Share of Agent, week of Oct 19
24%
n=600 · 95% range 20–27%
Too small to call vs Oct 12One reading a week, with the range the true number likely sits in. Where a line restarts, what we measure changed. See the answers behind it.
Nothing moved beyond the usual week-to-week wobble. That is a normal week.
Questions won, weekly from Aug 17 to Oct 19. Latest, week of Oct 19: 17%, n=30 · 95% range 7–34%. Too small to call vs Oct 12. 2 changes in how it is measured, so the line restarts: Aug 31, Oct 5. The table has every week.
| Week of | Reading | 95% range | n | Change |
|---|---|---|---|---|
| Aug 17 | 10% | 3–26% | 30 | First reading |
| Aug 24 | 10% | 3–26% | 30 | Too small to call vs Aug 17 |
| Aug 31 · line restarts | 3% | 1–17% | 30 | Not comparable: the questions changed |
| Sep 7 | 10% | 3–26% | 30 | Too small to call vs Aug 31 |
| Sep 14 | 10% | 3–26% | 30 | Too small to call vs Sep 7 |
| Sep 21 · model change | 13% | 5–30% | 30 | Not comparable: an assistant changed model |
| Sep 28 | 10% | 3–26% | 30 | Too small to call vs Sep 21 |
| Oct 5 · line restarts | 13% | 5–30% | 30 | Not comparable: measurement changed |
| Oct 12 | 13% | 5–30% | 30 | Too small to call vs Oct 5 |
| Oct 19 | 17% | 7–34% | 30 | Too small to call vs Oct 12 |
AI agent visits, weekly from Aug 17 to Oct 19. Latest, week of Oct 19: 156, n=156 · 95% range 132–183. Too small to call vs Oct 12. One change in how it is measured, so the line restarts: Oct 5. Weeks of Aug 17 to Aug 24: no reading these weeks. The table has every week.
| Week of | Reading | 95% range | n | Change |
|---|---|---|---|---|
| Aug 17 | — No reading this week | — | — | — |
| Aug 24 | — No reading this week | — | — | — |
| Aug 31 | 63 | 48–81 | 63 | First reading |
| Sep 7 | 95 | 76–117 | 95 | Too small to call vs Aug 31 |
| Sep 14 | 121 | 100–145 | 121 | Too small to call vs Sep 7 |
| Sep 21 | 143 | 120–169 | 143 | Too small to call vs Sep 14 |
| Sep 28 | 173 | 148–201 | 173 | Too small to call vs Sep 21 |
| Oct 5 · line restarts | 160 | 136–187 | 160 | Not comparable: measurement changed |
| Oct 12 | 170 | 145–198 | 170 | Too small to call vs Oct 5 |
| Oct 19 | 156 | 132–183 | 156 | Too small to call vs Oct 12 |
AI agent visits with a verified identity, weekly from Aug 17 to Oct 19. Latest, week of Oct 19: 133, n=133 · 95% range 111–158. Too small to call vs Oct 12. One change in how it is measured, so the line restarts: Oct 5. Weeks of Aug 17 to Aug 24: no reading these weeks. The table has every week.
| Week of | Reading | 95% range | n | Change |
|---|---|---|---|---|
| Aug 17 | — No reading this week | — | — | — |
| Aug 24 | — No reading this week | — | — | — |
| Aug 31 | 44 | 31–60 | 44 | First reading |
| Sep 7 | 59 | 44–77 | 59 | Too small to call vs Aug 31 |
| Sep 14 | 83 | 66–103 | 83 | Too small to call vs Sep 7 |
| Sep 21 | 93 | 75–114 | 93 | Too small to call vs Sep 14 |
| Sep 28 | 114 | 94–137 | 114 | Too small to call vs Sep 21 |
| Oct 5 · line restarts | 138 | 115–164 | 138 | Not comparable: measurement changed |
| Oct 12 | 147 | 124–173 | 147 | Too small to call vs Oct 5 |
| Oct 19 | 133 | 111–158 | 133 | Too small to call vs Oct 12 |
OpenAI agent visits, weekly from Aug 17 to Oct 19. Latest, week of Oct 19: 55, n=55 · 95% range 41–72. Too small to call vs Oct 12. One change in how it is measured, so the line restarts: Oct 5. Weeks of Aug 17 to Aug 24: no reading these weeks. The table has every week.
