AI is breaking your attribution – here’s what’s happening

Marketing used to have a paper trail. AI is erasing it.

Key takeaway: AI is breaking your attribution by hiding the middle of the customer journey inside AI conversations, enabling purchases without website visits, and blending paid placements into AI-generated answers in ways traditional analytics cannot track.

For twenty years, digital marketing ran on one basic assumption: you could trace a sale back to a click.

Someone searched Google, landed on your site, read a few pages, came back later, and bought. Analytics tracked every step. Attribution was never perfect, but it was good enough to justify a marketing spend.

AI has broken that model in three separate ways, and most businesses haven’t noticed yet because the sales are still coming in. They just don’t know why anymore.

• The middle of the AI-driven customer journey has gone dark.
• Agentic commerce could remove the journey altogether.
• AI advertising is creating a new attribution problem.
• Why this matters for your business.

The middle of the AI-driven customer journey has gone dark.

The old customer journey looked something like this:

1. Search Google
2. Click your website
3. Read a few pages
4. Come back later
5. Buy

Google Analytics could see most of that. Every click, every return visit, every conversion, all logged.

The AI-driven version looks like this:

1. Someone asks ChatGPT or Gemini a question
2. The AI researches multiple websites
3. It summarises what it finds
4. The user asks a follow-up question
5. The AI recommends one or two businesses
6. The user visits one website, or doesn’t visit any website at all

Everything between the first question and the final click is invisible to you.

Say someone asks: “Who are the best WordPress agencies for charities in Oxfordshire?”

ChatGPT might weigh up twenty agencies before it mentions one. You have no way of knowing whether your business was considered, whether it was ruled out, which competitors got a look, or why the AI picked someone else instead.

That whole middle section of the journey has vanished from your analytics. It’s the biggest single reason traditional attribution models are starting to fall apart.

This is also why most SEO content fails to convert. Content written purely for search engine rankings often misses the mark when AI systems are evaluating it for recommendations. The AI isn’t just looking for keywords. It’s looking for genuine expertise and clear answers.

Agentic commerce could remove the journey altogether.

Agentic AI doesn’t just answer questions. It carries out tasks on your behalf.

Ask an agent to “book me the cheapest four-star hotel in York next Friday” and it won’t hand you a list of options. It will compare hundreds of hotels, check the reviews, book the room, and pay with your saved card.

You never visit Booking.com. You never visit the hotel’s website. You never click an advert. The AI has become the customer.

According to Gartner’s research, agentic AI is expected to transform how business transactions occur, with AI agents handling increasingly complex purchasing decisions autonomously.

The same shift is coming for insurance, office supplies, software renewals, restaurant bookings, and plenty of other business services.

For businesses, this means the customer journey doesn’t just get harder to track. It might stop existing in any form you can measure. Instead of persuading a person to choose you, you’ll need to persuade the AI agent making the decision for them.

Consider what this means for your marketing:

• Your website might never receive a visit from the person who buys from you
• Your analytics will show zero traffic from transactions that actually happened
• Your conversion tracking becomes meaningless when the AI is the one converting
• Your remarketing lists won’t capture people who never touched your site

The entire measurement infrastructure that marketing has relied on for two decades assumes a human visits your website at some point. Agentic commerce breaks that assumption completely.

AI advertising is creating a new attribution problem.

AI platforms are starting to fold advertising into their answers: sponsored recommendations, promoted products, and paid placements sitting inside the response itself.

Someone asks: “Best project management software for agencies.” The AI comes back with ClickUp, Monday.com, and Asana. Maybe one of those three paid to be there.

If the user buys the software a week later, where did that sale actually come from? The AI recommendation? An old Google search? A LinkedIn ad they saw last month? A podcast they half-listened to? Or did they already know the brand before any of this happened?

Working out which marketing activity earned the credit is getting harder by the month. Last-click attribution, which was already a rough approximation of reality, no longer reflects how people actually make decisions.

The BBC has reported on how AI companies are exploring advertising models, and the implications for how brands compete for visibility inside AI-generated responses are significant.

This creates a cascading problem for marketing budgets. If you can’t attribute sales to specific channels, you can’t optimise spend. If you can’t optimise spend, you either waste money on channels that aren’t working or cut funding to channels that are.

Many businesses are already experiencing this. Their sales figures look healthy, but their attribution data makes no sense. Traffic is flat or down, yet revenue keeps growing. The explanation is often that AI-influenced decisions are driving purchases through paths that analytics simply cannot see.

Why this matters for your business.

All three of these trends point in the same direction.

In the old world, businesses optimised for search engines, customers visited websites, analytics tracked the clicks and conversions, and attribution was imperfect but understood well enough to work with.

In the AI world, customers ask AI before they ask Google, AI evaluates businesses before a human ever sees them, purchases can complete without a website visit, and most of the decision-making happens somewhere you can’t see.

This is exactly why AI visibility matters as its own discipline, separate from SEO. Ranking first on Google still counts, but it’s no longer the whole game. The bigger question is whether AI systems trust your business enough to cite it, recommend it, or act on it without a human checking your website at all.

Understanding how AI visibility works is becoming essential for any business that depends on being found online. The framework is different from traditional SEO because the evaluation criteria are different. AI systems assess expertise, trustworthiness, and clarity in ways that don’t map neatly onto keyword rankings.

The businesses that will thrive in this environment are the ones that recognise a fundamental shift is underway. Traffic numbers and click-through rates will become less reliable as success metrics. What matters instead is whether your business appears in AI-generated answers when potential customers ask relevant questions.

This also explains why SEO traffic without conversion alignment is wasted spend. If your content attracts clicks but doesn’t position your business as the answer to a specific problem, AI systems have no reason to recommend you. And increasingly, the AI recommendation is what leads to the sale.

The job isn’t just generating clicks anymore. It’s influencing decisions that happen before a website visit, or instead of one entirely.

If you want to know whether your business is showing up in AI answers, or find out why it isn’t, get in touch with Searchworthy.