# Martech AI Recommendation Index, run mai-20260908-0948-calibration_a

84 prompts, 6 models, 3,572 raw judge labels, 3,577 scored labels after normalization. Published weights: first choice +3, alternative +2, mention +1, soft negative -2, hard negative -3. First full run, 2026-09-08. This is run A of the calibration pair; nothing here has a noise floor yet.

Scored under vendor alias table version 2026-09-08.5. Raw judge labels are the permanent record; the numbers below are derived from them and are regenerated for every run whenever the alias table changes, so the whole series is always scored under one table version. Resolution of raw names: 3,454 by alias or canonical name, 68 by category-scoped alias, 33 from combined answers split into their products, 11 excluded by rule, 8 unresolved and counted under the raw name.

## Paraphrase stability

Share of categories where a model's first choice set on the direct prompt equals its set on the paraphrase.

| Model | Same | Different | Stability |
|---|---|---|---|
| Claude Opus 5 | 8 | 6 | 57% |
| Claude Opus 4.8 | 7 | 7 | 50% |
| GPT-6 Astra | 7 | 7 | 50% |
| GPT-5.6 Sol | 10 | 4 | 71% |
| Gemini 3.1 Pro | 3 | 11 | 21% |
| Perplexity Sonar Pro | 4 | 10 | 29% |

A looser measure, whether a direct first choice survives as a first choice or alternative on the paraphrase, is in the appendix.

## Standing by category

First choice share across the direct, paraphrase, budget and scale prompts and all 6 models. The top product's share is published alongside the label: contested means the top share is at or below 40%, and a category just above the line is not meaningfully different from one just below it. Consensus means every model's direct-prompt first choice is the same single product.

| Category | Top product | Top share | First choices | Status |
|---|---|---|---|---|
| ai-visibility-tracking | Peec AI | 25% | 28 | contested |
| attribution-mmm | Dreamdata | 30% | 27 | contested |
| b2b-intent-data | G2 Buyer Intent | 25% | 24 | contested |
| conversation-intelligence | Gong | 59% | 22 | clear leader |
| customer-data-platforms | Hightouch | 42% | 24 | clear leader |
| customer-support-helpdesk | Zendesk | 43% | 28 | clear leader |
| data-warehouse-reverse-etl | Hightouch | 40% | 43 | contested |
| email-marketing | HubSpot Marketing Hub | 67% | 18 | consensus |
| marketing-automation | HubSpot Marketing Hub | 64% | 22 | consensus |
| marketing-cms | Webflow | 44% | 27 | clear leader |
| product-analytics | Amplitude | 38% | 29 | contested |
| sales-engagement | Salesloft | 48% | 27 | clear leader |
| seo-content-optimization | Semrush | 30% | 30 | contested |
| webinar-virtual-events | Goldcast | 41% | 27 | clear leader |

