Organic growth in the AI era: should you care?

In the early 2000s, Amazon had a problem it didn't like admitting out loud. A rising share of the traffic that walked through its digital front door wasn't coming because customers typed "amazon.com." It was coming because Google sent them there.

So Amazon did something very Amazon: it measured the problem before it became one. The company built an internal metric called the "Google Reliance metric"¹ to track exactly how dependent Amazon was on Google to drive customers to its door. Nothing about the moment forced this. Google wasn't threatening Amazon at the time. Amazon simply felt uncomfortable with how much of its growth engine sat outside its own control, and treated that discomfort as reason enough to act. Leadership put a strategy in place, watched the number weekly, and kept working the dependency down until it sat at a level they were comfortable with.

Prime. Its own search and recommendation engine. A $31B+ ad business it didn't reveal publicly until 2022. A majority of product searches that now start on Amazon instead of Google. All of it traces back to that early discomfort, addressed deliberately, not to a crisis that forced their hand. That's the instinct worth borrowing: know precisely how reliant you are on someone else's algorithm, and treat the discomfort as reason enough to act, before anyone forces the issue for you.

Most D2C brands are living through a version of that same moment right now. Their organic growth rests on algorithms they don't control and haven't measured. Most just haven't named it yet.

What Google announced at I/O 2026

Google's I/O 2026 announcement was stark. The search box, largely unchanged for 25 years, is now an AI agent. It answers, synthesizes, and increasingly resolves the transaction itself, rather than just pointing you to a website. AI Mode has crossed a billion monthly users, with query volume more than doubling every quarter. Google is also rolling out "information agents" that keep working after you close the tab, and expanding agentic checkout. Search finds the option, compares the price, and hands you a direct link to finish booking with the provider. No ten blue links in between.

Google is no longer indexing pages and ranking them. It's reading the web, synthesizing an answer, and deciding which brand deserves to be in it. ChatGPT, Gemini, and Perplexity are doing the same from different angles. YouTube and Reddit have become two of the loudest trust signals feeding all of them, but they sit alongside review platforms like G2 and Trustpilot, comparison sites, forums, and press coverage. That's the wider web a model checks for corroboration before it decides who to vouch for.

This is the same shift Amazon was bracing for in 2003, except the intermediary isn't a ranking algorithm you can optimize with backlinks. It's a model deciding, in real time, who it trusts enough to vouch for.

But here's where the parallel breaks, usefully. Amazon's answer, back then, was to leave. It built Prime, its own search, its own ad business, its own direct demand, all so it needed Google less. And it never stopped watching that dependency. In July 2025, Amazon went dark on Google Shopping ads worldwide, then switched them back on a month later in every market except the US. Analysts read the exactly-one-month blackout as a controlled test of how much Amazon and its rivals still rely on Google traffic, followed by a market-by-market call on where that reliance was worth paying for. Even Amazon, at Amazon's scale, measures its dependency before it acts on it.

You can't copy the leaving part. You can't build your own ChatGPT, and you can't walk away from the engine that now decides which brand to recommend. So the move isn't escaping the intermediary. It's authoring what the intermediary reads. The owned asset changes shape: no longer a funnel you control end to end, but a reputation deliberately built across the web, consistent enough that whichever engine reads it lands on your version of the story. So the question stops being "how do we escape this?" and becomes "how do we become the source the engine reaches for?" And the answer starts with a word that sounds almost too human for a machine: trust.

You only recommend what you trust

You don't recommend a restaurant, a doctor, or a mattress to a friend unless you trust it. You have to trust it enough to stake your own credibility on it if it goes wrong. That's what a recommendation has always required: your reputation, on the line for someone else's experience.

AI engines are now doing exactly this, at scale, on every category of purchase. When AI Mode, ChatGPT, or Gemini answers "what's the best electrolyte mix for marathon training" or "which baby monitor doesn't leak data," it isn't ranking who paid the most or who has the slickest homepage. It's synthesizing what the web collectively seems to trust and vouch for: forums, reviews, comparison threads, YouTube reviews, Reddit arguments, product pages, press. If your brand doesn't show up consistently and credibly across that web of signals, the model has no reason to recommend you, no matter how good your product actually is.

This is the entire premise Citable is built on: we reverse-engineer how discovery actually works across Google (including AI Overviews and AI Mode), Gemini, ChatGPT, YouTube, and Reddit, and then build one coherent narrative across all of them, working backwards from the specific question or need you want your brand to be the trusted answer to. Not five disconnected campaigns. One story, told consistently enough that the machines reading the web start to vouch for you the way a trusted friend would.

