How Agentic AI Will Change Social Commerce

Three years ago, social commerce was framed as ‘the next big thing.’ Today, it is the thing.

Global social commerce sales were estimated at roughly $1.2 trillion in 2024, with multiple forecasts projecting $1.6+ trillion in 2025 and a path to $6 trillion to $27 trillion by the early 2030s, depending on methodology and scenario.

In the U.S., social commerce is on track to exceed $100 billion in annual sales by the mid-2020s, fueled heavily by TikTok Shop, Instagram Shopping, Facebook/Meta platforms and retail media tie-ups.

At the same time, influencer marketing has grown into a ~$24 billion industry in 2024, increasingly intertwined with shoppable video and live selling.

The core reasons social commerce is strategically important have remained the same over time:

• It’s where attention already lives.
• It compresses the funnel from discovery to purchase.
• It provides always-on social proof and feedback loops.
• It yields rich behavioral and interest signals that can power better targeting and personalization.

What has fundamentally changed is how shoppers get into those funnels. Generative and agentic AI are rapidly becoming the new ‘front door’ to digital and social commerce, steering traffic, shaping consideration sets, and increasingly executing transactions on behalf of customers.

How social commerce works today: Trends and implications

Social commerce continues to engage in three principal ways: content-driven, experience-driven and network-driven. Those pillars still hold – but the tactics under each have evolved.

1. Content-driven: Personalization, User-Generated Content (UGC), and AI-accelerated creators

• Personalization as table stakes. Customers now expect every surface – feed, search results, livestream, or AI recommendation – to reflect their preferences and context. Retailers that can unify first-party data and social signals to deliver truly relevant content still see outsized returns.
• UGC as the most trusted ad format. The logic remains unchanged: People trust individuals more than corporate brands. UGC and influencer content still outperform brand-only creative in driving conversion, particularly for younger cohorts. What’s new is that generative AI tools make it dramatically easier to remix UGC into multiple formats (short-form video, carousels, stories), and localize and personalize content at scale for micro-segments.
• Influencer ecosystems professionalize. The influencer marketing industry is now a multi-tens-of-billions market, and many creators operate as full-fledged media businesses, with in-house production and data teams. Micro- and nano-influencers remain critical for precise targeting and authenticity. Virtual and AI-native influencers are emerging as always-on, low-cost ‘faces’ for brands, particularly in fashion, beauty, gaming, and fandom communities.

2. Experience-driven: Live, immersive and conversational

• Livestream shopping becomes a standard format. Live commerce, pioneered at scale in China, has matured but is still under-penetrated in Western markets. Research suggests live commerce can account for 10% to 20% of all e-commerce in markets that fully embrace it, with conversion rates several times higher than traditional product detail pages. In the U.S., TikTok Shop and Instagram Live Shopping are the primary growth engines. ‘Live replays’ and shoppable highlights keep the content productive well beyond the live session.
• AR try-on and 5G-enabled experiences. Augmented reality try-on for beauty, home, and fashion has shifted from novelty to conversion driver. 5G (and soon 5G-Advanced) make high-fidelity, interactive experiences far more accessible in-feed.
• Conversational commerce 2.0 (now AI-native). Early chatbots have evolved into generative AI assistants that answer detailed product questions from reviews, specs, and policies; compare items and summarize trade-offs; and guide shoppers through bundles and projects (e.g. designing a room, planning a trip or stocking a dorm.)

3. Network-driven: Algorithms, communities, and social proof

• Algorithmic discovery is now co-piloted by AI. Shoppers increasingly arrive at social platforms via AI agents and search experiences. Adobe and others report AI-driven referral traffic to retailers has grown more than 10x in under a year, with some analyses citing ~1,200% YoY growth in 2024–2025.
• Communities remain the true moat. Discord servers, Reddit sub-communities, niche Facebook Groups, and private chat channels (WhatsApp, Messenger, Telegram) remain critical ‘whisper networks’ for products. Brands that equip their communities with early access, affiliate tools and AI-powered recommendation helpers will win more organic reach.

