How Glow25 Is Training 200+ Employees to Become AI Power Users

“My goal is to save people 3-4 hours per week.”

Akos Piffko has spent over a decade in eCommerce, co-founding and exiting multiple DTC brands including AVA & MAY. 

Now as Entrepreneur in Residence at Glow25 — a €100M+ collagen supplement brand — he's leading their company-wide AI transformation.

In this interview, Akos shares his approach to implementing AI across 200+ employees, the tools and frameworks he's using, and how he's turning every team member into an AI power user.

Q: Can you share your background and experience in eCommerce, including the brands you founded and sold?

I've spent 10-12 years in eCommerce, starting at Rocket Internet and FoodSpring. In 2015, I joined Invincible Brands where we were early adopters using data-driven influencer marketing to build three successful brands (Natural Mojo, Hello Body, I AM KAMU). Our centralized holding structure helped us reach €35M turnover with 20% EBITDA in year two, leading to a PE majority acquisition.

Next, I co-founded AVA & MAY, an online-first home wellness brand selling premium scented candles and diffusers. Promising early trajectory was slowed by post-Covid challenges, but the company was recently acquired by Candle Skandinavia.

I then joined Glow25, Europe's leading pro-aging supplement brand, as Entrepreneur in Residence and I lead selected key strategic initiatives including AI implementation.

Q: What does your day-to-day look like as Entrepreneur in Residence leading AI initiatives at Glow25?

I’m working as an individual contributor, which allows me to go deep into specific topics, something that was always difficult as a founder, because you're always putting out fires.

Glow25 is operating profitably at €100 million+ turnover with 150-200 employees. In the first few months, I helped to scope and staff the Business Operations department, which is tasked with ensuring the business engine can support the speed of growth.  

Slowly, I’m moving towards a full-time focus on AI implementation and adoption. 

Q: How are you structuring the AI training and implementation across teams?

1. Centralized guidance, decentralized enablement

To make AI adoption sustainable, I have recruited AI Ambassadors across all departments. They receive in-depth training and serve as the first point of contact within their functions.

While there is an escalation path to the AI Lab, the goal is to equip ambassadors to solve 60–70% of issues independently. Their training covers:

• Advanced prompting techniques
• Identifying and prioritising opportunities through a process-mapping framework
• Building assistants and workflows for tackling department-specific use cases

This structure empowers teams to scale knowledge and ensures AI skills become deeply embedded across the organisation.

2. Central Enterprise AI Platform

Secondly, we decided to implement a centralized enterprise AI solution called LangDock to streamline AI usage across the company. This solution allows users to select the LLM of their choice, build use-case specific assistants and workflows, while also allowing us to implement compliance and best-practise guardrails. LangDock is fully compliant with European regulations, giving us confidence in both data security and governance.

Q: Which use cases are you already deploying across the company?

We’ve built out agents to support a range of use cases across the business:

• Advertorial Writer: Which helps copywriters produce content 35-40% faster 
• Health Claims Checker: Validates legal compliance of any copy and marketing messaging 
• Automated Market Research: Workflows for international expansion (from standardized briefing creation to department specific research insight delivery)
• Pocket Customer: Acting as a digital twin of our ICP so we can always get feedback on what our customers may think 
• Prompt Engineer: To help formulate prompts 

Q: How do you see AI impacting eCommerce brands and what's your ultimate goal?

In the near term, our focus is on internal efficiency and processes. By organizing data better and ensuring teams can provide the right context, we create the foundation for every downstream activity.

The key lever until the end of this year is to increase horizontal efficiency, helping each person save 3–4 hours per week by reducing repetitive tasks. This reclaimed time can be invested in strategic thinking, creativity, and customer experience.

In the medium term, AI will influence how customers discover brands and how search dynamics evolve. I expect to start feeling the impact within the next year. Today, my goal is to make the organization ready for that future. 

In H1 next year I want everyone managing a team of assistants and orchestrating their own workflows. To achieve this vision we’ll need to upskill the team and drive change management initiatives. Overall I believe the future of work at an AI-first company might look quite different from what most of us are used to.