Designing seamless omnichannel experiences with AI

Omnichannel: From strategy to default

“Omnichannel used to be a defining strategy. Today, it’s the minimum expectation,” Gupta said. Wayfair, operating in the home and furnishings space, is leveraging large language models (LLMs) to deliver hyper-personalized experiences.

By combining customer session data with prior purchase context, Wayfair can now render dynamic, image-based recommendations. Imagine seeing your future living room before making a purchase.

With multimodal LLMs, customers aren’t just guided by text, they are visually inspired in real time. “We are creating imagery using AI that reflects a customer’s style and past preferences, dramatically enhancing product discovery,” Gupta added.

Inspiration to installation

For Purandar, the omnichannel challenge is magnified by the nature of home improvement journeys. A washing machine, for instance, might be bought due to a breakdown, a remodel, or for a new home–all requiring different kinds of support.

“Connecting the dots across digital, in-store, and call center journeys is where AI becomes crucial. Our store associate apps now use the same LLM tech as our online platforms to ensure consistency,” he explained.

Lowe’s recently launched Milo, its own agentic commerce platform, enabling natural language shopping and discovery. It is also pioneering Answer Engine Optimization, a strategy to structure product data to be AI-readable and discoverable, whether the query starts on their site or through an external agent like ChatGPT.

AI as an enabler

Basani offered a pragmatic lens: “AI isn’t the strategy, it’s a powerful enabler.” Best Buy is integrating AI across every phase of the customer journey, whether it’s pre-purchase discovery, in-home installations, post-sale services, and even call center support.

From an LLM-powered gift finder to AI-driven appointment scheduling and self-service troubleshooting, every touchpoint is being reimagined.

A key differentiator, he said, is Best Buy’s investments in data unification. “Our data–store, digital, call center, and in-home–all are now consolidated in Google Cloud, enabling models that deliver cross-channel intelligence.” This foundation is what makes personalized, context-aware engagement scalable and responsive.

Engineering for a new paradigm

Building these intelligent systems isn’t straightforward. Wayfair’s Gupta outlined the engineering dilemmas: choosing the right LLM model (balancing latency, cost, and accuracy), compressing A/B testing cycles, and adapting to the faster pace of generative experience development.

“We are rethinking testing, personalisation frameworks, and delivery architecture to align with how quickly GenAI is evolving,” he noted.

Deo, who leads Customer Engineering for AI at Google Cloud India, underscored the massive leap from traditional AI to agentic AI. “It’s not just about automating; it’s about creating new possibilities that simply weren’t feasible before.”

He mentioned a few areas where businesses must rethink their approach. First is contextual intelligence. “LLMs are stateless, so building rich customer context through unified data layers is critical. The better the prompt context, the better the outcome,” Deo said.

Besides, he believes GenAI doesn’t just improve existing systems like recommendation engines, it replaces them with deeper, multimodal logic. But expectations around latency, cost, and performance must be recalibrated accordingly.

From discovery to after-sales

Wayfair now lets shoppers preview purchases in photorealistic room mock‑ups, turning big‑ticket buying into a lifestyle fit check.

On the other hand, Lowe’s extends that visualization across laptop, phone and even Apple Vision Pro while feeding vendors ‘next best action’ prompts that strip friction from onboarding, logistics and services.

Best Buy is test‑driving an LLM‑powered search that invents dynamic, qualitative filters on the fly–goodbye static sidebars.

And Google Cloud’s multimodal workflow authenticates peer‑to‑peer listings: snap the device, box and invoice, cross‑match serial numbers and then auto‑fill rich metadata with a trust stamp without human intervention.

Together, these use cases redraw ecommerce norms.

The road ahead

What was once siloed–product catalogs, call center scripts or in-store experiences–is now becoming unified through AI.

Each of the leaders agreed that agentic AI is not just about efficiency, but empathy. By delivering context-aware, multimodal, and hyper-personalised engagement, global companies are not just selling products, they are building trust and delight at scale.

“This is the time to move beyond incrementalism,” Deo said. “Let’s use AI to create customer experiences that were unimaginable just a few years ago.”