Asked the same question twice, the models moved their own first choice 71% of the time. The leader still held in 23 of 30 readings (10 categories at three buyer sizes).
Ten categories were asked again three days after the edition run, with nothing changed: the same prompts, the same model versions, the same settings, one buyer segment, six framings, fourteen models, 840 answers. Whatever differs between the two runs is noise, and the noise is what this page measures. That is the reason the index asks every question six ways of every model rather than once: on its own, one ask moved 71% of the time; eighty-four, read together, are what the index reports.
Model by model
GLM 4.7 FlashX changed its first choice most often, 92% of its 40 pairs; Perplexity Sonar least, 39% of 38. Pooled over every model and question, 71%.
A flip is the same model, the same question, a different first choice a few days apart. It says how much one answer can be trusted on its own, and nothing about why the model answered as it did.
Category by category
| Category | Leader in the edition run | Share, run one | Share, repeat | Move | Leader |
|---|---|---|---|---|---|
| Contingent workforce man · Small business |  VectorVMS | 26% | 8% | -18 points | changed: SimpleVMS |
| Coaching platforms · Enterprise |  BetterUp | 46% | 31% | -15 points | held |
| Global payroll · Small business |  Deel | 33% | 46% | +14 points | held |
| Performance management · Mid-market |  Lattice | 43% | 56% | +13 points | held |
| Compensation benchmarkin · Enterprise |  Mercer Total Remuneration Survey | 44% | 33% | -11 points | held |
| Performance management · Enterprise |  Betterworks | 37% | 28% | -9 points | changed: Workday HCM Performance Management |
| Contingent workforce man · Enterprise |  SAP Fieldglass | 42% | 34% | -8 points | held |
| AI recruiting assistants · Enterprise |  Eightfold AI | 44% | 36% | -8 points | held |
| Global payroll · Enterprise |  G-P | 23% | 15% | -7 points | changed: Deel |
| Employee engagement surv · Small business |  Workleap Officevibe | 41% | 34% | -7 points | held |
| Compensation benchmarkin · Mid-market |  Payscale | 19% | 26% | +7 points | held |
| AI recruiting assistants · Mid-market |  Manatal | 31% | 24% | -7 points | held |
| Learning management syst · Enterprise |  Docebo | 40% | 46% | +6 points | held |
| Employee engagement surv · Mid-market |  Culture Amp | 37% | 31% | -6 points | held |
| Employee onboarding · Small business |  BambooHR | 34% | 28% | -6 points | changed: Gusto |
| Global payroll · Mid-market | Deel | 30% | 35% | +6 points | held |
| Coaching platforms · Mid-market |  Growthspace | 20% | 15% | -5 points | held |
| Learning management syst · Small business |  TalentLMS | 75% | 71% | -4 points | held |
| Contingent workforce man · Mid-market | VectorVMS | 16% | 20% | +4 points | held |
| Employee onboarding · Enterprise |  Workday | 47% | 50% | +3 points | held |
| Performance management · Small business |  Zoho People | 18% | 15% | -3 points | changed: 15Five |
| Learning management syst · Mid-market | TalentLMS | 51% | 48% | -3 points | held |
| Employee engagement surv · Enterprise | Culture Amp | 29% | 32% | +3 points | held |
| AI recruiting assistants · Small business | Manatal | 29% | 32% | +3 points | held |
| Benefits administration · Mid-market |  Gusto | 20% | 18% | -2 points | changed: Employee Navigator |
| Compensation benchmarkin · Small business |  Pave | 30% | 32% | +2 points | held |
| Employee onboarding · Mid-market |  Rippling | 25% | 26% | +1 point | changed: BambooHR |
| Benefits administration · Small business | Gusto | 62% | 61% | -1 points | held |
| Coaching platforms · Small business |  15Five | 11% | 10% | -1 points | held |
| Benefits administration · Enterprise |  ADP Workforce Now | 30% | 30% | +1 point | held |
In 23 of the 30 readings (10 categories at three buyer sizes) the same product led both runs. Where the leader changed, the two products were within 10 points of each other in the edition run.
The share floor
From the next edition on, a product's change in share counts as movement only when it is larger than 13 points, and a new leader is reported only when it clears the old one by more than that. Nine repeats in ten move a leader less. The floor is measured again with every edition and the method page carries the rule: how the floor is measured.
Cite this
HR AI Recommendation Index, October 2026 Edition: repeat measurement. hr-ai-index.com/research/repeat-measurement/. Published under CC BY 4.0. Every figure on this page is computed from the published edition and changes with it; the edition and its date are the citation.
The output is the models' output. Nothing here says the models can be steered, and nothing here is a recommendation by the index.