Content Platforms

Forrester's assessment of this market is published as Content Platforms, and the name is the first thing to understand, because it replaced Enterprise Content Management over successive editions.

That is a rename doing real descriptive work. Enterprise content management described a discipline: storing, classifying, and governing the documents an organisation accumulates. A content platform describes an environment on which other things are built. Retaining the old name would have been accurate and would also have undersold what the winning vendors are actually selling now.

Forrester's opening framing for the Q1 2025 edition is direct: the enterprise content management market has transformed, and today it is exemplified by AI-enabled cloud content platforms. Generic document management does not cut it.

Read that last clause carefully. Forrester is not saying document management stopped being necessary. It is saying that document management on its own stopped being the purchase.

The reversal the report keeps returning to

The most consequential sentence in the entire evaluation is Cheryl McKinnon's guidance to buyers, stated as an instruction: bring AI to your content rather than bringing your content to AI.

Most organisations are doing the second thing. The familiar pattern is that an employee takes a document, copies it, pastes it into a public generative AI tool, and asks a question. The document has now left every control the organisation built. Its contents may be used to train a model, retained by a vendor, or exposed to the next user who asks the right question.

The content platform vendors are betting the other direction. Because they already hold the content, they can embed the model inside the platform rather than exporting the content to the model.

The architecture that matters is retrieval-augmented generation, and specifically the way these vendors implement it: grounded in the organisation's own repositories, shielded from public models, and respecting the access controls and permission structures the platform already enforces.

The last part is the part to pause on. A general-purpose AI tool has no idea that a given user is allowed to see only three of forty documents in a matter. An assistant embedded in the content platform inherits those permissions by construction, because it reads through the same layer that already mediates every other request.

That is a materially different security posture from the paste-it-into-a-chatbot workflow, and it is the strongest concrete argument for this category existing at all.

The first evaluation set the baseline

The Forrester Wave: Content Platforms, Q1 2023 was the firm's first under the current name, authored by Cheryl McKinnon, and evaluated thirteen providers against twenty six criteria.

That edition named a field dominated by the enterprise content management incumbents and adjacent large platforms: Box, Google, GRM Information Management Services, Hyland, IBM, iManage, Laserfiche, M-Files, Microsoft, Newgen, OpenText, and others completing the thirteen.

The two years between editions are the story. Forrester's own summary is that the pace of innovation over that period was unprecedented, and attributes it to a single cause: generative AI transforming how content is created, consumed, and governed.

The companion Landscape report, published in Q3 2024, set the stage for the Wave, confirming that additional vendors beyond the evaluated dozen are notable for enterprise clients.

Inside The Forrester Wave: Content Platforms, Q1 2025

Published on 30 January 2025 and authored by principal analyst Cheryl McKinnon, the current evaluation scored twelve providers against twenty four criteria across current offering, strategy, and market presence.

Four were named Leaders: Box, Microsoft, Newgen, and OpenText.

Newgen's placement is the most instructive of the four for how this market is moving. The vendor's own summary of the report positions one product, NewgenOne, as a contextual content services platform, and the evaluation recognised it as a Leader on the strength of AI-driven automation applied to content-rich, document-centric processes.

OpenText's placement carries the historical weight of the category, a vendor Forrester credits with moving from a decades-long enterprise content management heritage into AI-powered content management without shedding the governance the old market was built on.

Five Strong Performers followed. Hyland and M-Files both announced Strong Performer placements, with M-Files' positioning built around metadata-driven intelligence and knowledge work automation. Doxis of SER Group also placed among the Strong Performers, a European vendor the report assesses for its intelligent content automation.

Three Contenders completed the field of twelve.

What the vertical argument is actually about

Forrester's buyer guidance devotes a full section to one idea: vertical expertise matters, and not everybody has it.

That is not the usual analyst filler about industry solutions. It is a precise observation about where the value of these platforms concentrates under AI.

A horizontal document management system can store contracts. A vendor that has spent a decade in banking has built the entity extraction that knows what a loan covenant is, the workflow that routes a breach of covenant to the right desk, and the compliance certification that lets a Swiss bank use it.

When generative AI lands on top of a content platform, that vertical depth is what the model grounds on. Embeddings do not help a bank if the platform cannot first identify which documents are credit agreements and which paragraphs are covenants.

The practical consequence for a buyer is that the shortlist should start from the vendors proven in your industry, not from the vendors with the highest aggregate score in a category that averages across every industry.

Forrester also flags that these vendors increasingly ship packaged applications and solution templates for specific industry use cases, which shortens the distance from purchase to value precisely where a customised deployment once stretched into years.

The governance problem is the AI problem

The honest limitation in this category is that the AI in these platforms is only as good as the metadata the organisation already refused to maintain.

The ECM era asked organisations to classify documents, apply retention policies, and structure their repositories. Most did a partial job. The sales team kept a shadow drive. The engineering team versioned by filename. Legal held the canonical copies nobody else could find.

A content platform embedding AI now inherits that state. An assistant grounded in a well-governed repository returns precise, permissioned, citable answers. An assistant grounded in fifteen years of unclassified folders returns confident-sounding errors, and the user cannot tell which they are receiving.

This is the version of the deployment gap that applies here. The vendors cannot fix what the organisation never built, and the AI amplifies rather than compensates for weak information governance.

McKinnon's instruction to bring AI to the content rather than content to AI works only if the content is worth bringing AI to.

Pricing is simplifying and that is not entirely good news

Forrester's guidance carries a warning embedded in a compliment: pricing models are simplifying, but becoming opaque.

A handful of vendors publish pricing for their most common subscription bundles. Most do not, and the standard shape is per-user, per-month with a choice of tiers labelled basic, intermediate, and advanced, with the advanced tier being where the generative AI capabilities live.

Self-hosted options add another layer, with pricing tied to API calls, storage volumes, or application-based parameters.

The risk is specific. The AI capabilities that justify the purchase are in the most expensive tier, the cost of those capabilities is not published, and the quantity that drives the cost, whether tokens, queries, or documents processed, is the quantity the buyer can least accurately estimate before deployment.

That is a budgeting problem disguised as a licensing model problem, and it is worth forcing into the open during evaluation rather than discovering it in the first bill.

Where this leaves a buyer

Three questions decide more than the tier placement.

Do your repositories already have the metadata and permissions a grounded assistant would need? If not, the strongest differentiator of these platforms will underperform for you, and the more valuable investment may be the cleanup that precedes the platform.

Which vendors have actually delivered in your industry, with the certifications and packaged processes to prove it, rather than claiming horizontal flexibility? The vertical question is where these platforms separate fastest.

And can you get an honest, written answer on what the AI tier costs at your projected volume, not at the demo's? The value is real and the pricing is opaque, and those two facts together require unusual diligence.

The category was renamed for a reason. What changed is that these systems stopped being the place content goes to rest and became the place content goes to work, with a model embedded in it that can only be as good as the organisation's own discipline allows.

The unstructured side of that governance problem is scored on its own terms in Document Mining And Analytics Platforms, where Forrester now states the category's price for the first time: about five cents a page at volume.

Forrester does not endorse any vendor named here, and tier placement should not be read as a recommendation to buy.