**One Size Fits None! **Operationalizing the Content Supply Chain with Adobe Workfront

My Adobe Summit 2026 presentation at AMA — Ask Me Anything

What an bash at the Adobe Summit! First time on the Summit stage. With audiences interested in how we implemented Adobe workfront for our usecase.

Here’s what I shared.

So, what was broken?

Let’s start at the beginning. We work with 100s of suppliers who send us product content — images, videos, 64 assets types specifically, product data, descriptions, and 400+ metadata per product. Before Workfront, the intake process was… let’s call it ‘creatively decentralized.’

Some suppliers emailed us; filled in custom forms; send via chat threads; custom workflows on monday.com. There was no standard process, no shared visibility, and no consistent way to know whether a submission was complete, correct, or even vaguely usable.

In short: plenty of content — just not always the right content, in the right shape, at the right time.

Beyond intake, we had a quality problem. Not every supplier has the same tooling. Many delivered rich content. But few lacked tooling to submit standardized content. We needed a way to validate and sanitize submissions — a quality gate that didn’t rely on someone manually checking every file.

And then integration complexity, with multiple source systems, multiple custom pipelines etc.

One size did not fit all — unless you enjoy bespoke tailoring at scale.

Enter Workfront

We needed a work management platform to handle content submissions end-to-end: track intake, enforce quality, adapt to supplier-specific requirements, and play nicely with both Creative Cloud and AEM. We leveraged Adobe Workfront.

We standardized on Adobe Workfront as our system of record.

A system of record. Not a system of ‘someone please find that email.’

The event-driven bridge (or: changing engines mid-flight)

Here’s where it gets interesting. We couldn’t just flip a switch and move every supplier to Workfront overnight. We needed a migration path that didn’t require everyone to change at the same time. The solution was an event-driven architecture using Adobe I/O AppBuilder.

Legacy submissions were funnelled into events and delivered to an AIO journalling queue. A runtime action in AppBuilder subscribed to those events and pushed the content into AEM DAM and Dynamic Media. I explained this in my previous article with code references. As suppliers were onboarded to Workfront; each workfront submissions were converted into events and published to the same queue — so the downstream pipeline process them similar to legacy.

We didn’t rip-and-replace. We ran a smooth merge lane until the new system had the traffic. We changed engines mid-flight — and still landed on schedule.

Did it work?

The project took ~10 months to setup Workfront projects, customforms, Fusion scenarios, integrations etc. Post launch, within two months, we were seeing real operational wins. 250+ suppliers were onboarded; workfront onboarding training was seemless; operations were simplified and feedback loop were closed quick.

Nothing builds confidence like fewer follow-ups and more ‘already done.’

Ask Me Anything — the Q&A

This is the part I was most looking forward to. Here’s what the audience asked, and what I actually said.

Q: After onboarding to Workfront, is this a one-time setup, or does it need ongoing developer maintenance?

Honest answer: all software needs some maintenance. We do receive requests to custom reporting, investigating content issues, onboarding new users, access issues, refining procedures. But, most Workfront maintenance are managed by Business practitioners through no-code configuration. Only new integrations, or updates to Fusion workflows, are picked by engineering team.

Verdict: we kept engineers focused on building, not babysitting.

Q: What were your key learnings from using Workfront?

Workfront hit the mark on work management. For Fusion and 3rd party integration, we spent extra time, required more effort than estimated. Fusion development is different from other coding IDE. And yes, we found a few product issues during configuration and duly reported them to Adobe.

Q: How much did you automate? What’s manual vs. automated today?

We launched with zero automation — intentionally. First we ran procedures manually and watched where friction appeared. Then we automated the repeatable stuff via Fusion: data validations, image quality checks, task routing. Today most of the workflow runs automatically. Humans stay in the loop for content approvals workflows. It is the same challenge to adopt AI into workstream, where to place the human in loop. We try, to keep, minimal human intervention and automate as much as possible.

Automate the repeatable. Keep humans where judgment actually matters.

Q: Where do you see AI fitting into your Workfront process?

AI is a major theme right now, we see three strong opportunities:

Generative content: Some suppliers lack creative toolset to produce rich assets or content. AI can generate supplemental content. Plus all content should be standardized to meet Brand expectations.

Quality Checks: Smarter quality checks beyond resolution — to validate relevance, image lighting, accessibility etc.

AI-assisted search: Search the data lake for relavant content using natural language queries.

Today, we’re using AI Assistant for data discovery and custom querying. We’re planning to expand into more agent-based capabilities as those mature.

The goal is better content — not more content, just louder.

The last thing I would like to say..

Standardize the intake. Validate the quality. Automate the repeatable. Sleep better.