Team ZoldFour months ago, no AI assistant had ever sent Team Zold a visitor. Now it's 9% of traffic.
Team Zold, a real estate team in a crowded, locally competitive market, started AI Engine Optimization with Detectabli on April 2, 2026. This is the whole engagement so far, measured month by calendar month: how search visibility grew 4.5×, how an AI referral channel appeared and kept climbing, and why the traffic that arrives now is worth far more than the traffic that left.
- 01
Team Zold received zero AI assistant sessions in April and May 2026, then 23 in June, 38 in July, and 13 in the first three days of August — 74 in total.
- 02
In that early-August window, AI assistants accounted for 9.4% of all site traffic, at 4.3 sessions per day against 1.2 per day in July.
- 03
Search impressions grew 345.1% comparing the last three months to the previous three (21.9K → 97.7K), driven by 83 data-driven pages carrying structured schema markup.
- 04
Site engagement rate rose from 5.4% in April to 25.0% in July and 31.2% in early August, while average engagement time went from 0s to 31s per session.
- 05
The Detectabli LLM dashboard recorded 100 referral sessions from 92 users across ChatGPT (67), Perplexity (30), and Claude (3) between May 6 and August 3, 2026.
- Client
- Team Zold — a real estate team in a crowded, locally competitive market.
- Engagement
- Four months of AI Engine Optimization with Detectabli, starting April 2, 2026.
- Measurement
- Month by calendar month, April 2 – August 3, 2026. No modelled or projected figures.
- Sources
- Google Search Console (search visibility), GA4 (engagement and channel data), and the Detectabli LLM referral dashboard (assistant-level referrals).
The AI channel, month by month.
The hatched bar is a partial month: 13 AI sessions in the first three days of August — a 4.3-per-day pace against 1.2/day in July, and 9.4% of all site traffic in that window. AI visitors in that window engaged at 46.2% and fired 8.3 events per session, double the site average.
The pattern matters more than any single number. Nothing in April, nothing in May while the content foundation went in, then a working channel in June that has grown every period since.
Assistants don't cite a site the day something publishes; they cite it once the material has been crawled, indexed, and judged reliable. That lag is exactly what these bars show, and it's why the channel is still accelerating.
How the visibility was built.
Since April 2, Detectabli has produced 83 pieces of data-driven content that now rank in Google: monthly market updates built from TRREB board data, in-depth neighbourhood guides, and head-to-head town comparisons, each carrying structured schema markup so search engines and AI assistants can parse and cite the numbers.
Those 83 pages have earned 45.9K impressions — 41% of everything the site received — and the flagship Richmond Hill market update, published April 8, is now the most-seen page on the site: 15,162 impressions at position 5.5, against 2,187 for the homepage.
110.7K impressions and 1,252 clicks since April 2. June peaked as spring content hit Google's AI Overviews.
Nearly 6× better than the April baseline. Average engagement time per session climbed alongside it: 0s → 7s → 10s → 14s → 31s.
Quality replaced volume.
Total sessions fell over the engagement — and that is the zero-click era working as designed. April's traffic was inflated by thousands of near-zero-engagement Direct sessions: 4,363 of them, engaging at 1.6%, effectively bots and unattributed noise. That noise washed out in early May.
What replaced it is smaller and dramatically better. Engaged sessions rose from 264 in April to 475 in July, up 80%, while raw sessions shrank — and every earned channel (Organic Search, Organic Social, AI Assistant) grew month over month.
Fewer sessions, more of them real. April's 264 sit on top of a traffic base several times larger than July's 475.
Start to now, metric by metric.
| Metric | Baseline | Latest | Change |
|---|---|---|---|
| Search impressions (3 mo vs prior 3 mo) | 21.9K | 97.7K | +345.1% |
| Search clicks (3 mo vs prior 3 mo) | 323 | 945 | +192.6% |
| Average CTR (3 mo vs prior 3 mo) | 1.47% | 0.97% | -34.0% |
| Average position (3 mo vs prior 3 mo) | 8.29 | 10.22 | -1.93 |
| Engagement rate (April vs July) | 5.4% | 25.0% | +19.6 pts |
| Engaged sessions (April vs July) | 264 | 475 | +79.9% |
| Avg. engagement time / session (April vs July) | 0s | 14s | +14s |
| Avg. engagement time / session (April vs Aug 1–3) | 0s | 31s | +31s |
| AI assistant sessions / month (April vs July) | 0 | 38 | New |
| AI assistant sessions (Apr 2 – Aug 3, total) | 0 | 74 | New |
| LLM referral sessions (May 6 – Aug 3) | 0 | 100 | New |
Comparison windows differ by metric and are stated in each row label. Search figures from Google Search Console; session and engagement figures from GA4; LLM referral sessions from the Detectabli LLM dashboard.