| Week of | Reading | 95% range | n | Change |
|---|---|---|---|---|
| Aug 17 | — No reading this week | — | — | — |
| Aug 24 | — No reading this week | — | — | — |
| Aug 31 | 20 | 12–31 | 20 | First reading |
| Sep 7 | 27 | 17–40 | 27 | Too small to call vs Aug 31 |
| Sep 14 | 39 | 27–54 | 39 | Too small to call vs Sep 7 |
| Sep 21 | 35 | 24–49 | 35 | Too small to call vs Sep 14 |
| Sep 28 | 49 | 36–65 | 49 | Too small to call vs Sep 21 |
| Oct 5 · line restarts | 55 | 41–72 | 55 | Not comparable: measurement changed |
| Oct 12 | 62 | 47–80 | 62 | Too small to call vs Oct 5 |
| Oct 19 | 55 | 41–72 | 55 | Too small to call vs Oct 12 |
Anthropic agent visits, weekly from Aug 17 to Oct 19. Latest, week of Oct 19: 33, n=33 · 95% range 22–47. Too small to call vs Oct 12. One change in how it is measured, so the line restarts: Oct 5. Weeks of Aug 17 to Aug 24: no reading these weeks. The table has every week.
| Week of | Reading | 95% range | n | Change |
|---|---|---|---|---|
| Aug 17 | — No reading this week | — | — | — |
| Aug 24 | — No reading this week | — | — | — |
| Aug 31 | 20 | 12–31 | 20 | First reading |
| Sep 7 | 27 | 17–40 | 27 | Too small to call vs Aug 31 |
| Sep 14 | 36 | 25–50 | 36 | Too small to call vs Sep 7 |
| Sep 21 | 42 | 30–57 | 42 | Too small to call vs Sep 14 |
| Sep 28 | 47 | 34–63 | 47 | Too small to call vs Sep 21 |
| Oct 5 · line restarts | 32 | 21–46 | 32 | Not comparable: measurement changed |
| Oct 12 | 33 | 22–47 | 33 | Too small to call vs Oct 5 |
| Oct 19 | 33 | 22–47 | 33 | Too small to call vs Oct 12 |
Other agent visits, weekly from Aug 17 to Oct 19. Latest, week of Oct 19: 18, n=18 · 95% range 10–29. Too small to call vs Oct 12. One change in how it is measured, so the line restarts: Oct 5. Weeks of Aug 17 to Aug 24: no reading these weeks. The table has every week.
| Week of | Reading | 95% range | n | Change |
|---|---|---|---|---|
| Aug 17 | — No reading this week | — | — | — |
| Aug 24 | — No reading this week | — | — | — |
| Aug 31 | 7 | 2–15 | 7 | First reading |
| Sep 7 | 24 | 15–36 | 24 | Up 17 vs Aug 31 |
| Sep 14 | 21 | 12–33 | 21 | Too small to call vs Sep 7 |
| Sep 21 | 31 | 21–45 | 31 | Too small to call vs Sep 14 |
| Sep 28 | 38 | 26–53 | 38 | Too small to call vs Sep 21 |
| Oct 5 · line restarts | 24 | 15–36 | 24 | Not comparable: measurement changed |
| Oct 12 | 25 | 16–37 | 25 | Too small to call vs Oct 5 |
| Oct 19 | 18 | 10–29 | 18 | Too small to call vs Oct 12 |
Perplexity agent visits, weekly from Aug 17 to Oct 19. Latest, week of Oct 19: 22, n=22 · 95% range 13–34. Too small to call vs Oct 12. One change in how it is measured, so the line restarts: Oct 5. Weeks of Aug 17 to Aug 24: no reading these weeks. The table has every week.