**ai-visibility-tracking**: Otterly 25%, Peec AI 25%, Profound 14%, Scrunch 11%, AthenaHQ 4%

**attribution-mmm**: Dreamdata 30%, HockeyStack 19%, Google Analytics 4 15%, Google Meridian 15%, CaliberMind 11%

**b2b-intent-data**: Apollo.io 25%, G2 Buyer Intent 25%, Bombora 17%, 6sense 8%, Warmly 8%

**conversation-intelligence**: Gong 59%, Avoma 18%, Fireflies.ai 18%, Fathom 5%

**customer-data-platforms**: Hightouch 42%, Segment 25%, Census 17%, RudderStack 12%, Tracardi 4%

**customer-support-helpdesk**: Zendesk 43%, Freshdesk 21%, Zoho Desk 21%, Pylon 11%, Jira Service Management 4%

**data-warehouse-reverse-etl**: Hightouch 40%, BigQuery 26%, Snowflake 23%, Census 9%, RudderStack 2%

**email-marketing**: HubSpot Marketing Hub 67%, Brevo 11%, Sender 11%, MailerLite 6%, Zoho Campaigns 6%

**marketing-automation**: HubSpot Marketing Hub 64%, Brevo 23%, ActiveCampaign 9%, MailerLite 5%

**marketing-cms**: Webflow 44%, HubSpot Content Hub 22%, WordPress 22%, Storyblok 7%, Contentful 4%

**product-analytics**: Amplitude 38%, Mixpanel 34%, PostHog 24%, Pendo 3%

**sales-engagement**: Salesloft 48%, Apollo.io 26%, Outreach 19%, Amplemarket 4%, Klenty 4%

**seo-content-optimization**: Semrush 30%, Clearscope 17%, SE Ranking 10%, Conductor 7%, Surfer SEO 7%

**webinar-virtual-events**: Goldcast 41%, Livestorm 15%, Zoho Webinar 15%, Demio 7%, ON24 7%

## Anti-recommendations

Products most often labeled soft or hard negative, with the number of models that did so at least once in the category, and the share of that product's labels in the category that were negative. A high negative share on a product with few labels is a warning; a low share on a product with many labels is salience, not sentiment.

- **ai-visibility-tracking**: Profound (4 models, 1 hard; 8 of 29 labels negative), Otterly (4 models; 4 of 24 labels negative), Ahrefs Brand Radar (3 models; 4 of 16 labels negative), Peec AI (3 models; 3 of 26 labels negative)
- **attribution-mmm**: Google Analytics 4 (4 models, 1 hard; 6 of 14 labels negative), Adobe Marketo Measure (4 models; 8 of 14 labels negative), Google Meridian (4 models; 4 of 21 labels negative), Meta Robyn (4 models; 6 of 16 labels negative)
- **b2b-intent-data**: 6sense (6 models, 2 hard; 12 of 32 labels negative), Demandbase (6 models, 2 hard; 9 of 28 labels negative), Bombora (5 models, 1 hard; 6 of 34 labels negative), ZoomInfo (4 models, 1 hard; 8 of 22 labels negative)
- **conversation-intelligence**: Otter.ai (6 models, 4 hard; 11 of 13 labels negative), Gong (5 models, 3 hard; 5 of 32 labels negative), Fireflies.ai (5 models, 2 hard; 8 of 24 labels negative), Fathom (5 models; 6 of 16 labels negative)
- **customer-data-platforms**: Segment (6 models; 9 of 33 labels negative), Adobe Real-Time CDP (5 models, 2 hard; 9 of 16 labels negative), Salesforce Data Cloud (4 models, 1 hard; 6 of 16 labels negative), mParticle (3 models; 3 of 11 labels negative)
- **customer-support-helpdesk**: Salesforce Service Cloud (6 models; 9 of 20 labels negative), Zendesk (6 models; 8 of 30 labels negative), Intercom (4 models; 5 of 27 labels negative), Kayako (2 models, 1 hard; 2 of 2 labels negative)
- **data-warehouse-reverse-etl**: Fivetran (5 models, 1 hard; 7 of 27 labels negative), Census (5 models; 5 of 34 labels negative), Amazon Redshift (4 models, 1 hard; 7 of 18 labels negative), Snowflake (4 models, 1 hard; 6 of 32 labels negative)
- **email-marketing**: Mailchimp (6 models, 3 hard; 10 of 20 labels negative), Constant Contact (5 models, 1 hard; 5 of 9 labels negative), HubSpot Marketing Hub (4 models, 1 hard; 6 of 26 labels negative), Brevo (3 models, 1 hard; 4 of 18 labels negative)
- **marketing-automation**: Adobe Marketo Engage (6 models, 3 hard; 11 of 29 labels negative), HubSpot Marketing Hub (6 models, 1 hard; 9 of 32 labels negative), Salesforce Marketing Cloud (5 models, 3 hard; 6 of 13 labels negative), Salesforce Marketing Cloud Account Engagement (5 models, 2 hard; 6 of 21 labels negative)
- **marketing-cms**: Wix (6 models; 6 of 15 labels negative), WordPress (6 models; 9 of 37 labels negative), Drupal (5 models, 4 hard; 7 of 11 labels negative), Contentful (5 models, 1 hard; 9 of 26 labels negative)
- **product-analytics**: Google Analytics 4 (6 models, 4 hard; 9 of 16 labels negative), Heap (5 models, 1 hard; 6 of 22 labels negative), Amplitude (5 models; 8 of 35 labels negative), PostHog (4 models, 1 hard; 4 of 34 labels negative)
- **sales-engagement**: Outreach (6 models, 3 hard; 12 of 32 labels negative), Salesloft (6 models, 2 hard; 10 of 33 labels negative), Apollo.io (5 models; 5 of 33 labels negative), Groove (3 models, 2 hard; 3 of 10 labels negative)
- **seo-content-optimization**: MarketMuse (4 models; 5 of 19 labels negative), Semrush (4 models; 6 of 28 labels negative), Surfer SEO (4 models; 5 of 26 labels negative), Ahrefs (3 models; 4 of 22 labels negative)
- **webinar-virtual-events**: Zoom Webinars (6 models, 1 hard; 10 of 35 labels negative), GoTo Webinar (5 models, 1 hard; 5 of 15 labels negative), ON24 (5 models; 11 of 26 labels negative), Microsoft Teams (4 models, 3 hard; 5 of 15 labels negative)

