Do AI Models Agree on Which Business Software to Recommend? October 2026
A 60-evaluation RecommendationIndex study measures how much four AI models agreed on CRM, project management, and email marketing recommendations.
This cross-sectional study compares 60 completed prompt-and-model evaluations, 287 observation rows, and 90 model-pair comparisons across CRM, project management, and email marketing.
Supported findings
- Across 90 model-pair comparisons, the average Jaccard overlap between recommendation sets was 0.540.
- Only 34 of 113 distinct prompt-company recommendation slots (30.1%) were recommended by all four models, although every prompt had at least one unanimous recommendation.
- Email marketing had the highest average pairwise overlap at 0.618, compared with 0.510 for CRM and 0.493 for project management.
- The strongest prompt-level agreement was 0.789 for both small-team project management and ecommerce email marketing; enterprise project management was lowest at 0.240.
- Anthropic and Gemini had the highest average pairwise overlap at 0.708; Anthropic and Perplexity had the lowest at 0.435.
All measured prompts
| Market and prompt | Average overlap | Distinct recommendations | Unanimous | Recommended by 3+ |
|---|
CRM
What CRM software platforms have the best AI features and automation capabilities? | 0.589 | 6 | 2 | 4 |
CRM
What is the best CRM software for a mid-sized company? | 0.411 | 6 | 1 | 1 |
CRM
What are the best alternatives to Salesforce and HubSpot CRM? Compare leading CRM platforms. | 0.667 | 8 | 4 | 4 |
CRM
What are the best enterprise CRM software platforms? Compare leading providers and explain why they are recommended. | 0.333 | 11 | 2 | 2 |
CRM
What are the best CRM software platforms for small businesses? Include recommended companies, rankings, and sources if available. | 0.548 | 9 | 3 | 4 |
Project Management
What is the best project management software for a remote team? | 0.394 | 7 | 1 | 3 |
Project Management
What are the best enterprise project management software platforms? Compare leading providers and explain why they are recommended. | 0.240 | 13 | 1 | 2 |
Project Management
What are the best project management software platforms for small teams? Compare leading options and explain why they are recommended. | 0.789 | 6 | 4 | 4 |
Project Management
What are the best project management software platforms for businesses? Include recommended companies, rankings, and sources if available. | 0.587 | 8 | 3 | 4 |
Project Management
What project management software platforms have the strongest AI features and automation capabilities? | 0.454 | 8 | 2 | 3 |
Email Marketing
What email marketing platforms have the strongest AI features and automation? | 0.700 | 5 | 2 | 5 |
Email Marketing
What is the best email marketing software for a growing small business? | 0.489 | 6 | 1 | 4 |
Email Marketing
What are the best email marketing platforms for ecommerce businesses? | 0.789 | 6 | 4 | 4 |
Email Marketing
What are the best email marketing platforms for businesses? Include recommended companies and sources. | 0.589 | 6 | 2 | 4 |
Email Marketing
Compare leading email marketing platforms for small and growing businesses. When is each platform appropriate, and what are the main reasons to choose one over another? Include recommended companies and sources if available. | 0.524 | 8 | 2 | 4 |
Recorded systems and model versions
- OpenAI: gpt-5-mini
- Anthropic: claude-sonnet-4-6
- Google Gemini: gemini-flash-latest
- Perplexity: sonar
Source provenance
| Market | Dataset | Measured UTC | Rows | SHA-256 |
|---|
| CRM | ri-research-crm-2026-10-08-57fc9473 | 2026-10-09T11:30:02+00:00 to 2026-10-09T12:10:15.904+00:00 | 96 | 29627b3bf40183b9210e72b7347a97f27b05a2c9604ec6eb802c8d378b32923e |
| Project Management | ri-research-project-management-2026-10-16-1f948d7c | 2026-10-16T12:10:03.459+00:00 to 2026-10-16T13:30:25.909+00:00 | 97 | cbec47f22fa515c07d8d02edf9d8117cf1081bc92a12405355a923bf31b7aeb9 |
| Email Marketing | ri-research-email-marketing-2026-10-16-1a8250a5 | 2026-10-16T12:10:35.208+00:00 to 2026-10-16T13:10:11.636+00:00 | 94 | 0339219834f4b4bbf4e5a9f11a1e8745d99fba0a0cc50146437695c165997df4 |
Methodology
Cross-sectional comparison of three exact October 2026 RecommendationIndex snapshots. No new model runs were initiated.
For each of 15 prompts, each model's recommendation set is the set of resolved companies whose observation row has recommended=true.
For each prompt and each of the six model pairs, Jaccard overlap equals intersection divided by union. The reported overall average is the unweighted mean of 90 prompt-by-model-pair comparisons.
A recommendation requires recommended=true on a resolved observation row.
A mention is any resolved company observation, whether or not recommended. Mentions are reported separately and are not agreement numerator values.
Citations are source entries attached to observation rows. Citation counts are retained separately and are not agreement numerator values.
Limitations
- This is a three-market, 15-prompt cross-section and does not represent all software categories or buyer questions.
- Jaccard overlap is sensitive to list length. Perplexity returned fewer recommendations on several prompts, so pairwise differences should not be interpreted as model quality.
- The study measures resolved company recommendations, not feature accuracy, product quality, customer adoption, or commercial market share.
- Model outputs are stochastic and time-sensitive; these findings describe the named snapshots and recorded model versions only.
- All prompt-level comparisons are weighted equally, regardless of category size or buyer demand.
- Mentions, citations, and recommendations are distinct. Citation availability is not used to infer whether a recommendation is well supported.
- Observed differences are descriptive; the data does not establish why models disagreed.
Related research
September–October recommendation stability
September three-market benchmark
Read the full methodology
Measure your AI recommendation visibility