Implementing Akeneo PIM: From Old Stumbling Blocks to New Features
Anyone planning an Akeneo PIM implementation today (PIM stands for Product Information Management) tends to hear the same warnings as a few years ago. Data modelling drags on for months. Supplier data arrives as Excel spreadsheets sent by email. A missing or incorrect attribute goes unnoticed until a customer finds it. Some of these warnings are now out of date. Not because the underlying problem has disappeared, but because Akeneo has translated it into concrete, dated product features over the past few years.
In short: Several of the recurring stumbling blocks in an Akeneo PIM implementation are now addressed by named features. Months-long data modelling is handled by the Data Architect Agent, announced in July 2025. Supplier data chaos is addressed by the AI-powered Supplier Data Manager, from 2023 and 2024. Invisible data quality is covered by Data Quality Insights, expanded step by step since 2020. Other stumbling blocks remain stumbling blocks because they are not a software question. No feature decides who in a company is responsible for which data field. This article sorts the features chronologically: then, now, and what Akeneo itself announces for the future. Every feature comes with a source and a date, so the claim can be checked.
What the Release History Shows
Akeneo does not release new features in a single annual leap. It releases them continuously. For 2024 alone, the official release notes list at least nine separately dated updates between February and December, plus the seasonally named main release "Spring 2024" in March of the same year. This cadence can be read almost like a log. Every new feature answers a question that apparently had no answer before. When a vendor announces in July 2025 that it is solving the biggest pain point in handling product data, as Akeneo put it when launching the Data Architect Agent, that statement says something about the state of things before July 2025. It also says something about how seriously the problem was taken internally.
For a company planning a PIM implementation with Akeneo, this way of reading the product is practical. Instead of relying on experience reports from three or five years ago, the release history shows which stumbling block is genuinely still a software problem and which one never was.
Then: Stumbling Blocks That Came With Every Akeneo PIM Implementation
Before looking at individual features, it is worth looking at the original promise. Akeneo consistently positions itself around the same three verbs: centralise, manage and enrich product data. As a starting point, the vendor names scattered product data on its own website: spreadsheets, ERP systems, e-commerce platforms and supplier portals, leading to inconsistent product descriptions, slow market launches and poor customer experiences. The stated goal behind this is not primarily a nicer database. It is consistent, compelling product experiences across every sales channel. Akeneo likes to call itself the "single source of truth" for product information. Seen from this angle, the three stumbling blocks below are not random examples. They are the specific points where this promise of centralisation most often broke down in practice.
Three problems came up regularly in projects, long before they had a product name attached to them.
Data modelling was a month-long project. Attribute models (the structure of individual product characteristics such as dimensions or material), families and categories had to be derived manually from ERP exports (Enterprise Resource Planning, the central inventory management system) or shop catalogues. The process tied up expertise, time and several rounds of coordination. Looking back, in the announcement of its own solution to this problem, Akeneo itself called this the biggest pain point in the entire handling of product data.
Supplier data arrived unstructured and scattered. In the press release for the Supplier Data Manager, Akeneo names three concrete symptoms in October 2023 that customers had reported before: chasing individual suppliers' catalogues through email inboxes, manually correcting unstructured files, and managing dozens of individual files with no shared format. The same three symptoms still appear, almost word for word, on the Supplier Data Manager's own product page today. That suggests this was not a one-off launch phrase but a problem description that still holds. A predecessor product already existed. Akeneo Onboarder let suppliers submit product information directly, but it was limited to simple collection.
Data quality gained a dedicated dashboard. Akeneo introduced Data Quality Insights with PIM 4.0 in February 2020. It assesses enrichment and consistency and highlights spelling and formatting issues. This does not imply that completeness checks or manual quality controls were impossible before then.
These three stumbling blocks share one thing. None of them was a fringe issue affecting a handful of projects. All three appear as the stated reason for a new feature in several independently written Akeneo announcements from different years. That points to recurring problems, not exotic ones.
Now: How Akeneo Has Turned These Problems Into Features
At a glance: which stumbling block is addressed by which feature today, what Akeneo documents about it, and what remains project work regardless.
- Complex data modelling is supported by the Data Architect Agent, introduced in July 2025. It suggests families, attributes and reference entities from a product export. Akeneo describes a reduction from months to days; this is a vendor claim, not a guaranteed project duration. What remains: answer follow-up questions, review the model and approve it before import.