## Leadership and criticism

First-choice share and negative label rate are independent, so every product with at least 10 labels in a category is placed on both axes. Cutoffs: leader at 30% or more of first choices, criticized at 25% or more negative labels. An endorsed leader is recommended and not criticized; a criticized default is recommended while the models list its problems; a criticized challenger is named mostly to be warned against.

| Category | Product | First-choice share | Negative rate | Labels | Quadrant |
|---|---|---|---|---|---|
| ai-visibility-tracking | Peec AI | 25% | 12% | 26 | accepted challenger |
| ai-visibility-tracking | Otterly | 25% | 17% | 24 | accepted challenger |
| ai-visibility-tracking | Profound | 14% | 28% | 29 | criticized challenger |
| ai-visibility-tracking | Scrunch | 11% | 7% | 15 | accepted challenger |
| ai-visibility-tracking | Semrush AI Visibility Toolkit | 4% | 14% | 21 | accepted challenger |
| ai-visibility-tracking | HubSpot AEO Grader | 4% | 18% | 11 | accepted challenger |
| ai-visibility-tracking | Ahrefs Brand Radar | 0% | 25% | 16 | criticized challenger |
| attribution-mmm | Dreamdata | 30% | 0% | 22 | accepted challenger |
| attribution-mmm | HockeyStack | 19% | 0% | 17 | accepted challenger |
| attribution-mmm | Google Meridian | 15% | 19% | 21 | accepted challenger |
| attribution-mmm | Google Analytics 4 | 15% | 43% | 14 | criticized challenger |
| attribution-mmm | Triple Whale | 4% | 27% | 15 | criticized challenger |
| attribution-mmm | Meta Robyn | 4% | 38% | 16 | criticized challenger |
| attribution-mmm | Measured | 0% | 7% | 14 | accepted challenger |
| attribution-mmm | Recast | 0% | 17% | 12 | accepted challenger |
| attribution-mmm | Rockerbox | 0% | 21% | 14 | accepted challenger |
| attribution-mmm | Northbeam | 0% | 36% | 14 | criticized challenger |
| attribution-mmm | Adobe Marketo Measure | 0% | 57% | 14 | criticized challenger |
| b2b-intent-data | G2 Buyer Intent | 25% | 8% | 25 | accepted challenger |
| b2b-intent-data | Apollo.io | 25% | 30% | 10 | criticized challenger |
| b2b-intent-data | Bombora | 17% | 18% | 34 | accepted challenger |
| b2b-intent-data | 6sense | 8% | 38% | 32 | criticized challenger |
| b2b-intent-data | Cognism | 4% | 7% | 14 | accepted challenger |
| b2b-intent-data | ZoomInfo | 4% | 36% | 22 | criticized challenger |
| b2b-intent-data | Dealfront | 0% | 0% | 14 | accepted challenger |
| b2b-intent-data | TrustRadius | 0% | 0% | 12 | accepted challenger |
| b2b-intent-data | Demandbase | 0% | 32% | 28 | criticized challenger |
| conversation-intelligence | Gong | 59% | 16% | 32 | endorsed leader |
| conversation-intelligence | Avoma | 18% | 7% | 27 | accepted challenger |
| conversation-intelligence | Fireflies.ai | 18% | 33% | 24 | criticized challenger |
| conversation-intelligence | Fathom | 5% | 38% | 16 | criticized challenger |
| conversation-intelligence | Clari Copilot | 0% | 9% | 22 | accepted challenger |
| conversation-intelligence | Salesloft Conversations | 0% | 14% | 14 | accepted challenger |
| conversation-intelligence | Chorus by ZoomInfo | 0% | 14% | 21 | accepted challenger |
| conversation-intelligence | Otter.ai | 0% | 85% | 13 | criticized challenger |
