Tech stack reality check
Benoit Jacquemont CTO at Akeneo says retailers and brands can unlock more value from the tools they already own and prepare for the AI commerce era
If your software spend is rising faster than your revenue, you are far from alone. Over the past decade, most retailers and brands have quietly assembled sprawling tech stacks based on ERP and MDM to manage operations, DAMs and CMSs to handle content, multiple generations of commerce platforms, an array of marketplace connectors and a forest of spreadsheets plugging the gaps in between. On paper, the stack looks powerful. In reality, teams still complain they don’t have the tools to deliver the rich, consistent experiences customers expect.
The core problem is rarely a lack of technology. It is a lack of usable, trustworthy, well-orchestrated product data. In an era where AI-powered discovery is rewriting the rules of commerce, efficiency, visibility and revenue are the main issues.
Today, buyers increasingly start their journey with conversational prompts to AI agents rather than keywords in a search bar. Instead of clicking through pages of results, they receive condensed, context driven answers that highlight a handful of options. If your product data is incomplete, inconsistent or inaccessible, you are unlikely to feature in those answers at all. In other words, poor product data no longer just hurts conversion on your own channels, it determines whether AI engines can see you in the first place.
The consequences of an overgrown, poorly integrated stack are visible everywhere. Tools bought for urgent projects now overlap or sit idle. Product data fragments across systems – attributes live in ERP, copy sits in CMS, imagery is scattered across one or more DAMs, translations live in spreadsheets and marketplace content is managed through separate portals. Local teams build their own mini systems or workarounds to keep up, often in Excel. IT teams spend an increasing share of their time patching integrations, maintaining brittle data flows and firefighting quality issues. Introducing a new channel or entering a new region can trigger months of rework and custom mapping, even when the core product range barely changes.
In theory, this is a problem of integration. In practice, it is a problem of architecture and ownership. Almost every system in your commercial ecosystem creates, transforms or consumes product information, but very few were designed to be the authoritative home of rich, contextual product data. ERPs focus on transactions, inventory and pricing. MDMs govern cross domain master data. Commerce platforms power front-end experiences. None of them are built to model nuanced product attributes, manage enrichment workflows, support localisation and channel-specific nuances at scale or feed clean, structured product data into AI engines in real time.
This is why product information management is no longer just a content or commerce capability; it has become the backbone of AI-enabled commerce. AI engines are not magical; they can only reason over the information you provide. If your product data is sparse or inconsistent, AI will produce weak, generic or misleading answers. If your data is rich, structured and contextualised, AI can interpret buyer intent and recommend your products with a high degree of relevance and trust.
At the same time, IT’s role is evolving from keeping systems stable to enabling AI readiness. Marketing, product and eCommerce teams still own the customer facing experience, but IT owns the infrastructure that makes AI commerce possible – the integrations that expose product data in real time, the APIs that allow AI agents to query and act, the governance frameworks that ensure quality and compliance and the overall architecture that must scale as new channels and interfaces emerge. In that sense, IT is increasingly on the hook for go to market success, whether they have been briefed that way or not.
A tech stack reality check is therefore the logical place to start. Instead of asking which tools should be added or replaced, IT and business leaders should first examine how product data moves through their existing systems. Mapping the product lifecycle (from onboarding and enrichment, through activation and syndication, to post-purchase updates) usually reveals multiple sources of truth, shadow workflows outside core systems and significant latency between a change being made and that change reaching every channel. Following a single SKU across channels exposes where systems are bypassed because they are too hard to use or too rigid to support new business needs. Quantifying the friction (time to market, error rates, return rates linked to bad information, manual hours spent on exports and imports) provides the foundation for any ROI discussion.
The conclusion many organisations reach is that they do not need more platforms, they need a central, purpose-built product data hub at the heart of the stack. This is the role PIM is designed to play by aggregating product information from ERPs, PLMs, supplier portals, DAMs and spreadsheets into a single, flexible product data model. It provides workflows, roles and validation rules that orchestrate how teams enrich and approve content, with data quality insights that highlight gaps before they propagate downstream. It connects, via APIs and connectors, to commerce platforms, marketplaces, retail media, CRM, marketing tools and AI engines, ensuring that every endpoint consumes consistent, complete and contextualised product information.
The immediate benefit is that existing systems start working better together. Commerce re-platforming becomes faster and safer because clean, unified product data is migrated once into Akeneo and then simply consumed by the new engine. Channel and market expansion become more repeatable, because channel specific and locale specific requirements can be modelled and managed centrally rather than reinvented each time. Tool rationalisation becomes easier as IT can replace fragile custom integrations and overlapping point solutions with a structured, API first approach.
The strategic benefit is that the organisation builds an AI-optimised backbone without ripping and replacing the entire stack. An open, API first PIM enablers product data to flow freely to emerging AI interfaces and agentic commerce platforms. Governance features ensure that what AI engines see is accurate, compliant and aligned with brand standards. Continuous improvement processes, supported by Akeneo’s data quality and activation capabilities, keep the product record fresh and relevant as markets, regulations and buyer behaviours evolve.
Commerce at the front end is shifting from search bars and storefronts to intelligent, adaptive systems that work on behalf of the customer. In that world, your products will only be as visible as your product data is structured. A tech stack reality check that starts with product data, with PIM at the centre as the product data hub, turns disconnected tools into a coherent ecosystem, frees teams from manual workarounds and unlocks the AI-driven revenue potential of the stack you already own.
