Enterprise AI assistants

ProductFirst-choice shareNegative rateLabelsQuadrantSince October 2026

01Microsoft 365 Copilot33%13%55endorsed leader▲+2Since October 2026: 31% → 33%, +2 points. Inside the 11-point floor: within noise.31% → 33%

02ChatGPT Lab in the set20%18%61accepted challenger▼−5Since October 2026: 25% → 20%, −5 points. Inside the 11-point floor: within noise.25% → 20%

03Glean8%15%33accepted challenger▲+2Since October 2026: 6% → 8%, +2 points. Inside the 11-point floor: within noise.6% → 8%

04Amazon Q6%6%16accepted challenger▼−4Since October 2026: 10% → 6%, −4 points. Inside the 11-point floor: within noise.10% → 6%

05Google Gemini for Workspace Lab in the set4%6%17accepted challenger▲+4Since October 2026: 0% → 4%, +4 points. Inside the 11-point floor: within noise.0% → 4%

06Claude for Work Lab in the set2%32%34criticized challenger▲+2Since October 2026: 0% → 2%, +2 points. Inside the 11-point floor: within noise.0% → 2%

Show the three products at 0%, ordered by negative rate

07Google Gemini Lab in the set0%25%24criticized challenger=heldSince October 2026: 0% → 0%, ±0 points. Inside the 11-point floor: within noise.0% → 0%

09Moveworks0%25%12criticized challenger=heldSince October 2026: 0% → 0%, ±0 points. Inside the 11-point floor: within noise.0% → 0%

08Salesforce Agentforce0%20%10accepted challenger=heldSince October 2026: 0% → 0%, ±0 points. Inside the 11-point floor: within noise.0% → 0%

The floor is 11 points of share, measured: how far the models move a leader on their own when the same questions are asked twice with nothing changed. A larger change is movement; a smaller one can't be told apart from repeat variation, and both are shown. Each index measures its own floor from its own repeat, over its own categories, so the floor differs from one index to the next. Both editions asked the same fourteen models. How the floor is measured

ChatGPT is made by OpenAI, whose model GPT-5.4 mini and GPT-6 Luna is in the set. On the four share prompts that model made it the first choice four times of 8; the other twelve models six times of 48. Google Gemini for Workspace is made by Google, whose model Gemini 3.5 Flash is in the set. On the four share prompts that model made it the first choice one time of 4; the other thirteen models one time of 52. Claude for Work is made by Anthropic, whose model Claude Haiku 4.5 is in the set. On the four share prompts that model made it the first choice zero times of 4; the other thirteen models one time of 52. Google Gemini is made by Google, whose model Gemini 3.5 Flash is in the set. On the four share prompts that model made it the first choice zero times of 4; the other thirteen models zero times of 52. Lab treatment is defined on the method page; the row is marked, not excluded.

Recommended versus criticized

Every product with at least 10 labels here, on both axes. The 30% line names a quadrant, not the verdict above: that one needs more than 40%.

Criticized challengerCriticized default

Negative label rate →

01

02

03

04

G05

C06

G07

08

09

Accepted challengerEndorsed leader

0%First-choice share → · lines at 30% share and 25% negative50%

Key

01Microsoft 365 Copilot33%

02ChatGPT20%

03Glean8%

04Amazon Q6%

05Google Gemini for Workspace4%

06Claude for Work2%

07Google Gemini0%

08Salesforce Agentforce0%

09Moveworks0%

Google's top ten for the three searches hold 19 sites; 1 of them are among the sites the models cited here (glean.com). The first result for “best enterprise ai assistant” is linkedin.com.