taught brands to think in pages, channels and clicks. AI is increasingly operating one layer above them: interpreting products, use cases, evidence and customer intent before the brand sees a session at all.

That changes what “AI readiness” should mean.

The conventional response has been to treat AI discovery as another optimisation problem: add schema, publish more content, monitor citations, improve product feeds, create FAQ pages and find ways to appear in ChatGPT, Gemini and other answer engines.

Those things can help. But they solve only one part of the problem.

A business can be visible to AI and still be poorly understood. It can be accurately recommended and still fail the customer's trust test. And it can influence a purchase through AI while its own attribution system credits Google, direct traffic, a marketplace or a store.

The more useful question is not simply:

How do we rank in AI?

It is:

Can machines understand the business accurately, can people believe the recommendation, and can the business learn from what happens next?

That is a commerce-system problem, not an SEO tactic.

The business problem: AI is changing the decision layer

Most commerce architecture assumes that the observable customer journey begins when someone arrives on a surface the business can measure.

A search click. An ad click. A marketplace visit. A website session. An email open.

AI weakens that assumption.

The customer can now do meaningful discovery and evaluation before any of those events occur. The AI may interpret a need, compare options, explain a product, surface evidence and shape the shortlist. The merchant may see only the downstream action.

That creates three gaps.

An understanding gap. Can the AI reconstruct what the product is, who it is for, when it is useful, what differentiates it and what evidence supports those claims?

A belief gap. Once the AI makes a recommendation, is there enough credible human proof for the customer to trust it?

A learning gap. If the recommendation changes consideration but the final visit arrives through search, direct, retail or a marketplace, can the business recognise that the journey changed?

Treating all three as “GEO” compresses a much larger operating problem into a marketing tactic.

Why conventional advice falls short

Most AI-discovery advice starts with the interface.

Optimise the page. Add structured data. Publish answer-friendly content. Track AI referrals. Increase mentions.

The problem is that interfaces will change.

The durable asset is not a perfectly optimised page for one answer engine. It is a business that can explain itself consistently across interfaces.

That requires more than content.

A product needs a clear identity. Its variants, ingredients or components, use cases, audiences, benefits, limitations, comparisons, evidence and relationships to other products need to be explicit enough that a machine does not have to infer the business from disconnected marketing copy.

Then the customer needs proof.

Reviews, FAQs, creator experience, community discussion, product evidence, comparison information, expert context and real usage stories become important because AI can compress research without eliminating uncertainty.

Finally, the business needs a way to learn even when the AI interaction itself is invisible.

If AI creates consideration and Google receives the final click, last-click attribution is technically correct and strategically incomplete.

The Operator Perspective: build for understanding, belief and learning

At Rekindle, we think a business should become understandable to its people, customers, systems and machines.

That idea becomes more useful as AI moves into commerce.

AI readiness is not a new content channel sitting beside SEO, paid media and marketplaces. It is a test of whether the business has enough structured commercial knowledge to be represented accurately outside the interfaces it controls.

The strongest architecture therefore has three layers.

1. Machine Understanding

Create clear entities and relationships around the things the business actually knows.

For a consumer brand, that can include:

Products and variants
Ingredients, materials or components
Benefits and supported claims
Use cases and occasions
Customer needs and audiences
Categories and collections
Comparisons and alternatives
Instructions, routines and usage
FAQs and objections
Evidence and trust signals
Availability, geography and fulfilment context

The goal is not to create a giant ontology for its own sake.

The goal is to reduce ambiguity.

If the business says a product is suitable for a particular use case, that relationship should exist consistently across the product page, structured data, editorial content, FAQs, feeds and other relevant surfaces.

2. Human Belief

Being understood is not the same as being believed.

AI can make a shortlist. The customer may still ask: does this actually work for someone like me?

That is where proof architecture matters.

Map important commercial claims to the evidence that supports them.

A useful proof system might connect:

Claim → Evidence → Source → Customer experience → Context → Limitation

That evidence can include verified reviews, credible third-party coverage, expert validation where appropriate, product testing, transparent comparisons, creator experience, community discussion and useful long-term customer stories.

This is not about manufacturing mentions or flooding communities with branded content. It is about making genuine evidence easier to find, understand and evaluate.

