# GTM AI Recommendation Index

> A published index of which go-to-market products frontier AI models recommend when a buyer asks, by category, with the method, the weights, the noise floor and every raw prompt and answer exposed. Published under David Pitta. Not a review site; no user ratings.

Latest edition: 2026-09-08. 14 categories, six prompt variants each (84 prompts), six models with web search enabled (Claude Opus 5, Claude Opus 4.8, GPT-6 Astra, GPT-5.6 Sol, Gemini 3.1 Pro, Perplexity Sonar Pro), a fresh session per prompt, every question asked on behalf of a mid-market B2B software company. A Claude Opus 5 judge labels every named product first_choice (+3), alternative (+2), mention (+1), soft_negative (-2) or hard_negative (-3). Raw judge labels are the permanent record; every number is derived through vendor table v2026-09-08.5. Noise floor: not yet measured; no trend is claimed.

## Categories

- [AI visibility and answer engine tracking tools](https://gtm-ai-index.com/marketing/ai-visibility-tracking/): Peec AI and Otterly tie at 25% of first choices, contested; next Profound 14%, Scrunch 11%
- [Attribution and marketing mix modeling](https://gtm-ai-index.com/marketing/attribution-mmm/): Dreamdata leads with 30% of first choices, contested; next HockeyStack 19%, Google Meridian 15%
- [B2B intent data providers](https://gtm-ai-index.com/sales/b2b-intent-data/): G2 Buyer Intent and Apollo.io tie at 25% of first choices, contested; next Bombora 17%, Warmly 8%
- [Conversational intelligence and call recording](https://gtm-ai-index.com/sales/conversation-intelligence/): Gong leads with 59% of first choices, clear leader; next Avoma 18%, Fireflies.ai 18%
- [Customer data platforms](https://gtm-ai-index.com/gtm-data/customer-data-platforms/): Hightouch leads with 42% of first choices, clear leader; next Segment 25%, Census 17%
- [Customer support and helpdesk](https://gtm-ai-index.com/customer/customer-support-helpdesk/): Zendesk leads with 43% of first choices, clear leader; next Freshdesk 21%, Zoho Desk 21%
- [Data warehouse and reverse ETL for marketing](https://gtm-ai-index.com/gtm-data/data-warehouse-reverse-etl/): Hightouch leads with 40% of first choices, contested; next BigQuery 26%, Snowflake 23%
- [Email marketing platforms](https://gtm-ai-index.com/marketing/email-marketing/): HubSpot Marketing Hub leads with 67% of first choices, consensus; next Brevo 11%, Sender 11%
- [Marketing automation and campaign management](https://gtm-ai-index.com/marketing/marketing-automation/): HubSpot Marketing Hub leads with 64% of first choices, consensus; next Brevo 23%, ActiveCampaign 9%
- [Content management systems for marketing sites](https://gtm-ai-index.com/marketing/marketing-cms/): Webflow leads with 44% of first choices, clear leader; next HubSpot Content Hub 22%, WordPress 22%
- [Product analytics](https://gtm-ai-index.com/gtm-data/product-analytics/): Amplitude leads with 38% of first choices, contested; next Mixpanel 34%, PostHog 24%
- [Sales engagement platforms](https://gtm-ai-index.com/sales/sales-engagement/): Salesloft leads with 48% of first choices, clear leader; next Apollo.io 26%, Outreach 19%
- [SEO and content optimization platforms](https://gtm-ai-index.com/marketing/seo-content-optimization/): Semrush leads with 30% of first choices, contested; next Clearscope 17%, SE Ranking 10%
- [Webinar and virtual event platforms](https://gtm-ai-index.com/marketing/webinar-virtual-events/): Goldcast leads with 41% of first choices, clear leader; next Livestorm 15%, Zoho Webinar 15%

## Headline metrics

- Paraphrase stability, Claude Opus 5: 8 of 14 categories kept the same first choice when the question was reworded
- Paraphrase stability, Claude Opus 4.8: 7 of 14 categories kept the same first choice when the question was reworded
- Paraphrase stability, GPT-6 Astra: 7 of 14 categories kept the same first choice when the question was reworded
- Paraphrase stability, GPT-5.6 Sol: 10 of 14 categories kept the same first choice when the question was reworded
- Paraphrase stability, Gemini 3.1 Pro: 3 of 14 categories kept the same first choice when the question was reworded
- Paraphrase stability, Perplexity Sonar Pro: 4 of 14 categories kept the same first choice when the question was reworded

## Method

- [Methodology](https://gtm-ai-index.com/methodology/): prompt design, model set with exact version strings, scoring rubric and weights, derived metric definitions, normalization rules, noise floor calibration, cadence, conflict policy, known weaknesses
- [Categories](https://gtm-ai-index.com/categories/): every category with its leader and standing
- [Index home](https://gtm-ai-index.com/): standing by category, paraphrase stability, leaders versus criticized defaults, anti-recommendations

## Data

- [Edition log](https://gtm-ai-index.com/data/): every edition with its raw judge export, full responses and report
- [Vendor table v2026-09-08.5](https://gtm-ai-index.com/downloads/2026-09/vendor_aliases.json): canonical names, aliases, category-scoped readings, exclusions, changelog
- [Prompts](https://gtm-ai-index.com/downloads/2026-09/prompts.json): all 84 prompts verbatim
- [Models](https://gtm-ai-index.com/downloads/2026-09/models.json): pinned model set, endpoints, search tools, list prices

## Disclosures

- The author operates Gane, which is building an operating system for B2B go-to-market. Marketing automation and email marketing, which its product overlaps with, are covered with a conflict disclosure on each page rather than excluded; the data there is collected, scored and published exactly like every other category.
- [Sponsors](https://gtm-ai-index.com/sponsors/): sponsorship funds categories, depth and editions, never placement; every sponsor is named on the page it funded
- The judge is an Anthropic model. Cited sources reflect live retrieval only. Category taxonomy is a judgment call. Reading a bare vendor name as its in-category product is a stated assumption and every instance is listed on the category page.