Q: How?
A: As a publisher, if you’re still treating AI as an experiment for the photo desk and imaging dept, you’re missing out on stunning new visual styles and losing momentum in capturing audience attention.
The job now isn’t to “try” AI; it’s to operationalise it
For publishers, this means turning text-to-image and AI powered editing into repeatable workflows to expand visual output across all platforms, cut turnaround times whilst maintaining your high levels of quality.
Image creation and editing have their own unique requirements and potentials for harnessing AI powered imaging tools, but how to implement such workflows?
Image creation: what to consider
For image creation, start with clear, low-risk use cases and pick sections where AI art can differentiate the content and stand out, being different to what you typically produce. Think illustration styles and concept art for features, hero images, infographics, even abstract backgrounds.
Choose a limited toolset; there’s a plethora out there but a good place to start would be Adobe Firefly, ChatGPT and Google’s Nano Banana which works best on ‘seed’ images (an input image used as the model’s visual starting point). You are likely to have your own preferred toolset, however it’s advisable to keep up to date with new offerings and what works best for you at the time.
Look for quality / consistency of your house styles in the results, the flexibility of imaging controls and optimally, integrations with existing tools in the organisation for faster onboarding. Start building prompt templates to form stylistic boundaries (types of content, moods, etc) and guardrails.
Define a targeted, lightweight workflow for the first few weeks: editorial writes a brief stating goals, audience and usage. The design team create a range of images using prompt templates. Follow this with editorial review for content and policy, then publish. Track responses, build feedback into the editorial review and fine tune. Store all assets (including prompts and versions) in your DAM.
Labelling these images as ‘AI-generated’ builds trust, draws attention and can turn the image into a talking point.
Plan a pilot for 8 to 12 weeks. The first 2 weeks are setup; toolset choice, prompt library creation, draft usage policy and training. Weeks 3-6, focus on a limited rollout, think a few sections with 10 to 20 images per week and track the metrics. Weeks 7-12, expand the usage, add sections or titles, adjust the prompt libraries and refine policy and audiences.
Metrics to review include: production speed (hours from brief to approved art), revision cycles, cost per final image (including human time) and, most importantly, the engagement lift on these sections; CTR, time on page for long-reads etc.
AI for image editing is already well established, with countless touchpoints throughout the process. The sheer power of AI processing comes to the fore when moving from manual creation to batch automation.
Adobe Photoshop balances creation and editing seamlessly, supported by a powerful range of AI-driven tools from its Firefly suite. Features like generative fill, object removal, and the harmonising tool enable sophisticated retouching with just a click, making AI imaging technology more accessible and approachable for users.
Image editing automation: what to consider
The most obvious benefit of batch automation is the automatic toning of photos, creating images that pop on screen and stand out in print. Teams typically see dramatically faster turnaround times and can redeploy imaging experts to higher value creative work. Additionally, automated background removal tools provide precise results in seconds with an extremely high auto pass rate.
Curate 50-100 images that represent your typical content, include diverse lighting, skin tones and subjects. A Photoshop specialist corrects these to create a ‘gold standard’ target. Tune the automation channels to match these and compare the resulting images on a calibrated monitor, tweaking the settings if needed.
Design the trial over a 6 to 8 week period. The first week is the internal setup and review process; build 3-5 toning channels for print, online and background removal use. Weeks 2-4 for first go-live and review. Process, print and publish online the first tranche of images via automation. Create a review panel (editorial, imaging, quality) for feedback and improvements on colour accuracy, cut-outs and overall reproduction. Weeks 5-8, optimise and expand, add content specific channels (food, fashion, sports etc) with tuned contrast, saturation and sharpening. Optimise output formats for online and production performance.
Short term metrics: turn around time, auto accept rate, engagement and conversion rates increase. Mid term metrics: image throughput growth, editorial workload reduction (manual edits required per image) and cost reductions across the imaging process.
Q: What are your three top tips?
- Start with tightly scoped pilots, then scale what works.
- Invest in people and embrace feedback loops. This is new, fluid and will evolve.
- Define success metrics at the outset and use to guide decisions.
Derek and the other contributors to our AI Special took part in an ‘AI Special – Q&A’ webinar on 18th November. You can watch a recording of the webinar by registering here
For over 25 years, Pixometry’s advanced image enhancement software has been powering the imaging workflows of publishers worldwide. Continuously evolving, our software now incorporates the latest AI imaging technologies to enhance and enrich images, perfect for engaging readers in print and digital media.
Email: derek.milne@pixometry.com
Website: www.pixometry.com
This article was included in the AI Special, published by InPublishing in October 2025. Click here to see the other articles in this special feature.
