# Thomas Ladefoged

Thomas Ladefoged speaks and writes about e-commerce content production, including how teams can use AI-generated imagery while maintaining product accuracy, creative quality, and consistency. Ladefoged's posts identify them with the Pixelz team.

- Profile: https://b2binfluencer.marketing/creators/thomas-ladefoged/
- LinkedIn: https://www.linkedin.com/in/thomasladefoged
- LinkedIn followers: 5,895
- Topics: artificial-intelligence, ecommerce
- Based on 11 dated posts, latest 2026-10-07

## Summary

Thomas Ladefoged is useful for production-focused talks on making AI product imagery accurate, polished, consistent, and ready for e-commerce.

- Best for: Content-production or retail-technology brands needing keynotes or webinars on AI product imagery workflows.
- Not for: Consumer campaigns needing shoppable social content and reach through an existing consumer following.
- Check first: Ask for past registration data, post-session survey results, and LinkedIn follower demographics.

### Takeaways

- Generating an image and having one ready to sell are two different jobs.
- AI images may need lighting, composition, product color, texture, and size attention after generation.
- A product-detail-page session drew questions on specifying, generating, delivering AI content and creative control.

## Main ideas

- **Generating an image and having one ready to sell are two different jobs**: Thomas Ladefoged argues that weak AI product images usually fail after generation, not during it. Successful images get further work: lighting adjusted, composition corrected, product colour, texture and size made accurate, and smaller retouching choices, such as adding grain. Weaker images show strange proportions, products that look slightly wrong, illogical lighting or odd posing. Pixelz built its history on work that starts once the image exists — retouching, quality, automation and consistency. AI has moved the difficult part earlier in the production process without changing what a finished e-commerce image has to do. Sometimes the answer is to regenerate; sometimes it is good post-production.
- **Let the audience's questions set the next session**: The volume and specificity of audience questions set the agenda for the follow-up, rather than another introductory talk. The May session raised practical questions about how AI-generated content is specified, generated and delivered; how teams maintain quality and product accuracy; and where creative control sits between AI, teams and brands. Questions outpaced the time available, so the second session was built around them, with a platform demonstration and an invitation to suggest further topics in the comments beforehand.
- **Ask the industry instead of guessing where it stands**: No two conversations about AI sound alike: some people are optimistic, some skeptical; some teams are already producing AI content, some are testing it, and some are watching from the sidelines. Thomas Ladefoged argues against assuming where any given team stands. Instead, they call for a practitioner survey on AI in e-commerce content creation, asking how teams use it, where they find value and what their biggest challenges are. Answers will be anonymous, with results promised in that year's E-commerce Visual Trends Report. The aim is to separate what is genuinely useful from what is simply generating headlines, rather than argue about AI as a category.
- **Revisit last year's predictions in public**: Their conference format looks backwards as well as forwards. At FLOW: London, Thomas Ladefoged opens the day with a keynote reviewing the major shifts of the past twelve months, revisiting the previous year's trends and predictions, and taking stock of where the industry stands. They then stay on stage to host much of the rest of the programme. The same emphasis on practical reckoning appears in a Creative Summit Europe post: the challenge is not listing AI's possibilities but making them work at scale, and comparing notes with other practitioners on what is and is not working.
- **Four questions run through the work: how to implement, scale, control quality and choose what to pursue**: The same operational questions recur across the posts: How do you implement AI in an existing production process? How do you scale it? How do you keep quality and consistency under control? And which new possibilities are worth pursuing? The product-detail-page webinar, industry survey and conference agenda are each organised around those questions rather than a product pitch, making them an agenda for discussion rather than a set of answers.

## Work to read

- [There is a lot of AI content out there right now](https://www.linkedin.com/posts/thomasladefoged_there-is-a-lot-of-ai-content-out-there-right-activity-7496113668871417858-lK9j)
- [Back in May, Pamela and Carlos gave a first look at how we're bringing AI assets to the PDP](https://www.linkedin.com/posts/thomasladefoged_back-in-may-pamela-and-carlos-gave-a-first-activity-7480554090519838721-Mdtl)
- [Love AI? Hate it? We want to know!](https://www.linkedin.com/posts/thomasladefoged_love-ai-hate-it-we-want-to-know-lately-activity-7475873132944265216-5SZq)
- [The FLOW: London agenda is out! And yes, you’ll see my name on there](https://www.linkedin.com/posts/thomasladefoged_the-flow-london-agenda-is-out-and-yes-activity-7513221168561152000-HUsL)
- [I’ll be joining the Creative Summit Europe this November](https://www.linkedin.com/posts/thomasladefoged_ill-be-joining-the-creative-summit-europe-activity-7505272744381542401-KJx8)

## Brand relationships

- **Pixelz** (employment)
- **FLOW: London** (mention)
- **Creative Summit Europe** (mention)

## Partnerships

The material reviewed shows a working pattern of appearances and joint content: a conference keynote with a stage-hosting role, conference attendance with the Pixelz team, question-led webinars, a customer survey feeding a published trends report, and a technical blog post on AI image post-production. For a buyer planning a webinar or content collaboration, sessions are built around questions gathered from creative and e-commerce teams, and event work is tied to dates announced in advance. Ask about pricing, deliverables, exclusivity, and disclosure terms.

## Alternatives

- Katrine Rasmussen: Pick Katrine Rasmussen when the deliverable is a room or an event slot rather than a single talk. The posts show event leadership: FLOW: London sold out more than six weeks before the date, with an agenda built from keynotes, workshops, demos and eight roundtable conversations. They also show a stage appearance at Visual 1st in San Francisco on a panel about GenAI taking over parts of the traditional photoshoot, alongside Caimera, Cloudinary and Titles. If the plan is to convene retail and creative-operations buyers or attach a brand to an existing event community, that is closer to what these posts describe than a standalone keynote or customer webinar.
- Richard Fox: Pick Richard Fox when the audience's obstacle is rollout rather than image quality: making AI predictable enough to run in production. Richard Fox's posts argue that AI should become 'boring' — with clear inputs, defined use cases, review, post-production and guardrails — so output can be repeated at scale. They describe workflows built into Pixelz AI Studio and a walkthrough during E-commerce Content Creation Week, and present an AI use case matrix that weighs each opportunity against what it takes to introduce it, making the case for smaller use cases that fit existing work. Strongest for a session aimed at internal scepticism or a stalled adoption programme.
- Greg Weyman: Pick Greg Weyman when the need is wider than AI imagery: a product-content system that cannot keep up with demand. Greg Weyman's posts look back on founding MarketThrive in 2018 after leading e-commerce for a manufacturer with thousands of SKUs, and report patterns recurring across more than 80 brands: resistance to change even when everyone knows the current way is failing; acquisitions splitting standards and systems; product data scattered across engineering and marketing; and change starting with one product pilot rather than a transformation. Best suited to a session on modernizing product content operations where AI is one part of the answer, not the headline.

---
AI-assisted summary of public posts. Audience demographics are not available. Engagement appears only where we checked recent posts. Method: https://b2binfluencer.marketing/methodology/