| Week of | Reading | 95% range | n | Change |
|---|---|---|---|---|
| Aug 17 | — No reading this week | — | — | — |
| Aug 24 | — No reading this week | — | — | — |
| Aug 31 | 7 | 2–15 | 7 | First reading |
| Sep 7 | 7 | 2–15 | 7 | Too small to call vs Aug 31 |
| Sep 14 | 12 | 6–21 | 12 | Too small to call vs Sep 7 |
| Sep 21 | 16 | 9–26 | 16 | Too small to call vs Sep 14 |
| Sep 28 | 17 | 9–28 | 17 | Too small to call vs Sep 21 |
| Oct 5 · line restarts | 21 | 12–33 | 21 | Not comparable: measurement changed |
| Oct 12 | 22 | 13–34 | 22 | Too small to call vs Oct 5 |
| Oct 19 | 22 | 13–34 | 22 | Too small to call vs Oct 12 |
Google agent visits, weekly from Aug 17 to Oct 19. Latest, week of Oct 19: 14, n=14 · 95% range 7–24. Too small to call vs Oct 12. One change in how it is measured, so the line restarts: Oct 5. Weeks of Aug 17 to Aug 24: no reading these weeks. The table has every week.
| Week of | Reading | 95% range | n | Change |
|---|---|---|---|---|
| Aug 17 | — No reading this week | — | — | — |
| Aug 24 | — No reading this week | — | — | — |
| Aug 31 | 5 | 1–12 | 5 | First reading |
| Sep 7 | 4 | 1–11 | 4 | Too small to call vs Aug 31 |
| Sep 14 | 7 | 2–15 | 7 | Too small to call vs Sep 7 |
| Sep 21 | 10 | 4–19 | 10 | Too small to call vs Sep 14 |
| Sep 28 | 9 | 4–18 | 9 | Too small to call vs Sep 21 |
| Oct 5 · line restarts | 14 | 7–24 | 14 | Not comparable: measurement changed |
| Oct 12 | 14 | 7–24 | 14 | Too small to call vs Oct 5 |
| Oct 19 | 14 | 7–24 | 14 | Too small to call vs Oct 12 |
Apple agent visits, weekly from Aug 17 to Oct 19. Latest, week of Oct 19: 6, n=6 · 95% range 2–14. Too small to call vs Oct 12. One change in how it is measured, so the line restarts: Oct 5. Weeks of Aug 17 to Aug 24: no reading these weeks. The table has every week.
| Week of | Reading | 95% range | n | Change |
|---|---|---|---|---|
| Aug 17 | — No reading this week | — | — | — |
| Aug 24 | — No reading this week | — | — | — |
| Aug 31 | 2 | 0–8 | 2 | First reading |
| Sep 7 | 3 | 0–9 | 3 | Too small to call vs Aug 31 |
| Sep 14 | 3 | 0–9 | 3 | Too small to call vs Sep 7 |
| Sep 21 | 6 | 2–14 | 6 | Too small to call vs Sep 14 |
| Sep 28 | 5 | 1–12 | 5 | Too small to call vs Sep 21 |
| Oct 5 · line restarts | 6 | 2–14 | 6 | Not comparable: measurement changed |
| Oct 12 | 6 | 2–14 | 6 | Too small to call vs Oct 5 |
| Oct 19 | 6 | 2–14 | 6 | Too small to call vs Oct 12 |
Microsoft agent visits, weekly from Aug 17 to Oct 19. Latest, week of Oct 19: 6, n=6 · 95% range 2–14. Too small to call vs Oct 12. One change in how it is measured, so the line restarts: Oct 5. Weeks of Aug 17 to Aug 24: no reading these weeks. The table has every week.
