Pinterest Launches AI Shopping Assistant While Embracing Open Source Models to Cut Costs

Pinterest Unveils AI Shopping Assistant with Voice-First Experience

Pinterest has launched its new AI-powered shopping assistant, marking a significant expansion of the visual platform's artificial intelligence capabilities. The Pinterest Assistant, now available in beta to US users aged 18 and above, represents the company's effort to transform how users discover and shop for products through conversational AI 1

1

3

The assistant operates through voice interactions, requiring users to hold down a microphone button to ask questions like "Show me holiday party dresses that fit my style" or "what outfits match this theme." Unlike traditional text-based search, the AI tool provides personalized recommendations based on users' saved collections, boards, and collages, as well as data from users with similar tastes 2

2

4

Source: The Verge

Open Source Strategy Delivers Major Cost Savings

During Pinterest's third-quarter earnings call, CEO Bill Ready highlighted a strategic shift toward open source AI models that has yielded impressive financial results. The company has been testing leading proprietary models against open-source alternatives and discovered significant cost advantages without sacrificing performance quality 1

1

Source: TechCrunch

"We are just getting tremendous performance from open source models specifically for Pinterest use cases on visual AI," Ready told investors. "In early testing, we're seeing orders of magnitude reduction in cost with comparable performance using fine-tuned open source models versus leading off-the-shelf proprietary models" 1

1

This cost optimization strategy comes at a crucial time for Pinterest, as the company faces revenue headwinds. Fourth-quarter revenue is expected to reach between $1.31 billion and $1.34 billion, slightly below analyst expectations, partly due to concerns about tariff impacts on the home furnishings category 1

1

Strong User Engagement and Search Growth

Pinterest's AI investments are paying dividends in user engagement metrics. The platform processed 80 billion monthly queries in the third quarter, representing a 44% increase year-over-year. Monthly active users climbed 12% to 600 million, with the company describing how all logged-in users generate "billions of first-party intent signals" that train its recommendation systems 5

5

Source: PYMNTS

The company's multimodal search capabilities, launched initially for women's fashion in May, combine text and image inputs for more precise results. Visual search drives most of the query growth, with users conducting more searches per session, indicating deeper engagement and stronger commercial intent 5

5

Related Stories

Pinterest CEO Downplays Agentic Shopping, Emphasizes AI-Enabled Assistance

Pinterest crosses $1 billion quarterly revenue as AI-powered visual search reshapes advertising

Pinterest Enhances Visual Search with AI-Powered Features for Personalized Shopping Experience

Pinterest AI Powers Record $1.18B Revenue as Competition Intensifies and Q3 Growth Slows

Technical Architecture and Competitive Positioning

The Pinterest Assistant utilizes a "multimodal" AI model that processes audio, image, and text inputs while maintaining a "visual-first" approach. The core AI model combines proprietary in-house technology built from user signal data with off-the-shelf AI models for basic language processing 3

3

This technical approach differentiates Pinterest from other AI shopping platforms. While Amazon offers Lens for visual product search and Walmart partners with OpenAI for ChatGPT integration, Pinterest's assistant leverages the platform's unique position as a visual discovery engine with deep personalization data 2

2

The company has also addressed concerns about AI-generated content flooding the platform by implementing content labeling and allowing users to "dial down" AI-generated material in their feeds 4

4