Inside Amazon's Approach to DAM, Metadata & AI

Why DAM Matters More Than Ever in the Age of AI: A Conversation with Amazon Devices

INSIDE AMAZON’S APPROACH TO DIGITAL ASSET MANAGEMENT, METADATA, AND AI

Adapted from an Orange Logic DAMazon community webinar featuring Pauline, Senior Digital Asset Manager at Amazon Devices. Questions have been consolidated and rewritten for clarity, while answers have been lightly edited to preserve Pauline’s original perspective and language.

The challenge: making DAM essential to the AI conversation

As organizations accelerate their adoption of artificial intelligence, the role of digital asset management (DAM) is being challenged and redefined. Software engineers are building new content tools, AI is changing how teams discover and use assets, and questions about the relevance of traditional metadata and governance are becoming increasingly common.

For Pauline, Senior Digital Asset Manager at Amazon Devices, these changes represent an opportunity for DAM practitioners to demonstrate their value in a new way. Rather than operating solely as custodians of digital assets, DAM teams have the opportunity to become essential partners in building the content infrastructure that enterprise AI requires.

In this conversation, Pauline shares her perspective on why metadata matters more than ever, how DAM teams can collaborate with technical stakeholders, and what it takes to make digital asset management an integral part of AI innovation.

Interview with Amazon Devices

Q: As AI becomes more prevalent, are you seeing people question whether traditional DAM is still relevant?

Pauline: I have a little list in my head of things I hear that really stick with me, particularly things that make me think, “Oh, that’s concerning someone said that about DAM.”

One thing I heard in the past year, when I was trying to justify a new DAM purchase, my second instance standing up Orange Logic, was, “Isn’t DAM old-fashioned?” It was questioning the whole idea of why we would need a DAM. I thought that was interesting and something I should really think about.

I think one of the reasons we’re hearing these questions about the fundamentals of DAM is that we have a growing group of stakeholders who are new to our world. They’re software engineers and technical product managers working on internal tools. They’re very smart about building products, but they don’t necessarily know DAM yet.

If they’re building a tool that has to do with their creative pipelines, they’re ready to build an asset repository. It’s just an item on their checklist: “I’ll build this tool, and it’ll have an asset repository.”

What they don’t know yet is the sheer amount of effort and will it takes to maintain an asset library and keep it from descending into chaos or atrophying.

That’s where we come in. They don’t know it yet, but we know it. We’re their best partners because we’ll do the hard, unglamorous work of building asset libraries and maintaining metadata for them.

Q: There’s a growing assumption that AI will eventually eliminate the need for metadata. What’s your perspective?

Pauline: I heard something just in the last week that has already stuck with me. I was talking to some teams about aligning metadata between tools, and someone said, “I want us to get to a point where I just want AI to figure out what product is what. I don’t want us to have to talk about the metadata.”

That one kind of got me.

I think we all know there’s a problem there because you can’t get to that point without the metadata.

Maybe in a future state of AI, metadata becomes more invisible to people once it’s been incorporated into the LLM and can identify some of the visual aspects of our products. But you need that metadata as a foundation.

AI has brought up so many questions about whether metadata is necessary. I think the answer is clearly becoming yes, it is.

Q: How is AI changing the governance challenges that DAM teams need to solve?

Pauline: AI is certainly bringing some new challenges into our work. How do we track what’s been made with AI or the use of synthetic humans? How do we make sure restricted assets aren’t going into the wrong tools?

But I think these are really just extensions of questions we’ve always dealt with through governance.

The biggest challenge is how, with AI, do we square governance with all the product development that’s going on?

This is where we might get a reputation as being a little bit old-fashioned. This world is filled with so much AI hype, and teams are rushing to create and release AI products. Leadership is rewarding them every time they do so.

We don’t want to just be on the sidelines saying, “Have you thought about asset versioning?” or “Sorry, those assets are restricted. You can’t use those.”

We know that’s not what a DAM is. We know it’s a dynamic and transformative space for assets that belongs within the layer of these tooling systems.

Q: What opportunities does this create for DAM teams to become more involved in AI development?

Pauline: I think this is actually a great opportunity.

We know how well DAM can fit into all these conversations about what’s being explored and developed. We have the opportunity to be seen as an essential piece of the technical infrastructure needed to deliver AI tools at scale.

One of our jobs as DAM practitioners has always been to explain DAM to our colleagues and leadership. What is DAM? Why does it matter?

But I think right now we need to make that one of our number one priorities.

I don’t think this is a bad place to be for us in our long-suffering, overlooked discipline. But we’re going to have to do a lot of work and prioritize that work to make sure we have a seat at the table.

We need to show up wherever these conversations about AI tools are happening. We need to start treating tech teams and software engineers like they’re our stakeholders too.

They’re not used to that because they’re used to having stakeholders, but maybe they’ll start to appreciate us at some point.

And we need to make sure our DAMs are ready to scale and ready to be this infrastructure that supports the work we’re going to take on in the future.

Q: How can DAM practitioners demonstrate their value to software engineers and technical product teams?

Pauline: We need to introduce ourselves.

We bring the assets together. We create that infrastructure of metadata that they’re going to need to build their products.

Some of my favorite meetings at the moment are when we get on the topic of metadata and the tech teams start asking me, “Oh, but can your DAM do X, Y, and Z here?”

And I can just say, “Yes. Yeah, it can do that. We already have a solution for that. You don’t need to build that.”

I think it’s so important that we have confidence in our DAM solutions when we’re doing this work.

Q: What role has your partnership with Orange Logic played in supporting that vision?

Pauline: I’ve been with Orange Logic for a number of years and have always appreciated what they’ve enabled us to do.

I think it’s so important that we have confidence in our DAM solutions when we’re doing this work. We need to make sure our DAMs are ready to scale and ready to be the infrastructure that supports what we’re going to take on in the future.

I appreciate the partnership and support of Orange Logic and what they’ve enabled us to do.

The takeaway: DAM’s next chapter is about enabling AI

Pauline’s perspective highlights an important shift in how DAM teams can position themselves within their organizations. As AI accelerates product development and content creation, the need for trusted, structured, governed content becomes more important, not less.

The opportunity for DAM practitioners is to move beyond explaining why asset management matters and demonstrate how existing DAM capabilities can help technical teams build faster, avoid duplicating infrastructure, and deliver AI-powered experiences at scale.

For organizations investing in enterprise AI, the question may no longer be whether they need DAM, but how effectively they can make DAM part of the infrastructure powering what comes next.