Winning AI Search Before Our Competitors Knew It Was a Channel

REDX sells seller leads to real estate agents. It's a crowded category: Vulcan7, Mojo, Landvoice, and a long tail of smaller players all compete for the same searches. Most of them had been publishing content for years.

When I joined, the site had a familiar shape: a broad ranking profile that converted narrowly, with 2,707 keywords stranded past position 51 and 228 pages splitting the traffic between them. We ranked position 63 for "real estate leads", the category head term, at a $6.18 cost per click.

The bigger problem was one nobody in the category was measuring yet. Agents were starting to ask ChatGPT which lead platform to use. Nobody knew who those answers named, because nobody was tracking it.

1. Optimize for traffic value, not traffic

I stopped reporting sessions as the headline metric and switched to estimated traffic value: what the organic traffic would cost to buy. It changed every prioritization decision. A 450-search-per-month keyword like "motivated seller leads" at an $8.23 CPC is worth pursuing ahead of a high-volume informational term with no commercial intent.

2. Consolidate instead of publishing more

Most content programs answer underperformance by publishing more. I did the opposite, consolidating overlapping pages and cutting content that ranked nowhere and served nobody. Pages generating organic traffic came down from 228 to 141.

3. Build the measurement layer for AI search first

Before optimizing anything for AI answers, I built the scoreboard: competitive share-of-voice tracking across ChatGPT, Gemini, Perplexity, and Google AI Mode, benchmarked against four named competitors. When platform-level LLM referral classification became available, I got it isolated in analytics so ChatGPT-sourced visits stopped hiding inside "direct."

You can't optimize a channel you can't see. Building the scoreboard first is why the results below are provable rather than anecdotal.

4. Systematize production

I documented the whole pipeline (topic research through publication) as an SOP, then automated the handoffs. Production time dropped 75%. That let the team execute the consolidation strategy at a pace that would otherwise have required more headcount.

AI answers now name us first

On non-branded category prompts, questions like "best real estate lead generation software," where the user hasn't mentioned any brand, REDX holds 86.9% share of voice across ChatGPT, Gemini, Perplexity, and Google AI Mode. The nearest competitor holds 42.9%.

More telling than the share is the average mention position of 1.33. When an AI answer lists lead generation platforms, REDX is almost always the first one named. Across the measurement window we recorded 410 citations across 1,423 AI answers, a 25.1% citation rate.

Citations nearly tripled in six months

AI visibility compounds as more of the site earns citations. Total citations across seven AI platforms went from 55 to 154 in six months, including new coverage in Google AI Overviews, Google AI Mode, and Grok where REDX previously had none. ChatGPT now cites 30 distinct REDX pages, up from 19. That is site-wide authority, not one lucky article.

ChatGPT became a measurable acquisition channel

Isolating LLM referrals surfaced 579 visitors from ChatGPT in the trailing 12 months, traffic that had previously been invisible inside direct. The quality signal is the interesting part:

• 52.7% bounce rate for ChatGPT-referred visitors vs. 86.0% for direct, the best of any external source
• 318-second average session vs. 159 seconds for direct
• 1,092 total visitors across all LLM sources

Rankings improved sharply, and the clicks went somewhere interesting

Average Search Console position improved 13.8 places, from 21.8 to 7.9, in six months. Over the same period, clicks stayed roughly flat at 8,200–8,700 per month.

That gap is the story of SEO in 2026. When you win the ranking but not the click, the click has usually been absorbed by an AI Overview. Because I'd built AI citation and share-of-voice tracking, I could show where the demand actually went: AI citations up 180% and 86.9% share of voice over the same window.

Across the full profile:

• Top-3 rankings tripled: 209 → 656 keywords (+214%)
• Top-10 rankings tripled: 493 → 1,485 keywords (+201%)
• Referring domains: 568 → 1,685 (+197%); Domain Rating 52 → 58
• Traffic per ranking page: +104%, on 38% fewer pages
• Annual organic traffic value: $263,978 → $597,089 (+126%), with a monthly peak of $71,267

That last figure is the one I'd put in front of a CFO. At roughly $597,000 a year, the organic channel is doing the work of a substantial paid search budget. Unlike paid, it keeps working when the spend stops.