From Order Taking to Operational Visibility

When Delta's digital content operation moved from an on-demand, order-taking model to a structured sprint-based workflow, the way the team talked about its work had to change too. Before, requests were handled mostly on their own. Work got assigned, finished, and published on individual timelines, and stakeholder communication happened mainly between the editor and whoever had asked for the work. The new model needed something different: stakeholders had to see the work as a system. What was in progress? What was coming next? When should they submit a request, and when could they expect it delivered? What had already shipped, and how did any of it ladder up to broader business goals? As communications lead for the new sprint management team, answering those questions landed on me.

The Challenge

There was no reporting process built for the new model. The information to communicate progress already lived in the team's workflow data, but pulling it out, making sense of it, and turning it into something worth reading would have taken a lot of manual work every time. I wanted to try something else: use the data the operation already generates as the foundation for an automated communications and reporting ecosystem. The point wasn't to produce one more status report. It was to make stakeholder communication timely, consistent, targeted, and genuinely useful.

The Approach

I started experimenting with AI-assisted development in Kiro, working from data exported out of Workfront, the platform the team uses to manage its content work. What grew out of that was a set of recurring reports and communications aimed at different audiences. Some go straight to stakeholders whose work just finished or is entering an upcoming sprint, while others give broader visibility to leadership, the Digital Experience organization, or the wider team. The reports don't just reprint the underlying data. AI interprets the information, picks out the relevant metrics and trends, summarizes the work, and wraps narrative context around what it all means. I also set the parameters and requirements that steer how the reports get generated, including:

• What information should be extracted
• How the data should be summarized
• How business priorities and enterprise goals should be reflected
• How communications should be structured for different audiences
• Delta standards for visual presentation and email design
• Editorial and language guidelines
• What information should intentionally be excluded

For example, early versions listed the individual employee names tied to completed work. After looking at the output with teammates, I decided identifying individuals wasn't necessary for what the reporting was for, so I wrote that into the requirements and later reports stayed focused on the work rather than the person who did it. The result wasn't a stack of AI-generated emails. It became a repeatable reporting system with defined rules, editorial judgment, business context, and human oversight baked into the process.

Extending the System Beyond Reporting

The same AI-assisted approach opened up another use: operational dashboards for content and experience components managed through AEM. Some of the data behind these experiences is stored as JSON, and while the raw data holds everything you need to understand what's live, it's miserable to read or compare by eye, especially across multiple environments. I built dashboards that turn that data into visual representations and make it easy to compare environments like staging and production. What used to mean manually parsing and diffing JSON now reads off a dashboard almost immediately.

Impact at a Glance

The Result

The biggest result wasn't the time saved. The new reporting ecosystem gave stakeholders a view into the content operation they'd never had. They can see where their requests stand, what shipped in the last sprint, and what's scheduled for the next one, and they can turn around and use that to brief their own leadership. The communications reached past the immediate audience, too. Leadership several levels above me started noticing the reports, sending feedback, and asking for more information and more capabilities. That told me something that mattered: the reporting wasn't administrative overhead. It was becoming a real source of organizational visibility.

What This Demonstrates

Operational Excellence

Recognizing that a change in operating model requires a corresponding change in the systems and communications that support it.

AI-Assisted Development

Using AI as a development and problem-solving partner to rapidly build practical tools around real operational data.

Automation

Turning recurring manual data compilation and analysis into repeatable workflows.

Stakeholder Enablement

Giving people the information they need to understand progress, anticipate outcomes, and communicate effectively with their own organizations.

Data Storytelling

Moving beyond raw metrics to create reporting that explains what the data means and why it matters.

Governance & Editorial Judgment

Establishing rules for what information should be included, how it should be presented, and what should intentionally be excluded.

The Bigger Lesson

The most valuable part of this work wasn't the technology. The technology let me build fast, but the real work was figuring out what information people needed, why they needed it, how different audiences would use it, and how to turn operational data into a story someone could act on. The reports people now rely on came out of that thinking, not out of the automation. The automation just made it repeatable.

Tools & Technologies