From Chatbots to Concierge: Inside Retail’s Agentic CXM Shift

For Mankani, integrated CXM in this context is not a buzzword; it’s a necessity. “Integrated CXM is the shift from fragmented point solutions to a unified orchestration layer,” he explains. The goal is to connect customer data, inventory and fulfilment into a single real‑time profile so that every touchpoint, whether it’s a WhatsApp chat, an in‑app browse, or a returns request, reads from the same source of truth and responds in a way that feels coherent and context‑aware.

From Passive Helpers to Journey Architects

Most retailers already use some form of conversational interface—chatbots, live chat, guided search—to make shopping easier. These tools have delivered real savings: one analysis estimates chatbots help retailers save around $439 million annually in service costs, while 77% of consumers say conversational interactions make shopping easier and faster. 

Yet many journeys still break down at critical moments: sizing uncertainty, confusing delivery options, opaque return policies, or generic recommendations that ignore context.

The agentic shift is less about adding one more assistant and more about redesigning how decisions are made across the journey. As an example, Mankani points to Merkle’s Experience Concierge for Retail, developed in partnership with Adobe, which uses specialised agents to behave more like a behind‑the‑scenes journey architect than a front‑of‑site helper.

“The experience offers up new opportunities of product comparison that is tailored to the individual via details of the products that matter most to the customer’s profile, which helps with conversion,” he says. Instead of static comparison grids, the experience surfaces the attributes that actually matter to that shopper—fit, material, warranty, or sustainability—based on their behaviour and history.

There is also “promotional placement in the experience to help expose new products a customer might not be aware of that the brand or retailer sells,” turning discovery into a more intentional, personalised form of merchandising.

At the purchase phase, a dedicated Purchase Architect agent performs a compatibility check to ensure that the products in the cart still match the original need or ask, and then “prompt[s] final suggestions right at the point of purchase that could be added to the cart to solve their initial problem even further,” Mankani explains. 

That combination of tailored comparison, intelligent promotion and last‑mile basket optimisation helps explain why Experience Concierge is targeting 15–30% lifts in conversion, 10–25% growth in average order value and 20–40% reductions in customer service costs.

Inside an Agentic Retail Experience

To make the mechanics of agentic CX more concrete, Mankani explains the three pillars Experience Concierge works on.

• Agentic AI acts as the orchestration engine. It deploys six specialised agents to analyse intent, filter options, validate compatibility, and coordinate next‑best actions across the journey.
• Conversational AI adds personality and emotional intelligence, adjusting tone, vocabulary, and guidance based on individual customer signals and confidence levels, meeting exploratory shoppers with detailed explanations and ready‑to‑buy customers with streamlined confirmations.
• Generative UI creates the immersive delivery layer: a dynamic stage that assembles personalised rich content, product imagery, and tutorials in real time, adapting as customers interact.

Those six agents map directly to classic journey stages. The Intent Validator anchors awareness, clarifying what the customer is actually trying to achieve. A Product Curator narrows the field during discovery, while a Solution Builder assembles combinations that make sense at consideration. 

At purchase, the Purchase Architect checks compatibility and proposes thoughtful add‑ons; a post‑purchase Execution Guide helps customers get value from what they bought; and a Lifecycle Navigator looks across loyalty, nudging timely re‑engagement. 

To prevent this from fragmenting into six separate experiences, orchestration is handled by design. “The Intent Validator orchestrates across all agents, maintaining unified context by continuously referencing customer intent, profile data, and session history through a central CDP,” Mankani says. 

Every agent action flows through this layer, which means decisions throughout the journey are grounded in the same customer data and original intent—so what starts as a search for “back‑to‑school laptop” doesn’t end with a cart that looks like it belongs to a completely different shopper. 

Beyond the Chatbot at the Edge

In many organisations, AI is still framed as “another assistant” or a widget sitting on the edge of the site. Mankani is careful to draw a sharp distinction between that mindset and the agentic approach. “Traditional edge chatbots function as static assistants constrained to scripted decision trees and read‑only FAQ lookup, offering little beyond guiding users to external links or escalating tickets,” he notes. 