Getting the basics right (harder in a multi-platform world)

Many of the operational challenges documented in the original paper still exist – only now they are amplified by more channels, more formats, and AI-accelerated traffic patterns:

• Inventory and orders sync across marketplaces, social shops, live events, and agent surfaces (e.g., ChatGPT, retail-owned agents).
• Engagement → conversion. Turning likes, views, and followers into repeat buyers remains hard; attribution is even more complex when AI agents sit between the shopper and storefront.
• Technology and integration. New touchpoints – agent APIs, social checkout, live commerce tools – put pressure on monolithic commerce platforms.
• Content operations. The need to keep feeds fresh across multiple social channels plus AI surfaces pushes brands toward generative and modular content pipelines.
• Measurement and experimentation. Traditional last-click models are insufficient. Retailers must understand which creators and agents are driving incremental demand, how AI-referred traffic behaves versus organic/direct traffic, and how social commerce contributes to lifetime value, not just first orders.

A composable, headless tech architecture is even more important now: It allows retailers to plug into new social formats and AI agent protocols without full replatforms, and to swap components (search, recommendations, payments, streaming) as innovation accelerates.

The next frontier: Agentic commerce

What is agentic commerce? Agentic commerce is commerce orchestrated by AI agents that can autonomously do the following:

• Discover products and offers across sites and apps. Compare options based on multi-objective constraints (price, sustainability, delivery time, brand preferences). Initiate and complete transactions on a shopper’s behalf (within guardrails).
• Handle post-purchase tasks like tracking, returns, and re-ordering.

McKinsey estimates that by 2030, agentic commerce could drive up to $1 trillion in orchestrated revenue in U.S. B2C retail and between $3 trillion to $5 trillion globally.

Separately, Morgan Stanley forecasts that by 2030 nearly half of U.S. online shoppers could be using AI shopping agents, adding over $100 billion in incremental e-commerce sales.

Early signals from the market

Major retailers and platforms are rolling out AI shopping agents: Walmart’s ‘Sparky’ is a multimodal, generative AI shopping assistant that helps customers plan, compare, and purchase products across categories in the Walmart app. Shoppers also can now browse and buy Walmart products directly inside ChatGPT, turning a general-purpose AI into a shopping front-end. Amazon’s Rufus, Target, Shopify, and others are deploying similar agents.

–AI-driven referrals have exploded, with AI tools influencing an estimated hundreds of billions of dollars in global online sales during the 2024 holiday season.

Will agentic commerce really take off?

Short answer: Yes – though unevenly and in phases. Here are the reasons why:

1. Time and cognitive pressures: Comparing dozens of products, reading reviews, and managing subscriptions is tedious. Agents can compress that overhead dramatically.
2. Better matching of needs and inventory: Agents can search across retailers and social platforms instantly, finding combinations (bundles, alternatives, substitutions) that humans would never see.
3. Retailer benefits:

–More structured, query-rich demand signals and higher conversion from better recommendations.

–New ad and sponsorship formats (e.g., “sponsored prompts” inside agents, which at least one major retailer is already testing).

–Consumer openness: Various surveys show strong interest in AI for shopping inspiration, comparisons, and even returns/claims handling, especially among Gen Z and Millennials.

What could slow or shape adoption

–Trust and transparency. Shoppers will demand clarity: “Is this recommendation best for me or because someone paid for placement?” Models that bias too aggressively toward sponsored results will erode trust.

–Data access and interoperability. Agents need robust product data (attributes, images, availability, prices) and APIs across retailers and social platforms. Fragmented or low-quality feeds will limit agent performance.

–Risk and safety. Fraud, mis-orders, and dark-pattern upsells are amplified when decisions are automated. Regulations may require explicit consent, audit trails, and consumer override rights.

–Retailer- vs. third-party-owned agents. Research suggests consumers currently trust retailer-owned agents (e.g., a Walmart or Target assistant) significantly more than third-party agents that sit between them and the seller.

Most likely scenario

Agentic commerce becomes mainstream over the next 5 to 7 years, but mostly in ‘centaur’ mode – humans plus agents – not full automation. Shoppers will do the following:

• Use agents for search, curation, and planning.
• Retain final approval for higher-stakes or higher-price purchases.
• Let agents fully automate low-risk, repeatable purchases (e.g., household replenishment, groceries, consumables).

How generative and agentic AI will change social commerce

Agentic commerce doesn’t replace social commerce – it plugs into it. The two will increasingly form a single ecosystem.

Discovery and inspiration will move from feeds to ‘conversations.’