| customer-data-platforms | Hightouch | 42% | 8% | 24 | endorsed leader |
| customer-data-platforms | Segment | 25% | 27% | 33 | criticized challenger |
| customer-data-platforms | Census | 17% | 0% | 11 | accepted challenger |
| customer-data-platforms | RudderStack | 12% | 0% | 23 | accepted challenger |
| customer-data-platforms | Tealium | 0% | 14% | 14 | accepted challenger |
| customer-data-platforms | mParticle | 0% | 27% | 11 | criticized challenger |
| customer-data-platforms | Salesforce Data Cloud | 0% | 38% | 16 | criticized challenger |
| customer-data-platforms | Adobe Real-Time CDP | 0% | 56% | 16 | criticized challenger |
| customer-support-helpdesk | Zendesk | 43% | 27% | 30 | criticized default |
| customer-support-helpdesk | Zoho Desk | 21% | 0% | 15 | accepted challenger |
| customer-support-helpdesk | Freshdesk | 21% | 12% | 26 | accepted challenger |
| customer-support-helpdesk | Jira Service Management | 4% | 27% | 15 | criticized challenger |
| customer-support-helpdesk | Front | 0% | 0% | 11 | accepted challenger |
| customer-support-helpdesk | HubSpot Service Hub | 0% | 7% | 27 | accepted challenger |
| customer-support-helpdesk | Intercom | 0% | 19% | 27 | accepted challenger |
| customer-support-helpdesk | Help Scout | 0% | 20% | 15 | accepted challenger |
| customer-support-helpdesk | Salesforce Service Cloud | 0% | 45% | 20 | criticized challenger |
| data-warehouse-reverse-etl | Hightouch | 40% | 6% | 35 | endorsed leader |
| data-warehouse-reverse-etl | BigQuery | 26% | 9% | 35 | accepted challenger |
| data-warehouse-reverse-etl | Snowflake | 23% | 19% | 32 | accepted challenger |
| data-warehouse-reverse-etl | Census | 9% | 15% | 34 | accepted challenger |
| data-warehouse-reverse-etl | RudderStack | 2% | 6% | 18 | accepted challenger |
| data-warehouse-reverse-etl | dbt | 0% | 4% | 23 | accepted challenger |
| data-warehouse-reverse-etl | Airbyte | 0% | 15% | 20 | accepted challenger |
| data-warehouse-reverse-etl | Segment | 0% | 25% | 12 | criticized challenger |
| data-warehouse-reverse-etl | Fivetran | 0% | 26% | 27 | criticized challenger |
| data-warehouse-reverse-etl | Amazon Redshift | 0% | 39% | 18 | criticized challenger |
| data-warehouse-reverse-etl | Databricks | 0% | 44% | 16 | criticized challenger |
| email-marketing | HubSpot Marketing Hub | 67% | 23% | 26 | endorsed leader |
| email-marketing | Brevo | 11% | 22% | 18 | accepted challenger |
| email-marketing | MailerLite | 6% | 17% | 12 | accepted challenger |
| email-marketing | Salesforce Marketing Cloud Account Engagement | 0% | 0% | 10 | accepted challenger |
| email-marketing | Customer.io | 0% | 0% | 10 | accepted challenger |
| email-marketing | ActiveCampaign | 0% | 5% | 21 | accepted challenger |
| email-marketing | Adobe Marketo Engage | 0% | 18% | 11 | accepted challenger |
| email-marketing | Klaviyo | 0% | 24% | 17 | accepted challenger |
| email-marketing | Mailchimp | 0% | 50% | 20 | criticized challenger |
| marketing-automation | HubSpot Marketing Hub | 64% | 28% | 32 | criticized default |
| marketing-automation | Brevo | 23% | 0% | 13 | accepted challenger |
| marketing-automation | ActiveCampaign | 9% | 14% | 22 | accepted challenger |
| marketing-automation | Klaviyo | 0% | 15% | 13 | accepted challenger |