| Week of | Reading | 95% range | n | Change |
|---|---|---|---|---|
| Aug 17 | — No reading this week | — | — | — |
| Aug 24 | — No reading this week | — | — | — |
| Aug 31 | 1 | 0–6 | 1 | First reading |
| Sep 7 | 2 | 0–8 | 2 | Too small to call vs Aug 31 |
| Sep 14 | 2 | 0–8 | 2 | Too small to call vs Sep 7 |
| Sep 21 | 2 | 0–8 | 2 | Too small to call vs Sep 14 |
| Sep 28 | 6 | 2–14 | 6 | Too small to call vs Sep 21 |
| Oct 5 · line restarts | 6 | 2–14 | 6 | Not comparable: measurement changed |
| Oct 12 | 6 | 2–14 | 6 | Too small to call vs Oct 5 |
| Oct 19 | 6 | 2–14 | 6 | Too small to call vs Oct 12 |
Meta agent visits, weekly from Aug 17 to Oct 19. Latest, week of Oct 19: 2, n=2 · 95% range 0–8. Too small to call vs Oct 12. One change in how it is measured, so the line restarts: Oct 5. Weeks of Aug 17 to Aug 24: no reading these weeks. The table has every week.
| Week of | Reading | 95% range | n | Change |
|---|---|---|---|---|
| Aug 17 | — No reading this week | — | — | — |
| Aug 24 | — No reading this week | — | — | — |
| Aug 31 | 1 | 0–6 | 1 | First reading |
| Sep 7 | 1 | 0–6 | 1 | Too small to call vs Aug 31 |
| Sep 14 | 1 | 0–6 | 1 | Too small to call vs Sep 7 |
| Sep 21 | 1 | 0–6 | 1 | Too small to call vs Sep 14 |
| Sep 28 | 2 | 0–8 | 2 | Too small to call vs Sep 21 |
| Oct 5 · line restarts | 2 | 0–8 | 2 | Not comparable: measurement changed |
| Oct 12 | 2 | 0–8 | 2 | Too small to call vs Oct 5 |
| Oct 19 | 2 | 0–8 | 2 | Too small to call vs Oct 12 |
AI agent visits to your web, weekly from Aug 17 to Oct 19. Latest, week of Oct 19: 146, n=146 · 95% range 123–172. Too small to call vs Oct 12. One change in how it is measured, so the line restarts: Oct 5. Weeks of Aug 17 to Aug 24: no reading these weeks. The table has every week.
| Week of | Reading | 95% range | n | Change |
|---|---|---|---|---|
| Aug 17 | — No reading this week | — | — | — |
| Aug 24 | — No reading this week | — | — | — |
| Aug 31 | 46 | 33–62 | 46 | First reading |
| Sep 7 | 64 | 49–82 | 64 | Too small to call vs Aug 31 |
| Sep 14 | 86 | 68–107 | 86 | Too small to call vs Sep 7 |
| Sep 21 | 105 | 85–128 | 105 | Too small to call vs Sep 14 |
| Sep 28 | 124 | 103–148 | 124 | Too small to call vs Sep 21 |
| Oct 5 · line restarts | 151 | 127–178 | 151 | Not comparable: measurement changed |
| Oct 12 | 160 | 136–187 | 160 | Too small to call vs Oct 5 |
| Oct 19 | 146 | 123–172 | 146 | Too small to call vs Oct 12 |
AI agent visits to your API, weekly from Aug 17 to Oct 19. Latest, week of Oct 19: 0, n=0 · 95% range 0–4. Too small to call vs Oct 12. One change in how it is measured, so the line restarts: Oct 5. Weeks of Aug 17 to Aug 24: no reading these weeks. The table has every week.