3. Business Learning

The third layer is the easiest to miss.

If AI increasingly influences decisions upstream, brands need to stop asking only how much directly attributed AI revenue they received.

Instead ask:

Are branded searches changing?
Are customers arriving with different levels of product knowledge?
Are particular questions, use cases or comparisons appearing more often?
Are AI-referred sessions behaving differently?
Are customers entering through marketplaces or retail after AI-mediated discovery?
Which products and claims are repeatedly represented by AI systems?
Where are recommendations inaccurate, incomplete or unsupported?

This is where AI discovery meets Customer Intelligence.

The business needs enough customer and commercial memory to interpret the downstream signals, even when it cannot observe every upstream interaction.

A practical framework: Understand → Prove → Observe → Learn

Use four questions to diagnose AI readiness.

Understand

What must an AI understand to recommend us accurately?

Identify the products, customer needs, use cases, claims, comparisons and relationships that matter commercially. Check whether those facts are explicit, consistent and accessible across the business.

Prove

What would a skeptical customer need to believe the recommendation?

For every important claim, identify the strongest credible evidence. Find the gaps where the brand is asking customers—or machines—to take unsupported language on trust.

Observe

What signals would tell us AI influenced the journey even without a referral click?

Define leading and downstream indicators rather than relying on one attribution bucket.

Learn

How does that evidence change what the business does next?

Feed inaccurate AI representations, emerging customer questions, validation behaviour and downstream commercial signals back into product knowledge, content, merchandising, customer experience and decision making.

That creates a loop:

Business knowledge → AI interpretation → Customer validation → Customer action → Business learning → Better business knowledge

The advantage compounds because each cycle improves both how the business is represented and how well it understands the customer journey.

What leaders should do next

Start smaller than an AI-transformation programme.

Choose one commercially important product family or customer use case.

Map the entities and relationships an informed salesperson or product expert would need to explain it properly. Audit whether that knowledge exists consistently across your website, feeds, FAQs and editorial content. Then map the evidence a skeptical customer would need to validate the recommendation.

Next, identify what you can realistically observe after AI-mediated discovery. Do not wait for perfect attribution. Establish a baseline for AI referrals, branded search, direct traffic, product-page entry behaviour, marketplace movement, customer questions and any available AI citation/mention signals.

Then create an operating rhythm for reviewing what changed.

The objective is not to win a weekly “AI visibility score.”

It is to make the business progressively clearer, more believable and more capable of learning.

The takeaway

AI does not remove the need for brand, customer understanding or strategy.

It raises the value of clarity.

The brands best positioned for AI-native commerce will not simply be the ones that publish the most answer-engine content. They will be the businesses whose products and knowledge can be understood accurately, whose claims can be validated credibly, and whose systems can learn from customer journeys that increasingly begin outside the surfaces they control.

AI may build the shortlist. Proof earns confidence. Customer Intelligence tells you what happened next.

Operator Session

If AI discovery is becoming important to your business but the work is fragmented across SEO, ecommerce, product data, customer experience and analytics, an Operator Session can help pressure-test the system rather than another isolated tactic.

We can map what AI needs to understand, where customer proof is weak, what signals you can realistically observe, and which capability is worth building next.

FAQ opportunities

What does it mean for a business to be understandable to AI? It means the business represents important products, use cases, claims, relationships, evidence and context clearly enough that AI systems do not have to reconstruct the commercial truth from fragmented pages and inconsistent language.

Is AI visibility the same as SEO or GEO? No. Search and answer-engine optimisation can improve discoverability, but durable AI readiness also depends on structured business knowledge, credible evidence and the ability to learn from AI-influenced customer behaviour.

Why does human proof matter if AI already recommends the product? A recommendation reduces search effort but does not remove uncertainty. Customers may still validate claims through reviews, communities, creators, comparisons and other trusted evidence before buying.

How should brands measure AI-influenced commerce? Direct AI referrals are one signal, not the whole journey. Brands should also watch branded search, direct traffic, entry-page behaviour, marketplace/retail movement, customer questions and other downstream changes that may follow AI-mediated discovery.

What is proof architecture? A structured way of connecting important commercial claims with credible supporting evidence, sources, customer experience, context and limitations so both people and machines can evaluate them.