- Supplier data in emails and individual files can be brought together and processed in Supplier Data Manager. According to the documentation, AI mapping suggestions appear only with a confidence score above 70 per cent. This measures confidence in a suggestion, not a measured percentage match. Teams must still review and correct suggestions and configure the workflow for their catalogue.
- Data quality is assessed by Data Quality Insights through enrichment and consistency indicators. AI-Enhanced Enrichment is a separate capability: the October 2025 updates expanded AI enrichment to number and yes/no attributes and technical information from assets. Teams must still define required fields and verify factual accuracy.
Months became days. The Data Architect Agent, launched in July 2025, accepts a product export in CSV or XLSX format, can optionally be given a short context prompt, and automatically generates families, attributes, reference entities (related objects such as brands or materials with their own characteristics) and asset families from it. Users answer the AI's follow-up questions and refine the result before it is imported into the live environment. According to the official product documentation, the language model used is Gemini Flash 2.5. Data sent to Google is said to be neither stored on its servers nor used to train its models. This is a point Swiss companies with their own data protection requirements regularly ask about first. Akeneo itself quantifies the effect as a reduction from months to days.
Email attachments became an automated channel. The Supplier Data Manager, introduced in October 2023 as an evolution of Onboarder, brings supplier data together in a single portal instead of scattered inboxes. Shortly before that, in September 2023, Akeneo had acquired the AI platform Unifai, which specialises in automatically matching, cleaning, categorising and enriching supplier data. In its announcement, Akeneo cited results from customer projects, figures reported only by the vendor itself and not independently verified. At the wholesaler Rexel, manual effort is said to have dropped by 80 per cent and time to market to have been cut to a third. At the sports retailer Intersport, time spent on product data processing is said to have halved. In March 2024, at the Spring release presented at the PX World customer conference, this Unifai technology was fully integrated into the product as "Supplier Data Manager AI", alongside generative AI for product descriptions and AI-powered translation into more than 50 languages.
The product documentation explains the controls in more detail: classification can be automated or reviewed individually; extraction and mapping allow manual corrections. Mapping suggestions are shown only above 70 per cent confidence. This does not prove that data is error-free. Review procedures must reflect business risks and the quality of the source material.
Assessing quality and enriching data are different tasks. Data Quality Insights evaluates product information and highlights quality issues. The generative capabilities expanded in October 2025 belong to AI-Enhanced Enrichment, rather than a new version of the Data Quality Insights dashboard. They can populate number and yes/no attributes and extract technical information from PDFs or other assets. The same update renamed Asset Manager to Akeneo DAM and added AI keyword tagging.
One line, not three separate features. Across all three stumbling blocks, the same sequence repeats: capture data, normalise it, model it, enrich it, check it, approve it, and make it usable per channel. None of the features fully replaces one of these steps. Each one shortens or standardises it. That is what distinguishes this from a plain feature list. The Data Architect Agent models, the Supplier Data Manager captures and cleans, Data Quality Insights checks. The same review-and-approve logic runs through all three, just at a different point in the data flow.
Arcmedia uses Akeneo as its primary PIM platform in customer projects and works as an Akeneo Solution Partner. The specific shape of features such as the Data Architect Agent or AI-powered data enrichment depends on the Akeneo package booked in each case and is not assumed as standard for every project.
Tomorrow: What Akeneo Itself Announces Next
Akeneo publishes a product roadmap describing capabilities and development priorities. It mentions custom prompts in Supplier Data Manager and extraction of product information from assets. A Register Interest form alone proves neither a future release date nor the absence of an available capability.
Check the specific product documentation for availability. Custom classification and extraction prompts are documented in Supplier Data Manager: instructions can be added under Workflow > Settings > Steps. The roadmap alone is therefore insufficient evidence that the function is still in development or that the page is outdated.
Asset extraction also needs to be assessed by data type and scope. In October 2025, Akeneo already documented technical values extracted from PDFs and assets, alongside selection of predefined reference values such as colours or materials. This establishes neither every possible extraction capability nor the blanket unavailability of colour values. Project decisions should rely on current documentation, the purchased package and a test with your own data.
