Last year I worked with a newsroom on automating a part of their AI workflow. Going into it, I thought I knew exactly what the engagement would look like. They needed a technical fix: automate payment failure notifications and expiration reminders so their one-person membership team could focus on the million other things that it takes to build and engage a community.
However, what I did not expect was that a few exploratory conversations would reveal a much richer picture—one that led to a diagnostic memo I hadn't planned to write, and a clearer sense of what AI can (and can't) do for under-resourced newsrooms.
Why I like to talk
Before we jumped into conversations about the automation, I asked broader questions about their membership strategy, content workflow, and decision-making processes. I just wanted to have the broader context in which this automation would work. But this technical scoping turned into a revealing conversation about capacity, ambition, and the invisible friction points in their operation.
The team had a clear vision for growth, but resource constraints meant they were constantly choosing between publishing, member engagement, and strategic planning. They were collecting valuable member feedback and engagement metrics but lacked capacity to analyze it systematically or integrate insights into decisions. Without consistent frameworks or KPIs, membership decisions were being made reactively rather than strategically.
I realized the automation request was a symptom of a larger pattern. So I did something I hadn't planned: I wrote a diagnostic memo.
The memo nobody asked for
Look, I'm aware of how this sounds. "Consultant shows up to fix one thing, decides to lecture others on how to do everything else".
But the things is, the patterns were so clear that not writing them down felt like a missed opportunity. I mean what else am I to do with my time and the copious notes I sometimes take? So I spent an evening turning our conversations into a memo that reflected on what I had heard, clarified what the observations might mean, and suggested... wait for it, both AI-enabled and non-AI interventions for each pain point.
Revolutionary, I know.
A few examples of what emerged:
“Capacity to analyse reader revenue”
What this could mean: Data exists but is underutilized due to limited time or analytical expertise. Reader revenue opportunities may be hidden in existing datasets.
Possible interventions:
- AI-enabled: data analysis, recommendations, and trend detection for open feedback like comments, emails, surveys.
- Non-AI: define your KPIs, why they matter, and refine how you collect those. Focus on max 2 KPI for at least two quarters to test different approaches.
“Would also like to use AI for brainstorming”
What this could mean: The team is open to experimenting with AI, starting from low-risk creative tasks.
Possible interventions:
- AI-enabled: brainstorming partner for campaigns, benefit design, comms drafts with guardrails.
- Non-AI: human review and approval process, clear style/tone guidelines.
...and so on.
What this process reconfirmed for me was that AI can help, but it's rarely the whole answer. Often, the non-AI intervention, such as defining clear frameworks, simplifying processes, is more impactful.
I promise this isn't some elaborate ploy to upsell strategy consulting. Sometimes a Google Doc with clear KPIs does more than the fanciest automation.
What this taught me about AI in Newsrooms
This experience reinforced three convictions we have about AI in journalism:
1. Conversation before prescription
I could have shown up with a toolkit of AI solutions. Instead, I showed up with curiosity. And curiosity is not killing this cat! The diagnostic memo emerged from that. That's how you find interventions that actually matter.
Also, let's be honest, asking questions makes you look smart while buying time to figure out what you're actually going to do.
2. AI doesn't have to be all-or-nothing
You don't have to use AI to convert your content into 6 different formats for 3 different audiences across 4 different platforms. Just a simple automation that saves your team time so they can focus on audience-first journalism can be sufficient.
3. Sometimes the best intervention isn't AI at all
Several valuable recommendations in the memo have nothing to do with technology:
- Define your KPIs and stick with them for two quarters
- Create a Membership Operating System document
- Build an experimentation backlog with clear criteria for scaling or killing initiatives
These are foundational practices. AI can accelerate them, but it can't replace them.
For this newsroom, I did build a Make a Make automation workflow that is triggered when a payment fails or a membership nears expiration. Instead of replacing human judgment, the system generates a draft email in the membership manager's inbox. From there it is up to the person behind the desk to hit send, ignore, or edit. But what I also gave them is the Diagnostic Memo, because, again, it would have been a waste not to. To do anything with it or not is up to them but much like what I say about using AI to make your data actionable, I hope to have given them a tool for future decision making.
A word of advise for newsrooms
Start with these questions:
- What repetitive tasks are draining your team's capacity? Start there with simple automation.
- What data do you already have that you're not using? AI can help analyze it, but first define what you want to know from that data.
- What decisions feel like guesswork right now? Build the frameworks first, then add AI-powered insights.
And if you're talking to consultants, look for people who listen before they prescribe. The best solutions emerge from understanding your specific context, not from a pre-packaged playbook.
*Yes, I'm aware I just spent 937 words describing my own consulting approach. The irony is not lost on me.
At First Draft Media, we believe in meeting newsrooms where they are. Whether it's building automation workflows, developing strategic frameworks, or identifying opportunities for AI-light interventions, our approach starts with listening. If you're interested in exploring how we can support your newsroom, get in touch.