Clio and LexisNexis drew 89.1% of the 118.82M visits to these five legal platforms. Harvey and Legora are climbing from far below: up 361.4% and 686.2% in a year, against 11.2% and 12.0% for the incumbents.
We analysed twelve months of SEMrush Traffic Analytics data for three AI-native legal platforms and two established legal software incumbents: visits, engagement, channel mix, monthly trends for every channel, and AI referrals by assistant. Every figure below is computed from the SEMrush exports and compared with the prior twelve months or across the two halves of the year.
- 5 competitors
- Period: Nov 2025 to Oct 2026
- Compared with Nov 2024 to Oct 2025
- Source: SEMrush
Change in visits, Nov 2025 to Oct 2026 against the prior twelve months
+686.2%
+361.4%
+12.0%
+11.2%
-25.7%
An order of magnitude separates the incumbents from the AI natives, and the gap is closing fast. Legora grew to 7.9 times its prior-year visits, Harvey to 4.6 times.
Visits in each twelve-month period, with change year on year
- Nov 2024 to Oct 2025
- Nov 2025 to Oct 2026
Clio drew 67,013,260 visits from November 2025 to October 2026, 8.2 times Harvey and 36.4 times Spellbook. LexisNexis drew 38,885,804. Together the two incumbents added 10,916,971 visits in the year, yet their share of the set fell from 95.4% to 89.1%, because the AI natives grew far faster from a smaller base.
Harvey went from 1,777,108 visits to 8,200,139 (+361.4%), lifting its share from 1.8% to 6.9%. A year earlier Clio drew 33.9 times Harvey’s visits; now it draws 8.2 times. Legora grew 686.2% to 2,879,351, and is now 35.1% of Harvey’s size. Both had their busiest month last: Harvey drew 1,203,925 visits in October 2026, 5.3 times its November 2025 figure of 226,903, and Legora 419,036, 4.3 times its 98,058.
Spellbook’s figure needs care. Visits to spellbook.legal fell 25.7% in the year, but the fall is concentrated after April 2026: 306,627 visits that month, 25,707 in June and 30,400 in October (-90.1%). Every channel fell together, which is consistent with Spellbook moving its site to another domain rather than with lost demand. Spellbook now also operates spellbook.com, which this pack does not cover, so the page treats spellbook.legal as a partial view of the company.
Five platforms, sorted by visits
| Competitor | Visits, full year | YoY | Share of set | Second half vs first | Largest channel | Organic search visits | Paid search visits |
|---|---|---|---|---|---|---|---|
| Clioclio.com | 67,013,260 | +11.2% | 56.4% | +37.2% | Direct (53.7%) | 16,305,203 | 4,820,945 (7.2%) |
| LexisNexislexisnexis.com | 38,885,804 | +12.0% | 32.7% | +7.1% | Direct (44.6%) | 11,518,249 | 2,354,351 (6.1%) |
| Harveyharvey.ai | 8,200,139 | +361.4% | 6.9% | +93.4% | Direct (41.6%) | 2,571,868 | 983,121 (12.0%) |
| Legoralegora.com | 2,879,351 | +686.2% | 2.4% | +143.6% | Direct (41.8%) | 798,887 | 403,147 (14.1%) |
| Spellbookspellbook.legal only, see note | 1,841,733 | -25.7% | 1.6% | -81.5% | Paid search (26.1%) | 319,498 | 481,216 (26.1%) |
Source: SEMrush Traffic Analytics, Overview, monthly Visits and Traffic Journey. Second half vs first compares May to Oct 2026 with Nov 2025 to Apr 2026. Channel shares are of each platform’s Traffic Journey total. Organic search visits are Traffic Analytics channel visits, not Domain Overview organic traffic. Spellbook’s row covers spellbook.legal only; its second-half fall coincides with a likely move to spellbook.com, which is not in this pack. Authority Score is not part of this edition.
Four findings that hold across the category
Direct traffic leads at four of the five. Spellbook alone leaned on paid media, for 42.3% of its visits.
