Webinar Recap: The Future of AI Agents for Retailers and Marketplaces

AI is changing the way we shop.

AI agents can resolve support issues end-to-end, guide shoppers to the right products, and complete actions (like reorders, exchanges, and cart updates) in one conversation.

In our recent webinar, The Future of AI Agents for Retailers & Online Marketplaces, we covered what "agentic commerce" actually means, where it creates value first, and how companies like Meesho, Immobiliare, Deliveroo, and CARS24 are using AI agents.

Why this matters

Traditional online shopping can be overwhelming. You navigate multiple filters or search for products, you endlessly scroll and read reviews, you compare products, and eventually you might make a purchase. 

Agentic commerce changes that. 

We’re already seeing this drive measurable conversion at scale. Amazon’s AI shopping assistant, Rufus (embedded in the Amazon app and site), reported that shoppers who engage with Rufus are 60% more likely to complete a purchase than shoppers who don’t. It’s a strong signal that conversational assistance isn’t just a support layer - it’s becoming a new path to purchase.

Instead of doing all the work yourself, intelligent assistants help you find what you need, answer questions in real time, and can even complete actions on your behalf.

Three things are driving this shift

1. Customer behavior – People expect instant, personalized help that feels like a real conversation, not a support ticket.
2. Market size – Industry projections suggest agent-mediated transactions could hit $136 billion by 2025, reaching $3–5 trillion globally by 2030.
3. Rising expectations – 81% of shoppers prefer brands that personalize their experience. 64% will leave after just one impersonal interaction.

You can give every customer a personalized experience, even when you're serving millions of them.

Demo: Personalized shopping and intelligent customer support

This demo shows how AI agents can guide customers through personalized shopping experiences from product discovery to in-store pickup coordination.

Scenario: A shopper wanted help decorating a dorm room with Taylor Swift-themed items.

What the agent did:

• Offered language flexibility - switched from Spanish to English mid-conversation without losing context
• Suggested initial dorm decor options (LED fairy lights, area rug, posters)
• Navigated directly to specific product pages on request
• Added multiple items to cart seamlessly during the conversation
• Proactively recommended complementary Taylor Swift merchandise (vinyl record, tote bag, journal, fleece throw, phone case) based on customer interest
• Checked real-time inventory at the customer's local Denver store
• Provided store address and specific aisle locations for each item
• Sent an SMS with complete shopping list and aisle locations to the customer's phone
  

Why it matters: The agent recognizes customer preferences and builds a personalized shopping experience around them.

Demo: Order support and customer service

This demo shows how AI agents can handle post-purchase support issues, including order tracking, replacement processing, and policy enforcement.

Scenario: A customer's running shoes were marked as delivered but never arrived. She wanted a replacement in a different color.

What the agent did:

• Authenticated the customer and pulled up order history automatically
• Identified the specific order (running shoes, $79.99, delivered January 25th)
• Offered to send a replacement immediately
• Explained system limitations when asked to change the color on an existing order
• Provided clear options: reorder in original color or cancel and place new order
• Declined the discount request politely while explaining company policy
• Processed the replacement order in real-time

Why it matters: The agent resolves common support issues without human intervention - reducing support ticket volume and response times. 

Success Stories 

Meesho: Scaling Multilingual Support Across India

Meesho, one of India's largest e-commerce marketplaces, needed to deliver empathetic, human-like support at scale across multiple languages and use cases.

They built voice agents with ElevenLabs to automate customer support in both Hindi and English, handling high-volume queries about order statuses, delays, and cancellations.

The results:

• 60,000+ customer calls handled per day in Hindi and English
• Significant reduction in customer call costs
• Reduced average handling time

"Our goal was to create a voice experience that truly feels human—and ElevenLabs helped us achieve that. The clarity, warmth, and tone of the voice played a key role in building trust with our users at scale"  - Siddharth Gupta, GM - New Initiatives at Meesho.

Immobiliare: Converting More Leads in Italy's #1 Real Estate Marketplace

Immobiliare, Italy's largest real estate marketplace with 1.2M+ listings, faced a common problem: sellers were overwhelmed by spam while high-intent buyers received no response.

They deployed a 24/7 AI voice agent that answers questions and qualifies buyers on behalf of sellers.

The results:

• 70% seller adoption rate (up from 42% pre-launch)
• 80% positive user feedback from buyers
• Higher quality leads generated, as buyers were more willing to share contact info with the voice agent than with text forms

"70% of our sellers have already opted into the agent so it can respond to listing-related questions in real-time and identify high intent leads for sellers in an automated way" - Paolo Sabatinelli, Chief Product Officer at Immobiliare.

Deliveroo: Streamlining Operations Across Their Network

Deliveroo, a global food delivery platform with 176K partners, used AI agents to tackle three operational challenges: rider onboarding drop-offs, outdated restaurant hours, and slow activation of partner tags.

Their AI agents re-engage inactive rider applicants, verify restaurant status in real time, and guide partners through tag activation.

The results:

• 81% of riders reached and 30% reactivated within seven days
• Agents reached 75% of restaurant sites and confirmed accurate hours
• 86% of partner sites successfully activated after agent contact

"ElevenLabs Agents transformed our operations—giving us real-time visibility across our network and enabling faster, more consistent engagement with riders and restaurants" - Amy Marangon, Senior Manager, Operations Strategy at Deliveroo.

CARS24: Building Trust in Used Car Sales

CARS24, a leading used-car platform with 4,000+ employees and 200+ customer hubs, faced a unique challenge: 70% of their buyers are purchasing their first vehicle, and phone calls help build trust in an industry that has historically struggled with it.

They integrated ElevenLabs Agents to automate their contact funnel, send automated reminders, and reduce wait times without compromising customer experience.

The results:

• 30% of inbound calls (3 million minutes) automated by AI agents
• 45% of sales now assisted by agents with reduced wait times
• 50% reduction in calling costs

"The used car industry has struggled with trust for generations. Our voice AI solution is changing that reality. With ElevenLabs, every interaction builds confidence instead of doubt” - Head of AI and Innovation at CARS24.

Best Practices for Deploying Retail Agents 

The five core principles for deploying retail agents effectively:

1. Start small, then scale
   Begin with a single, high-impact flow like returns, delivery status, or product discovery. Measure containment and satisfaction before expanding.
   
2. Integrate your data early
   Connect key systems that allow the agent to take real action. Without access to catalog, inventory, and order data, conversations remain static.
   
3. Design with multimodality in mind
   Voice-first doesn’t mean voice-only. Use voice when it’s the fastest path to resolution, and support chat/messaging when that’s the better experience - while keeping one consistent agent behavior across channels.
   
4. Set strict guardrails
   Define escalation paths, tone limits, refund policies, and restricted topics. Test with adversarial prompts before going live.
   
5. Measure continuously
   Treat agents like software - version, test, deploy, and monitor. Success is not launch, but improvement over time.
   

Watch the Full Session 

The Future of AI Agents for Retailers and Marketplaces

Watch the full session here. 

Voice-first AI is changing how customers shop and engage. Retailers like Meesho and Deliveroo have not only reduced costs - they’ve created faster, more personal experiences that improve resolution and trust at scale.

Agentic commerce is no longer theoretical. It’s live, measurable, and ready to scale.