What This Means for Your Akeneo PIM Implementation Today
For a project team, this first means something practical. Before assuming "Akeneo can't do that" or "this will inevitably take months for us", it is worth checking the current release notes, because that assumption has repeatedly proven outdated over the past few years. At the same time, the Data Architect Agent itself shows the limit of this development. According to the official documentation, the result of the automatic modelling is explicitly presented for refinement before import, not adopted directly. The AI delivers a draft, not a decision.
That applies precisely to the stumbling blocks that even the fastest automation cannot solve. Who in the company is responsible for which data field, which department approves a product description, and which system wins in a conflict remain organisational decisions, not a matter of features. An Akeneo PIM implementation that clarifies this responsibility before the project starts benefits more from the new features today than one that leaves it open and hopes the software will quietly sort it out.
Conclusion
Akeneo's release history shows how specific capabilities support data modelling, supplier data and quality assessment. The Data Architect Agent suggests models; Supplier Data Manager consolidates and processes supplier information; Data Quality Insights evaluates enrichment and consistency. AI enrichment is a complementary capability. Automation does not replace expert review or guarantee project duration. Roadmap interest forms are not evidence of availability. Current documentation, package scope and clear ownership remain essential.
FAQ
What is the Akeneo Data Architect Agent?
An AI feature launched in July 2025 that automatically creates a data model, with families, attributes and reference entities, from an uploaded product export. The result can be refined before import. According to the official documentation, the language model used is Gemini Flash 2.5.
What is the difference between Akeneo Onboarder and the Supplier Data Manager?
Onboarder let suppliers submit product information directly through a portal. The Supplier Data Manager, introduced in 2023, builds on that and adds automated AI functions for matching, cleaning, categorising and enriching that data.
Since when has Akeneo PIM had Data Quality Insights?
Since February 2020, introduced with Akeneo PIM 4.0. Data Quality Insights assesses enrichment and consistency with a colour-coded score. AI enrichment, including number and yes/no attributes, belongs to the complementary AI-Enhanced Enrichment capability.
Which AI features does Akeneo announce for the future?
The roadmap mentions Supplier Data Manager prompts and asset extraction. Prompts and several extraction capabilities are already documented. An interest form alone does not reliably distinguish available features from future extensions. Check the exact capability and package scope in the current help documentation.
Does an Akeneo PIM implementation with these features also resolve questions of ownership within the company?
No. Who is responsible for a data field and who grants approval remains an organisational decision. None of the features described here makes that decision on the company's behalf.
Planning an Akeneo PIM implementation?
Arcmedia has been supporting companies with digital commerce projects since 1995, from analysis through implementation to ongoing operations. Our experts are happy to advise you on your project.
Always up to date with the latest trends and tips: subscribe to our newsletter.
Sources
- Akeneo, Serenity updates 2024, overview page (help.akeneo.com)
- Akeneo, press release "Akeneo Introduces First AI Data Architect Agent and New PX Insights Capabilities...", 10 July 2025
- Akeneo, official Data Architect Agent product documentation (help.akeneo.com)
- Akeneo, press release "Akeneo Launches New Supplier Data Manager Offering...", 17 October 2023
- Akeneo, Akeneo Onboarder product documentation (help.akeneo.com)
- Akeneo, press release "Akeneo Acquires AI Platform Unifai...", 19 September 2023
- Akeneo, press release "Akeneo Unveils its Spring 2024 Release...", 5 March 2024
- Akeneo, press release "Akeneo PIM 4.0 Delivers New Product Experience Management Platform", 6 February 2020
- Akeneo, Serenity updates October 2025 (help.akeneo.com)
- Akeneo, official product roadmap (akeneo.com/akeneo-product-roadmap)
- Akeneo, Supplier Data Manager: AI FAQ (help.akeneo.com)
- Akeneo, Supplier Data Manager: AI agent, advanced usage (help.akeneo.com)
- Akeneo, Supplier Data Manager: extraction (help.akeneo.com)
- Akeneo, Data Architect Agent: 2025 updates (help.akeneo.com)
- Akeneo, Supplier Data Manager product page (akeneo.com)
- Akeneo, homepage (akeneo.com)
- Akeneo, "What's new in Akeneo PIM Serenity" overview page (help.akeneo.com)
- Arcmedia, Product Information Management service page