| marketing-automation | Salesforce Marketing Cloud Account Engagement | 0% | 29% | 21 | criticized challenger |
| marketing-automation | Adobe Marketo Engage | 0% | 38% | 29 | criticized challenger |
| marketing-automation | Salesforce Marketing Cloud | 0% | 46% | 13 | criticized challenger |
| marketing-automation | Mailchimp | 0% | 58% | 12 | criticized challenger |
| marketing-cms | Webflow | 44% | 7% | 29 | endorsed leader |
| marketing-cms | HubSpot Content Hub | 22% | 17% | 29 | accepted challenger |
| marketing-cms | WordPress | 22% | 24% | 37 | accepted challenger |
| marketing-cms | Storyblok | 7% | 13% | 15 | accepted challenger |
| marketing-cms | Contentful | 4% | 35% | 26 | criticized challenger |
| marketing-cms | Sanity | 0% | 16% | 25 | accepted challenger |
| marketing-cms | Squarespace | 0% | 33% | 12 | criticized challenger |
| marketing-cms | Wix | 0% | 40% | 15 | criticized challenger |
| marketing-cms | Sitecore | 0% | 50% | 10 | criticized challenger |
| marketing-cms | Drupal | 0% | 64% | 11 | criticized challenger |
| marketing-cms | Adobe Experience Manager | 0% | 69% | 13 | criticized challenger |
| product-analytics | Amplitude | 38% | 23% | 35 | endorsed leader |
| product-analytics | Mixpanel | 34% | 12% | 34 | endorsed leader |
| product-analytics | PostHog | 24% | 12% | 34 | accepted challenger |
| product-analytics | Pendo | 3% | 16% | 25 | accepted challenger |
| product-analytics | Heap | 0% | 27% | 22 | criticized challenger |
| product-analytics | Google Analytics 4 | 0% | 56% | 16 | criticized challenger |
| sales-engagement | Salesloft | 48% | 30% | 33 | criticized default |
| sales-engagement | Apollo.io | 26% | 15% | 33 | accepted challenger |
| sales-engagement | Outreach | 19% | 38% | 32 | criticized challenger |
| sales-engagement | HubSpot Sales Hub | 0% | 0% | 26 | accepted challenger |
| sales-engagement | Groove | 0% | 30% | 10 | criticized challenger |
| seo-content-optimization | Semrush | 30% | 21% | 28 | endorsed leader |
| seo-content-optimization | Clearscope | 17% | 12% | 24 | accepted challenger |
| seo-content-optimization | Conductor | 7% | 17% | 12 | accepted challenger |
| seo-content-optimization | Surfer SEO | 7% | 19% | 26 | accepted challenger |
| seo-content-optimization | Frase | 3% | 13% | 15 | accepted challenger |
| seo-content-optimization | Ahrefs | 3% | 18% | 22 | accepted challenger |
| seo-content-optimization | MarketMuse | 3% | 26% | 19 | criticized challenger |
| webinar-virtual-events | Goldcast | 41% | 6% | 17 | endorsed leader |
| webinar-virtual-events | Livestorm | 15% | 0% | 20 | accepted challenger |
| webinar-virtual-events | Demio | 7% | 16% | 19 | accepted challenger |
| webinar-virtual-events | Zoom Webinars | 7% | 29% | 35 | criticized challenger |
| webinar-virtual-events | ON24 | 7% | 42% | 26 | criticized challenger |
| webinar-virtual-events | Webex Webinars | 0% | 25% | 12 | criticized challenger |
| webinar-virtual-events | Cvent | 0% | 31% | 13 | criticized challenger |
| webinar-virtual-events | GoTo Webinar | 0% | 33% | 15 | criticized challenger |
| webinar-virtual-events | Microsoft Teams | 0% | 33% | 15 | criticized challenger |
| webinar-virtual-events | RingCentral Events | 0% | 41% | 17 | criticized challenger |