• From scrolling to asking: Instead of passively scrolling feeds, shoppers will increasingly say, “Help me find a dress like this one on TikTok, under $80, that ships by Friday.” Agents will use social content (images, videos, captions) as input and return a curated set of options, each with shoppable links – often back into social environments.
• Social content as training and ranking data: Engagement on TikTok, Instagram, YouTube, and livestreams becomes an important signal set for agents. Products with strong authentic engagement and low return/fraud rates get preferential ranking. Negative sentiment and scam reports can rapidly demote products across both social and agent surfaces.

AI-native social storefronts

• Shoppable AI chats inside social apps. Social platforms are already embedding their own AI co-pilots into messaging and feeds. These agents will turn creator content into dynamic product catalogs (e.g., “Show me everything in this video that I can buy”). They can also help viewers recreate a look or project with budget, style and brand preferences.
• Cross-surface continuity. As partnerships between retailers and AI model companies mature, we’ll see flows where inspiration happens on social (e.g., TikTok video) and research and comparison happen in an AI assistant (e.g., ChatGPT).

Creator economy 2.0: AI-amplified social sellers

Generative and agentic AI will reshape the creator and social-seller landscape in the following ways:

• Co-creation and production. Creators will lean on AI for scripting, editing, localization, thumbnail design, and dynamic creative variants – reducing production time and increasing throughput.
• AI brand and virtual influencers. Brands will run persistent, AI-driven personas that interact with fans in comments and DMs, host live or semi-live shopping events with synthetic hosts, and adapt tone and content to specific micro-communities.
• Agent-to-agent commerce. On the horizon are creator-owned agents negotiating bundles, inventory blocks, or affiliate rates directly with retailer agents, with minimal human involvement.

Conversational service inside social commerce 

• Always-on service in DMs and comments. Advanced conversational AI will answer product and policy questions in-thread and resolve simple issues (status, returns, replacements) without channel-hopping; escalate gracefully to human agents for complex cases.
• Reputation management at AI speed. Because social complaints and reviews are public, retailers will deploy agents to detect emerging issues early and reach out proactively to affected customers.

Risk, fraud, and authenticity

Generative AI doesn’t only empower retailers – it also empowers bad actors (deepfake ads, scam stores, fake social proof). Recent reports already highlight the rise of AI-powered holiday shopping scams and lookalike storefronts, especially on social platforms.

Retailers and platforms will need robust defenses:

– Verified creator programs and product authenticity badges

– Model-based detection of synthetic or manipulated content.

– Strong identity, payments, and reputation infrastructure for agents.

– Clear transparency on sponsored content and agent bias.

Strategic roadmap for retailers

To win in a world where social, generative, and agentic commerce intersect, retailers should focus on six priorities:

1. Make data the foundation. Unify first-party data across stores, e-commerce, apps, and social touchpoints. Feed that data into both personalization engines and AI agents.
2. Design for both humans and agents. Treat AI agents as a new type of ‘customer’: provide clean, well-documented APIs, structured product data, and clear policies for recommendations. Optimize product detail pages, feeds, and content not only for social algorithms but for AI agents and retrieval systems.
3. Invest in retailer-owned agents and surfaces. Double down on AI assistants as the central orchestrator of customer journeys, including those that start on social. Integrate social proof (UGC, reviews, creator content) directly into agent experiences.
4. Modernize the tech stack (composable and headless). Build modular services for product, pricing, promotions, checkout, streaming, and messaging. Ensure easy integration with social APIs, AI agent protocols, and emerging standards in agentic payments and identity.
5. Reimagine measurement and incentives. Attribute value fairly across creators, social platforms, and AI agents. Create incentive models that reward true incremental performance, not just last-click wins.
6. Lead on responsible, trustworthy AI. Commit to transparent disclosures when AI is used in recommendations, pricing, or content creation. Put guardrails around sensitive use cases and vulnerable segments. Collaborate with regulators and industry consortia to set standards for agent conduct, advertising, and consumer protection.

Social commerce has moved from a promising experiment to a core, multi-trillion-dollar commerce channel. At the same time, a new wave – agentic commerce powered by generative AI – is rapidly reshaping how shoppers discover, decide and transact.

In this new landscape:

• Social platforms remain the cultural and discovery layer.
• Creators and communities remain critical trust brokers.
• Retailer-owned AI agents emerge as the orchestration layer between shoppers, social content, and supply chains.

Retailers that combine deep customer understanding, operational excellence, and responsible innovation in generative and agentic AI will not only ride the social commerce wave – they will help define the next era of digital commerce itself.