From 0% to 65% AI visibility in two weeks

Customer: Daylyte - a bootstrapped, two-founder UK hydration brand Category: Vitamin & electrolyte drinks (UK) Timeframe: April - July 2026 Product: HeyOtis AI Search Visibility tracking

The 30-second version

Daylyte is a still-water vitamin and electrolyte drink - no sugar, no artificial sweeteners, none of the chalky texture of the powders. It's run by two founders with no PR agency, no retail listings, and no advertising budget.

By 19 June 2026, when someone asked ChatGPT or Google's Gemini a buying question in their category, Daylyte was mentioned zero times out of 40 tracked questions. They had fallen off the map.

Two weeks later, Daylyte was the single most-recommended brand in its entire category - named in 65% of ChatGPT answers, ahead of established players like VITHIT.

They did it with free, founder-led work - no ad spend, no agency - guided by exactly what HeyOtis's tracking showed them to fix.

At a glance

ChatGPT visibility is the headline figure because it's the engine that surfaces this category most; pooled visibility blends ChatGPT and Gemini (see "The honest bit" below).

The challenge: a slow slide to zero

HeyOtis tracks Daylyte's "AI visibility" - how often the brand appears when real buyers ask ChatGPT and Gemini questions like "non-chalky vitamin drinks with no sugar" or "buy sugar-free electrolyte drinks for workouts." We ask 40 of these buyer questions across both engines on a repeating schedule, and record which brands get named, where they rank, and which web pages the AI cited as its sources.

Between late April and mid-June, Daylyte's number went one direction - down:

This wasn't a sudden ban or a broken website. It was a quiet, steady disappearance - the kind you can't see happening without measurement.

The diagnosis: one point of failure

Because HeyOtis records the sources behind every AI answer, we could see exactly why. The finding was stark:

Across every run, the only web page the AI ever cited that mentioned Daylyte was Daylyte's own site - not a single shop listing, review, article, or forum post. Their entire presence in AI answers balanced on a single page winning a single search.

Meanwhile the brands holding their ground were cited from 5-10 different websites each. As the category got more crowded through spring, Daylyte's one page kept losing the ranking contest - until the AI stopped surfacing it at all.

We also ruled out the scary explanations: the site wasn't down, wasn't blocking AI crawlers, and was properly indexed. Their store was healthy and unusually well-built. The problem was simply discoverability - and that's fixable with content the founders controlled.

The prescription was deliberately founder-sized and free:

1. Give the AI more of your own pages to find - turn one citable page into ten, each shaped like a question customers actually ask.
2. Get them indexed - submit to Bing (which powers ChatGPT's search) and Google, so the new pages can actually be seen.
3. Earn a few outside mentions over time - reviews, Reddit, cheap shop listings.

What the founders did

They executed the highest-leverage move fast. On 21-22 June they published nine articles on their own site - comparison pages (Daylyte vs VitHit, vs Vitamin Well, vs Phizz) and niche guides (the office fridge, daily hydration, the 3pm slump, low-sodium) - each written answer-first, the way AI search likes to read a page, with clean structured data behind it. Then they got them indexed.

That's it. No agency. No ad budget. Content they wrote, aimed precisely at the questions HeyOtis showed were winnable.

The result: invisible to #1

The next measurement runs told the story:

On 2 July, Daylyte wasn't just back - it was the most-visible brand in the whole category, ahead of its nearest rivals:

And the quality of those mentions was high: an average rank of 2.4, a top-3 placement in 55% of tracked questions, and a #1 spot in nine questions - including both the niche they're genuinely best at ("non-chalky vitamin drinks with no sugar") and generic head terms ("compare prices for zero-sugar electrolyte drinks") they were never expected to reclaim. When the AI named them, it was positive 88% of the time.

The mechanism was exactly what the diagnosis predicted. The number of times AI answers cited Daylyte's own site went from 1 in mid-June to 27 on 2 July and 53 on 3 July - the nine new articles doing precisely the job they were built for.

The honest bit (why this is repeatable, not luck)

Good case studies don't oversell, so two caveats that actually make Daylyte's win more credible, not less:

• This was a content-and-indexing win, not a PR win. Daylyte's recovery came almost entirely from getting their own pages found. The next chapter - earning third-party mentions (reviews, Reddit, shop listings) - is still ahead of them. There's more upside in the tank.
• It's mostly a ChatGPT story so far. Google's Gemini is only just starting to surface Daylyte (5-7.5%). That's the next engine to crack, and the same playbook applies.

In other words: a single, focused, free move - publishing question-shaped pages and getting them indexed - took a brand from invisible to category-leading in two weeks. And they've still only played the first move.

What marketers can take from this

• You can't fix what you can't see. Daylyte's decline was invisible until it was measured. AI visibility is now a channel you can track like any other.
• AI answers are earned with content, not ad budget. The single highest-leverage move was writing pages that answer real buyer questions - then making sure Bing and Google could find them.
• Answer the question, don't just describe the product. The pages that won were shaped like the questions buyers ask an AI, with the answer up top.
• Being cited from one place is fragile; the goal is to be corroborated in many. Daylyte's next gains will come from outside mentions - and so will yours.

"We could see Daylyte's presence in AI answers balancing on a single page. The moment that page lost a search, they vanished. The fix wasn't budget - it was giving the AI more of their own pages to find, and making sure they got indexed. Two weeks after they did it, they were the #1 brand in their category."

• HeyOtis

Methodology

HeyOtis measures AI visibility by asking two AI engines - ChatGPT (GPT-5) and Google Gemini - a fixed set of 40 real buyer questions for the brand's category, on a repeating schedule (11 runs between 24 April and 3 July 2026). For every answer we record which brands were named, at what rank, and which web pages the AI cited as its sources. Visibility % is the share of question-and-engine opportunities in which the brand appeared. "ChatGPT visibility" isolates the GPT-5 engine; "pooled visibility" blends both engines across all 80 opportunities per run. All figures in this case study are drawn directly from Daylyte's live measurement data.

Published with Daylyte's permission. Figures current as of 3 July 2026.