## Treatment of lab-affiliated products by model

Self-preference is only measurable with a control group: how the lab's own model treats its products compared with how every other model treats the same products. Anthropic, OpenAI have no products in the covered categories with enough labels to measure, so nothing is claimed about them.

**Google products** (13 products, 117 labels). Cells are positive/negative/all labels.

| Product | Claude Opus 5 | Claude Opus 4.8 | GPT-6 Astra | GPT-5.6 Sol | Gemini 3.1 Pro | Perplexity Sonar Pro |
|---|---|---|---|---|---|---|
| BigQuery | 5/1/6 | 5/1/6 | 5/1/6 | 6/0/6 | 6/0/6 | 5/0/5 |
| Google Analytics 4 | 6/1/7 | 4/2/6 | 2/2/4 | 3/3/6 | 2/3/5 | 2/4/6 |
| Google Meridian | 5/1/6 | 4/0/4 | 3/1/4 | 4/1/5 | 1/1/2 |  |
| Google Search Console | 3/0/3 | 1/0/1 | 1/0/1 | 2/0/2 | 1/0/1 | 2/0/2 |
| Google Meet | 1/0/1 |  |  | 0/1/1 | 2/0/2 |  |
| Looker | 2/0/2 |  |  |  |  | 0/1/1 |
| Google LightweightMMM |  |  |  | 0/1/1 | 1/0/1 |  |
| YouTube Live | 1/0/1 |  |  |  | 1/0/1 |  |
| Looker Studio | 1/0/1 |  |  |  | 0/1/1 |  |
| Google Marketing Platform |  |  |  | 0/1/1 |  |  |
| Attribution 360 | 0/1/1 |  |  |  |  |  |
| Google Keyword Planner | 1/0/1 |  |  |  |  |  |
| Adometry | 0/1/1 |  |  |  |  |  |

Mean label weight on Google products: Gemini 3.1 Pro +1.00 over 19 labels; the other 5 models +0.95 over 98 labels. Weights run from +3 (first choice) to -3 (hard negative). A gap under about half a point on this few labels is not a finding.

## Discontinued products recommended

Positive labels on products the catalog marks discontinued. A model offering a shut-down product is a retrieval failure worth naming.

- Google LightweightMMM (Deprecated by Google in favor of Meridian.) labeled alternative by Gemini 3.1 Pro on the negative prompt in attribution-mmm
- Grouparoo (Open source reverse ETL, shut down 2023.) labeled alternative by Perplexity Sonar Pro on the negative prompt in data-warehouse-reverse-etl

## Normalization

Unresolved, 8 labels counted under the raw name, for human review:

Adobe Marketing Campaign Analytics in attribution-mmm (2), Salesforce in attribution-mmm (2), Adobe in ai-visibility-tracking (1), Microsoft’s Sales Agent in conversation-intelligence (1), Native HubSpot/Salesforce attribution in attribution-mmm (1), Oracle in data-warehouse-reverse-etl (1)

Excluded by rule, 11 labels not scored:

Google Ads (2; ad platform self-reported attribution, not a vendor in the category), Meta Ads Manager (2; ad platform self-reported attribution, not a vendor in the category), Amazon (1; ad platform self-reported attribution, not a vendor in the category), Excel (1; general-purpose tool, not a vendor in the category), Google geo experiments/Ads lift studies (1; ad platform lift study, not a vendor in the category), LinkedIn Ads (1; ad platform self-reported attribution, not a vendor in the category), Meta Conversion Lift (1; ad platform lift study, not a vendor in the category), Native Google, Meta, TikTok and Amazon attribution (1; ad platform self-reported attribution, not a vendor in the category), TikTok (1; ad platform self-reported attribution, not a vendor in the category)

Category-scoped resolutions, 68 labels. A bare vendor name was read as that vendor's product for the asked category, which assigns specificity the model did not write; each one is listed so the assumption is visible:

HubSpot in email-marketing read as HubSpot Marketing Hub (10), HubSpot in marketing-automation read as HubSpot Marketing Hub (9), Salesloft in conversation-intelligence read as Salesloft Conversations (7), Semrush in ai-visibility-tracking read as Semrush AI Visibility Toolkit (7), Ahrefs in ai-visibility-tracking read as Ahrefs Brand Radar (6), Zoom in webinar-virtual-events read as Zoom Webinars (6), HubSpot in customer-data-platforms read as HubSpot (4), Oracle in customer-data-platforms read as Oracle Unity (3), Outreach in conversation-intelligence read as Outreach Kaia (3), HubSpot in ai-visibility-tracking read as HubSpot AEO Grader (2), HubSpot in attribution-mmm read as HubSpot Marketing Hub (2), Salesforce in sales-engagement read as Salesforce Sales Engagement (2), Adobe in customer-data-platforms read as Adobe Real-Time CDP (1), Clari in conversation-intelligence read as Clari Copilot (1), HubSpot in sales-engagement read as HubSpot Sales Hub (1), HubSpot (Marketing Hub) in marketing-automation read as HubSpot Marketing Hub (1), HubSpot's native tooling in conversation-intelligence read as HubSpot Sales Hub (1), SAP in customer-data-platforms read as SAP Customer Data Platform (1), Salesforce in customer-data-platforms read as Salesforce Data Cloud (1)

Combined answers split into one label per product: Account Engagement/Pardot, AppsFlyer / Adjust, Clari + Salesloft, Clari Copilot / Salesloft Conversation Intelligence, Cvent + ON24, Cvent / Bizzabo, Google Meridian / Meta Robyn, Hugo/Astro, Klaviyo/Bloomreach, Marketo / Pardot (Salesforce Account Engagement), Salesforce Marketing Cloud & Marketing Cloud Account Engagement (formerly Pardot), Salesforce Marketing Cloud (SFMC) & Pardot, Salesforce Marketing Cloud / Account Engagement, Salesforce Marketing Cloud/Account Engagement, Salesforce Marketing Cloud/Pardot, Salesforce/Adobe, Sitefinity/Kentico

## Appendix: loose paraphrase stability

- Claude Opus 5: a direct first choice kept as first choice or alternative in 13 of 14 categories (93%); differing under the strict rule: attribution-mmm, b2b-intent-data, customer-data-platforms, customer-support-helpdesk, seo-content-optimization, webinar-virtual-events
- Claude Opus 4.8: a direct first choice kept as first choice or alternative in 13 of 14 categories (93%); differing under the strict rule: ai-visibility-tracking, b2b-intent-data, customer-data-platforms, data-warehouse-reverse-etl, product-analytics, seo-content-optimization, webinar-virtual-events
- GPT-6 Astra: a direct first choice kept as first choice or alternative in 12 of 14 categories (86%); differing under the strict rule: attribution-mmm, b2b-intent-data, customer-support-helpdesk, data-warehouse-reverse-etl, product-analytics, sales-engagement, seo-content-optimization
- GPT-5.6 Sol: a direct first choice kept as first choice or alternative in 13 of 14 categories (93%); differing under the strict rule: ai-visibility-tracking, data-warehouse-reverse-etl, marketing-cms, seo-content-optimization
- Gemini 3.1 Pro: a direct first choice kept as first choice or alternative in 13 of 14 categories (93%); differing under the strict rule: ai-visibility-tracking, attribution-mmm, b2b-intent-data, conversation-intelligence, customer-support-helpdesk, data-warehouse-reverse-etl, marketing-cms, product-analytics, sales-engagement, seo-content-optimization, webinar-virtual-events
- Perplexity Sonar Pro: a direct first choice kept as first choice or alternative in 11 of 14 categories (79%); differing under the strict rule: ai-visibility-tracking, b2b-intent-data, conversation-intelligence, customer-support-helpdesk, data-warehouse-reverse-etl, marketing-cms, product-analytics, sales-engagement, seo-content-optimization, webinar-virtual-events