Here's what that looks like in practice, not in theory. Say you sell a protein supplement and want to own the question "is this safe to take daily." That answer doesn't live on your product page. It lives across a Reddit thread where real users compare side effects, a YouTube review that shows the label under a loupe, a comparison article that cites a lab test, and whatever Google's AI Overview decides to stitch together from all three. Citability means making sure your brand's actual position on that question is the one consistently represented, accurately, in every one of those places, so that when five different AI engines independently go looking for an answer, they all land on the same trustworthy version of your story, sourced from five different corners of the web. That's orchestration. It's not something you get from optimizing one channel in isolation.

First, figure out if this is even your problem yet

Before you spend a rupee on this, do what Amazon did: measure your actual dependency. Whether you're the marketing lead, the founder, the CFO, or on the board, ask yourself the same four questions, with numbers, not impressions. How much of the brand's growth is organic. How much of website traffic is organic. How much of YouTube traffic is organic. How much of actual sales can be attributed to organic.

If you don't know, that's the first problem to fix. Not because not knowing is embarrassing, but because you can't judge urgency off a guess. If you do know, and the number is low, don't treat that as a reason to leave it alone. Treat it as headroom: organic is underleveraged for you, and there's more upside sitting there than there is for a brand where it's already close to maxed out. Double down, and get help if you need it.

Because if any real share of discovery is digital (people searching before they buy, comparing on YouTube, checking Reddit before checking out, asking ChatGPT "what should I buy"), this is already happening to you, regardless of what that number turns out to be. The only question is whether you show up in the answer, or whether a competitor does.

Why this quarter, not "eventually"

AI Mode queries are doubling every quarter, per Google's own I/O 2026 disclosure. That's not a slow drift you can watch from the sidelines; it's compounding in real time, actively redirecting volume that used to land on ten blue links into a single AI-synthesized answer.

The bigger reason to move now: trust in these systems compounds fast. Unlike SEO's backlink authority, which took years to build and even longer to unseat, a citation advantage can lock in within weeks. Once a model's sources settle on who it treats as credible for a given question (whose review it cites, whose comparison it references, whose Reddit thread it treats as ground truth), that pattern reinforces itself on the next few retrieval cycles, not the next decade. That's the sharper problem, not a softer one. The window to be first is short. Once a competitor is already the cited answer, you're not filling a vacuum anymore. You're trying to dislodge a source the model already trusts, which is a harder climb than being first.

And your customers aren't waiting for you to be ready. They're asking ChatGPT and AI Mode these questions today. Every day your brand isn't part of that answer is a day a competitor's is.

Be careful who you hire to fix this

A lot of SEO agencies have rebranded overnight as "GEO" agencies. Some are genuinely retooling. Many are relabeling the same old backlink-and-keyword playbook and hoping the acronym does the convincing. Don't hire based on client logos from 2018. This is a materially more technical problem than classic SEO. It involves understanding retrieval, how models weight and cite sources, how citation-worthy content actually gets surfaced in an AI Overview versus a traditional SERP, and how YouTube and Reddit function as trust signals that LLMs treat almost like reference checks. Ask any agency pitching you GEO what signals they're actually optimizing for (branded mentions across trusted third parties, structured content, review platforms) and how they measure whether it's working. If the answer is still backlinks and keyword density, that's your answer. They're not yet equipped to navigate you through this shift.

You don't need Reliance's budget to win this

If you're Reliance, or any brand with a functionally unlimited budget, you can attack every channel simultaneously and brute-force your way to visibility. Citable works fine for you too. But most D2C brands don't have that budget, and they don't need it. What they need is focus. Five specific, correctly sequenced things, nailed in the next two weeks: the handful of moves that actually move the needle on whether AI engines trust and cite you, executed well enough that the compounding starts. Citability compounds the way SEO once did, except faster, and on a quarterly clock instead of a decade-long one.

Amazon didn't wait for Google to become a threat before it acted. It got uncomfortable early, watched the metric weekly, and worked the dependency down while it still had the choice to. You have that same choice right now, except the clock is moving on a quarterly cycle, not a decade-long one. Know where you actually stand, and pick five things to do about it in the next two weeks.

That's the work. That's Citable.

¹ Reported by The Information; cited in Trung Phan's SatPost (Workweek), "Amazon's $31B ad business, explained," Feb 19, 2022.