AI brand recommendations vary by platform and week: Study

ChatGPT, Claude, Gemini, and Perplexity disagree on the top software brands in four of five categories, according to a new study by AI visibility platform Envoyra.

The study analyzed 2,720 answers over four weeks. It also found that 79% of question-and-assistant combinations had at least one change in recommended brands during the four weekly tests.

Why we care. One AI assistant may frequently recommend your brand, while another rarely mentions it. Recommendations also change over time. That makes it important to track your brand’s visibility across multiple platforms and test regularly.

AI assistants disagreed on top brands. Zendesk was the only brand to rank first across all four assistants, leading in customer support AI. In the other four categories, assistants favored different brands.

• CRM: HubSpot led in three assistants, while Gemini favored Zoho CRM.
• Meeting assistants: ChatGPT favored Otter.ai, Claude and Perplexity favored Fireflies.ai, and Gemini favored Zoom.
• Marketing automation: ChatGPT favored HubSpot, Claude favored Klaviyo, and Gemini and Perplexity favored ActiveCampaign.

AI visibility. ChatGPT and Gemini favored Semrush, while Claude and Perplexity favored Profound.

• Claude recommended Profound in 89 of 96 answers. Gemini never mentioned it.
• Claude also recommended Otterly.AI and Peec AI 75 times each. Gemini didn’t mention either brand.

AI recommendations changed frequently. Only 144 of 680 question-and-assistant combinations returned the same brand list in all four weekly tests. The other 536 (79%) changed at least once.

• HubSpot appeared in 305 of 368 eligible CRM answers. Its weekly counts ranged from 75 to 77.
• Intercom appeared in 62 of 96 customer support answers across the four weeks.

Brand recognition didn’t guarantee recommendations. Envoyra also tested whether assistants could correctly identify 30 software brands by name.

• Assistants correctly identified Keap in all 24 recognition tests but recommended it in just three of 400 eligible CRM answers.
• Profound showed the opposite pattern. Assistants recommended it in 136 of 384 category answers but correctly identified it by name in only five of 24 recognition tests.
• Some assistants confused Fathom with financial-reporting software and Profound with a developer tool from Profound Logic.

About the data. Envoyra tested 170 buyer-style questions per assistant each week from Sept. 13 to Oct. 4. ChatGPT (GPT-4o) and Perplexity (sonar-pro) used web search. Claude (Sonnet 4.6) and Gemini (2.5 Flash) did not.

• Envoyra competes in the AI visibility category included in the study. The company didn’t publish its test questions. Results reflect specific API models and settings, so they may differ from what users see in consumer-facing assistants. The separate brand-recognition test included six answers per brand per assistant.