| Week of | Reading | 95% range | n | Change |
|---|---|---|---|---|
| Aug 17 | — No reading this week | — | — | — |
| Aug 24 | — No reading this week | — | — | — |
| Aug 31 | 0 | 0–4 | 0 | First reading |
| Sep 7 | 0 | 0–4 | 0 | Too small to call vs Aug 31 |
| Sep 14 | 0 | 0–4 | 0 | Too small to call vs Sep 7 |
| Sep 21 | 0 | 0–4 | 0 | Too small to call vs Sep 14 |
| Sep 28 | 0 | 0–4 | 0 | Too small to call vs Sep 21 |
| Oct 5 · line restarts | 0 | 0–4 | 0 | Not comparable: measurement changed |
| Oct 12 | 0 | 0–4 | 0 | Too small to call vs Oct 5 |
| Oct 19 | 0 | 0–4 | 0 | Too small to call vs Oct 12 |
AI agent visits to your MCP, weekly from Aug 17 to Oct 19. Latest, week of Oct 19: 10, n=10 · 95% range 4–19. Too small to call vs Oct 12. One change in how it is measured, so the line restarts: Oct 5. Weeks of Aug 17 to Aug 24: no reading these weeks. The table has every week.
| Week of | Reading | 95% range | n | Change |
|---|---|---|---|---|
| Aug 17 | — No reading this week | — | — | — |
| Aug 24 | — No reading this week | — | — | — |
| Aug 31 | 17 | 9–28 | 17 | First reading |
| Sep 7 | 31 | 21–45 | 31 | Too small to call vs Aug 31 |
| Sep 14 | 35 | 24–49 | 35 | Too small to call vs Sep 7 |
| Sep 21 | 38 | 26–53 | 38 | Too small to call vs Sep 14 |
| Sep 28 | 49 | 36–65 | 49 | Too small to call vs Sep 21 |
| Oct 5 · line restarts | 9 | 4–18 | 9 | Not comparable: measurement changed |
| Oct 12 | 10 | 4–19 | 10 | Too small to call vs Oct 5 |
| Oct 19 | 10 | 4–19 | 10 | Too small to call vs Oct 12 |
Fixes shipped, weekly from Aug 17 to Oct 19. Latest, week of Oct 19: 2, n=2 · 95% range 0–8. Too small to call vs Oct 12. The table has every week.
| Week of | Reading | 95% range | n | Change |
|---|---|---|---|---|
| Aug 17 | 0 | 0–4 | 0 | First reading |
| Aug 24 | 1 | 0–6 | 1 | Too small to call vs Aug 17 |
| Aug 31 | 1 | 0–6 | 1 | Too small to call vs Aug 24 |
| Sep 7 | 1 | 0–6 | 1 | Too small to call vs Aug 31 |
| Sep 14 | 1 | 0–6 | 1 | Too small to call vs Sep 7 |
| Sep 21 | 1 | 0–6 | 1 | Too small to call vs Sep 14 |
| Sep 28 | 0 | 0–4 | 0 | Too small to call vs Sep 21 |
| Oct 5 | 0 | 0–4 | 0 | Too small to call vs Sep 28 |
| Oct 12 | 0 | 0–4 | 0 | Too small to call vs Oct 5 |
| Oct 19 | 2 | 0–8 | 2 | Too small to call vs Oct 12 |
Fixes confirmed working, weekly from Aug 17 to Oct 19. Latest, week of Oct 19: 0, n=0 · 95% range 0–4. Too small to call vs Oct 12. The table has every week.
| Week of | Reading | 95% range | n | Change |
|---|---|---|---|---|
| Aug 17 | 0 | 0–4 | 0 | First reading |
| Aug 24 | 0 | 0–4 | 0 | Too small to call vs Aug 17 |
| Aug 31 | 0 | 0–4 | 0 | Too small to call vs Aug 24 |
| Sep 7 | 1 | 0–6 | 1 | Too small to call vs Aug 31 |
| Sep 14 | 1 | 0–6 | 1 | Too small to call vs Sep 7 |
| Sep 21 | 1 | 0–6 | 1 | Too small to call vs Sep 14 |
| Sep 28 | 0 | 0–4 | 0 | Too small to call vs Sep 21 |
| Oct 5 | 1 | 0–6 | 1 | Too small to call vs Sep 28 |
| Oct 12 | 0 | 0–4 | 0 | Too small to call vs Oct 5 |
| Oct 19 | 0 | 0–4 | 0 | Too small to call vs Oct 12 |