“Experience Concierge, in contrast, is a multi‑agent orchestration solution woven into your data, content, commerce, and fulfilment systems,” he says. Rather than advising from the sidelines, the agents “actively validate intent, curate compatible products against live inventory, assemble complete solutions, predict needs before checkout, and coach execution post‑purchase.”

They operate continuously across the full journey, referencing unified customer profiles and real‑time signals, so the system is optimising for an outcome, not just a conversation. 

For Mankani, the litmus test is whether the experience is ‘task‑complete’ or merely ‘conversation‑complete’. “The result is task‑complete rather than conversation‑complete: customers get validated, bundled, fulfilled solutions—not just recommendations,” he says. “Experience Concierge doesn’t sit at the edge asking ‘can I help?’—it’s embedded in the engine, autonomously removing friction and driving measurable outcomes like higher conversion, larger baskets, and fewer returns.” 

As external AI platforms send more referral traffic to retailer sites—Adobe data, for instance, shows AI‑driven traffic to US retail sites growing several‑fold year‑on‑year, with conversion rates increasingly outperforming non‑AI referrals—the pressure to have this kind of agentic infrastructure in place will only increase.

Measuring Whether Journeys are Truly Getting Better

If journeys are changing this fundamentally, measurement has to evolve beyond the standard funnel. Conversion rate and basket size still matter, but on their own, they don’t reveal whether experiences are genuinely more fluid or meaningful. 

“Beyond standard top‑line KPIs like conversion rate and basket size, measurement in an integrated CXM environment focuses on behavioural and context‑driven signals that reveal true journey fluidity,” Mankani says. 

His team tracks metrics like Resolution Velocity and Zero‑Click Rate “to evaluate how rapidly complex customer intents are satisfied without forcing manual site navigation or live‑agent intervention.” They measure Context Continuity to verify that user preferences persist seamlessly across channels, and monitor real‑time Sentiment Trajectory throughout the conversation rather than relying purely on delayed CSAT surveys. 

Finally, they analyse Predictive LTV Delta “to gauge long‑term brand loyalty across regional sub‑markets”, a way of asking not just whether a journey is converted, but whether it is compounding the relationship.

What this Means for MENA Retailers

Zooming out, agentic commerce is still in early innings, but adoption curves are steep. Estimates suggest that nearly 40% of consumers already use AI for product discovery and comparison, with more than half of Gen Z shoppers comfortable buying via AI in some form. 

Surveys consistently show that fast, conversational support is now a baseline expectation, not a differentiator: 77% of consumers say conversational experiences make shopping easier, and 60% are more likely to buy from brands that offer live, chat‑based assistance.

In MENA, where digital retail is layered over complex legacy infrastructure and diverse consumer habits, Mankani sees integrated CXM as the operating model that can turn these macro trends into advantage. Connecting data, decisioning and execution under a unified orchestration layer is how retailers move from simply “adding AI” to building journeys where agents quietly do the heavy lifting: recognising intent, resolving friction, and coordinating actions across systems.

The risk for brands is not that agentic experiences will arrive—they already have—but that they will arrive unevenly. Those who treat AI as a peripheral assistant may find themselves optimised for yesterday’s search‑driven funnel. Those who build agentic, integrated CXM systems will be better positioned for a world where shoppers increasingly expect not just information, but intelligent action on their behalf.

Adobe X Merkle is the strategic partner for Vibe Marketing Tech Fest (VMF), happening on 9–10 September 2026, at the Hilton Palm Jumeirah, Dubai. At the event, Merkle & Adobe will showcase Merkle’s Experience Concierge for Retail – an agentic, AI-powered shopping assistant built on the Adobe Experience Cloud. 

Edmund Atcheson, Head of Innovation, MEA at Adobe and Nick McCarthy, Chief Growth Officer, EMEA at Merkle will take the VMF stage to share insights on how brands can rebuild their digital experience strategy for a world where AI is reshaping customer discovery, engagement and expectation, in their session, “The End of the Website as We Know It: Rebuilding Brand Experience for Humans and Agentic AI”. 

For more information, visit Vibe Marketing Tech Fest (VMF) Dubai.