Direct is the largest channel at Clio (53.7%, the highest share in the set), LexisNexis (44.6%), Legora (41.8%) and Harvey (41.6%). LexisNexis draws the largest share from referrals, 13.7%.
Paid search, paid social and display together brought spellbook.legal 42.3% of its visits, with paid search its single largest channel at 26.1%. The two fast growers buy more than the incumbents but far less than Spellbook did: 16.9% at Legora and 14.5% at Harvey, against 8.9% at Clio and 8.6% at LexisNexis.
Share of visits by channel, Nov 2025 to Oct 2026
- Direct
- Organic search
- Paid search, paid social and display
- All other channels
Harvey and Legora grew in every channel between the two halves of the year. Nothing was carried by one lever.
Between November 2025 to April 2026 and May to October 2026, Legora’s organic search visits rose 137.3%, its paid search 152.0% and its direct visits 156.1%. Harvey’s direct visits rose 106.6%, its paid search 98.1% and its paid social 472.4%, the steepest channel rise in the set. Visits Harvey received from ChatGPT rose 429.5%.
The incumbents grew more slowly on the same comparison. Clio’s total visits rose 37.2% and LexisNexis’s 7.1%, with LexisNexis paid search essentially flat (+0.2%). Growth spread across every channel, rather than concentrated in paid media, points to broad demand rather than bought traffic, though this data cannot separate the two.
Source: SEMrush Traffic Analytics, monthly Visits, Organic Search, Paid Search, Paid Social, Direct and ChatGPT trends. Modelled estimates.
A Harvey visit now lasts 23:18, up 147.9%, and is 92.6% as long as a visit to Clio.
Harvey’s visits went from 09:24 and 3.1 pages to 23:18 and 7.9 pages, and its bounce rate fell 9.87 points to 48.00%. A profile like this is consistent with lawyers working inside a product hosted on the domain, not only reading marketing pages. Legora’s bounce rate fell 9.20 points too, to 51.80%.
Clio still holds visitors longest (25:10, 10.0 pages, the lowest bounce rate at 37.25%), but its visits shortened 24.1% from 33:10. Spellbook’s visits lengthened 203.4% to 07:26 while its bounce rate rose 4.47 points to 71.18%, the highest of the five. LexisNexis sits at 07:01.
Average visit duration, Nov 2025 to Oct 2026, with change year on year
25:10
23:18
08:37
07:26
07:01
Organic search visits rose at four of the five between the halves, 30.0% across those four. Legora’s more than doubled.
Legora’s organic search visits rose 137.3%, Harvey’s 64.7%, Clio’s 39.5% and LexisNexis’s 7.6%. Only spellbook.legal fell (-81.7%), in step with every other channel on that domain. Search demand reaching these sites is growing, not shrinking.
AI assistants referred 728,274 visits across the set in the year, 0.61% of all visits. Per 10,000 visits Harvey draws the most (107.2) and Clio the fewest (49.9). ChatGPT is the largest named assistant for all five.
Correlation, not causation. This page reports AI referrals and organic search movements side by side and makes no claim that one caused the other. This data cannot show cause. AI Overview keyword coverage is not part of this edition.
Change in organic search visits, May to Oct 2026 against Nov 2025 to Apr 2026
+137.3%
+64.7%
+39.5%
+7.6%
-81.7%
Six gaps. Two are open here.
When AI assistants send lawyers somewhere, they still send them to LexisNexis.
Named AI assistants sent the five platforms 253,796 visits in the year. LexisNexis took 45.9% of them, 3.3 times Harvey’s count, and drew the most visits from ChatGPT (75,501, against 21,486 for Harvey and 3,664 for Legora). Clio drew the most from Claude, 19,918.
Harvey converts its audience into AI referrals at the highest rate in the set (107.2 per 10,000 visits), but in absolute terms the incumbents still own the assistants’ answers. The AI natives are selling AI to lawyers while the assistants lawyers use point elsewhere. Whoever becomes the default citation for legal AI questions inside those assistants takes a channel no one in this set has locked up.