Every channel, full engagement.
| Channel | Sessions | Engaged | Eng. rate | Avg. eng. time | Events / session |
|---|---|---|---|---|---|
| Direct | 8,571 | 537 | 6.27% | 2s | 3.17 |
| Organic Search | 1,508 | 735 | 48.74% | 24s | 4.76 |
| Organic Social | 275 | 54 | 19.64% | 3s | 3.45 |
| Referral | 199 | 121 | 60.80% | 16s | 6.59 |
| AI Assistant | 74 | 27 | 36.49% | 31s | 6.69 |
| Site total | 10,518 | 1,468 | 13.96% | 6s | 3.53 |
GA4, April 2 – August 3, 2026. Unassigned (30 sessions) omitted. AI Assistant visitors spend 5× longer actively engaged than the site average and fire nearly twice the events per session — the highest-intent traffic on the site after Referral.
Which assistants send the traffic.
| Source | Sessions | Users | Eng. rate | Avg. engagement time |
|---|---|---|---|---|
| ChatGPT (chatgpt.com) | 67 | 64 | 26.9% | 1m 13s |
| Perplexity (perplexity.ai) | 30 | 26 | 63.3% | 12m 25s |
| Claude (claude.ai) | 3 | 2 | 33.3% | 46s |
Detectabli LLM referral dashboard: 100 LLM referral sessions from 92 users, measured at the source level — a wider net than GA4's channel grouping. A Perplexity referral averages twelve and a half minutes on site: someone mid-decision, not someone browsing.
Reading the numbers in a zero-click world.
Two things on a classic SEO dashboard point down this period: blended CTR (1.47% → 0.97%) and average position (8.29 → 10.22). Both are the expected signature of expansion, not decline.
In April, Team Zold appeared almost exclusively for its own name — queries it ranked #1 for, with high CTR and nowhere to grow. Today it surfaces across thousands of competitive informational queries in its market, where Google's AI Overviews answer at the top of the page and absorb most clicks. Impressions grew 4.5×, absolute clicks still nearly tripled (323 → 945), and the brand is now in the source material assistants draw from.
The bottom line: In four months Team Zold went from a site only its own clients searched for to one that Google shows 30K+ times a month, that ChatGPT and Perplexity cite and send visitors to, and whose engagement rate has climbed every month of the engagement. The sessions that disappeared were the ones nobody would want. The ones that replaced them read guides for twelve minutes and come back.
Working with Gabe and Lucas at Detectabli has been an outstanding experience. Coming into an industry I knew nothing about, their depth of product knowledge and ability to translate complex concepts into clear, actionable guidance made all the difference. They took the time to understand our goals, built a strategy around them, and executed with a level of conscientiousness and work ethic that's genuinely rare.
Since partnering with them, our views and engagement have noticeably increased, and we've already received messages from new clients directly tied to the work they've put in. Their response time is exceptional, their communication is consistent, and their commitment to delivering results shows up every single day.
Detectabli has been a true value-add to our business, and Gabe and Lucas have earned every bit of this recommendation.
How long did it take before AI assistants started sending Team Zold traffic?
Two months. Team Zold recorded zero AI assistant sessions in April and May 2026, the period in which the content foundation was built. The first assistant referrals landed in June (23 sessions), followed by 38 in July and 13 in the first three days of August. Assistants cite a site once its material has been crawled, indexed, and judged reliable — not on the day it publishes.
How much of Team Zold's traffic now comes from AI assistants?
74 AI Assistant sessions between April 2 and August 3, 2026. In the first three days of August, AI assistants accounted for 9.4% of all site traffic, at a pace of 4.3 sessions per day against 1.2 per day in July.
Which AI assistants cite and refer to Team Zold?
ChatGPT, Perplexity, and Claude. Between May 6 and August 3, 2026, the Detectabli LLM referral dashboard recorded 100 referral sessions from 92 users: 67 sessions from ChatGPT, 30 from Perplexity, and 3 from Claude.
Is AI assistant traffic actually worth having?
On this site it is the highest-intent traffic after Referral. AI Assistant sessions engaged at 36.49% against 13.96% site-wide, averaged 31 seconds of engagement against 6 seconds site-wide, and fired 6.69 events per session against 3.53. Perplexity referrals averaged 12 minutes 25 seconds on site.
Why did Team Zold's click-through rate and average position get worse?
Because the site expanded beyond its own brand terms. Blended CTR fell from 1.47% to 0.97% and average position moved from 8.29 to 10.22, but impressions grew 4.5× and absolute clicks nearly tripled from 323 to 945. In April the site ranked #1 almost exclusively for its own name; today it surfaces across thousands of competitive informational queries where AI Overviews answer at the top of the page. Falling CTR and position are the arithmetic of ranking for far more queries, not a decline in performance.
What did Detectabli actually do for Team Zold?
Detectabli produced 83 pieces of data-driven content that now rank in Google — monthly market updates built from TRREB board data, in-depth neighbourhood guides, and head-to-head town comparisons — each carrying structured schema markup so search engines and AI assistants can parse and cite the numbers. Those 83 pages earned 45.9K impressions, 41% of everything the site received.
- Google Search Console — April 2 to August 2, 2026
- Google Analytics 4 — April 2 to August 3, 2026
- Detectabli LLM referral dashboard — May 6 to August 3, 2026
Figures are reported as measured, with no modelling or projection. Comparison windows are stated wherever two periods are compared.
Wanttobethenextstory?
If you're using Detectabli to drive real outcomes, we'd love to write it up. Reach out and we'll handle the heavy lifting.