Organic search for legal software is growing, and the AI natives hold a small slice of it.
The three AI natives drew 11.7% of the set’s organic search visits in the year; Clio and LexisNexis drew 88.3%. Harvey’s organic search visits were 15.8% of Clio’s. Between the halves Clio alone added 2,688,525 organic visits, more than Harvey (628,474) and Legora (325,195) combined.
The growth is real at the AI natives, but the incumbents are taking most of the new search volume in absolute terms. A legal AI platform that publishes for the research and practice questions lawyers already search would be contesting demand that is growing rather than shrinking.
How Legora is outgrowing Harvey at a third of its scale
In full report
What Spellbook’s domain move did to its search and paid traffic
In full report
Why Clio’s visits shortened while its traffic grew
In full report
Which incumbent channels the AI natives can realistically take
In full report
Get the full legal AI report
- All six gaps, with the figures behind each and the move we would make
- Twelve months of monthly traffic by channel for every platform
- A month by month read of Harvey’s and Legora’s rise, channel by channel
- AI referral sources for each platform, assistant by assistant
Tell us your vertical and we will send the full legal AI report.
We send the report by email and follow up once to offer a walkthrough.
Method and provenance
- Source
- SEMrush Traffic Analytics: Overview, Traffic Journey, monthly Visits, monthly channel trends (direct, organic search, referral, paid search, organic social, paid social, email, display, ChatGPT) and AI traffic by source exports for harvey.ai, legora.com, spellbook.legal, clio.com and lexisnexis.com.
- Market
- Worldwide, all devices. Figures are not limited to any one country.
- Competitor set
- Three AI-native legal platforms (Harvey, Legora, Spellbook) and two established legal software companies that sell AI alongside much broader products: Clio in practice management and LexisNexis in legal research and publishing. They are not of equal scale: Clio drew 56.4% of the set’s visits and LexisNexis 32.7%, and both domains carry traffic well beyond their AI products.
- Spellbook
- This pack covers spellbook.legal only. From May 2026 every channel at that domain fell together, which is consistent with a move to another domain. Spellbook now also operates spellbook.com, which is not in this pack, so Spellbook’s figures here understate the company and its year-on-year change should not be read as lost demand.
- Modelled data
- Visit, engagement and channel figures are SEMrush modelled estimates, not first-party analytics from any of these companies. They are suitable for comparing competitors with each other and over time, not for stating any company’s exact traffic. Visits include every subdomain, so customers signed in to web apps hosted on these domains count alongside visitors to marketing pages.
- AI exposure
- AI referral visits and organic search movements are reported as correlation only. Nothing on this page claims that AI assistants or AI Overviews caused a change in organic or any other traffic. AI Overview keyword coverage is not part of this edition. AI referral totals use the Traffic Journey AI traffic channel. The by-source export names 20 sources for 253,796 of those visits (34.8%), 61.7% of them from ChatGPT; SEMrush does not name a source for the rest. Google AI Mode visits appear as google.com in that export, and Claude visits combine claude.ai and downloads.claude.ai.
- Period
- Nov 2025 to Oct 2026, compared with Nov 2024 to Oct 2025. Half-year comparisons set the first half of the period (Nov 2025 to Apr 2026) against the second (May to Oct 2026) and are not seasonally adjusted. The monthly exports also carry a November 2026 row; it is an incomplete month and is excluded throughout.
- Rounding
- Visits are shown as whole numbers, except totals shown in millions to two decimal places. Percentages and growth rates are computed from the raw counts and shown to one decimal place (AI share of visits and bounce rates to two, as SEMrush reports bounce rates), so shares may not sum to exactly 100%. Bounce rate changes are in percentage points. Channel shares are of each platform’s Traffic Journey total, which differs from Overview visits by at most 0.39%.
- Not in this edition
- Authority Score, organic keyword positions, branded and non-branded search, backlinks, referring domains, paid search creative, AI Overview coverage, top pages by share of traffic, and any figures for spellbook.com.