"True AI should learn, adapt and improve, not just display data," suggests Stelios Herodotou, Group CTO, Photos Photiades Group

"Many so-called “AI dashboards” and chatbots add-ons in software are simply rule-based alerts dressed in modern UI. True AI should learn, adapt and improve, not just display data," Stelios Herodotou, Group Chief Technology Officer, Photos Photiades Group suggests.

In a recent interview with GOLD magazine, Herodotou also talks about the company's adoption of AI and other new technology and the wider impact on the industry.

What are the specific AI tools that you have adopted? How well did they integrate with your existing infrastructure? Did they require specific customisation to work effectively?

We’ve integrated a diverse range of AI tools into our ecosystem. Some are stand-alone applications, while others are embedded within larger platforms already in use or newly introduced. Our approach wasn’t confined to a specific team or hierarchy. Instead, these tools have become part of our daily operations across various departments and employees.

Key tools currently in use include general platforms such as ChatGPT, Perplexity, Gemini, DeepSeek and Claude, productivity platforms like Tactiq, SlideGPT and Contract Fast Check, and visual generators which are Canva, HeyGen, Sora, DALL·E and Midjourney. To ensure seamless adoption, we invested early in extensive management training and repeated sessions in our Academy.

Additional light customisation was made to some of the tools to reflect our workflows and terminology. These tools didn’t just “plug and play”; many were adapted to align with the guidelines of our Cybersecurity Officer and other document formats and approval flows. Additional API integration is also under development.

Have these AI deployments led to revenue growth, cost savings or improved customer retention? Can you quantify these improvements? And how long will it take before you see a measurable return on your investment?

Quantifying AI’s financial impact is a complicated task. It requires many assumptions, especially when outcomes extend beyond direct revenue. However, we approached this in a simpler way. In June 2023, we started a series of training sessions across the office personnel in our organisation and in January 2024, we conducted an internal survey to assess AI adoption among the office personnel of Photos Photiades Breweries.

The results were revealing: 88% of employees were using AI tools; 15% used them daily (avg. 30 minutes saved per day); 38% weekly (avg. 60 minutes saved); and 46% monthly (avg. 60 minutes saved). By calculating saved hours and applying our average hourly labour cost, we estimated a total productivity gain valued at €93,000.

Against a total investment of about €15,000 in AI licences and development, we achieved an ROI of 520%, or €5.20 returned for every €1 invested. And this was only from initial usage metrics. As AI becomes further embedded in our processes, we expect the ROI to grow much higher.

What have been the biggest bottlenecks in integrating AI solutions? How did you overcome these challenges?

In early 2023, guidance on AI implementation, especially tailored to our industry, was limited. We had to experiment internally to learn correct and effective prompting techniques and overcome concerns about potential errors.

Several months later we received proper training and expert consultation. Another big hurdle we had to overcome was resistance to change. Some employees were hesitant to engage with AI, fearing data disclosure and job displacement. We addressed this directly by launching regular internal training sessions every Friday afternoon, fostering open discussion and positioning AI as a tool to enhance human capabilities, enabling faster and better work rather than replacing skilled personnel.

Technically, we faced data fragmentation mostly on industrial data. Information was spread across several legacy systems and formats, making meaningful AI training difficult. To solve this, we have started working on a centralised data lake with the aim to use unified data governance paving the way for future AI analytics.

As AI continues to transform the workplace, how has the Group adapted? How have employees responded and what steps have you taken to support this transition?

Our adaptation started just 6 months after ChatGPT’s launch. By mid-year, the executive management commissioned a comprehensive assessment of AI’s potential impact, which is fully aligned with our philosophy of continuous growth and innovation.

A cross-functional task force explored the technology and presented findings, which led to formal structured training programmes. Rather than replacing roles, we’ve seen remarkable productivity gains across the organisation. Administrative tasks are increasingly automated, allowing employees to focus on higher-value work. We’ve also introduced new roles, such as AI ambassadors, who identified new productivity opportunities and helped teams implement time-saving workflows.

To support this shift, we have invested in employee upskilling, created an internal AI training and development team and established knowledge-sharing hubs. Transparency and inclusion were key in training sessions. Everyone had a voice in shaping how AI fits into their day-to-day work and we were impressed to see the utilisation by all age ranges.

In your industry, what types of decisions can realistically be entrusted to an algorithm? And do you think that AI will ever be capable of handling the novelty and ambiguity of executive-level decision-making?

AI today is highly effective for operational and tactical decisions, such as inventory rebalancing across our regional warehouses, dynamic pricing models for B2B customers and distribution route optimisation based on weather, fuel costs and order volume.

When it comes to strategic decisions, like market expansion, mergers or brand pivots, AI can provide supporting insights but human judgment remains essential. These decisions often rely on nuance, experience and cultural context, areas where AI still lacks.

Therefore, I see AI evolving into an executive co-pilot, helping leaders stress-test scenarios and surface blind spots. But in the Food & Beverage industry, I don’t foresee it fully replacing C-level decision-making in the next decade.

Which AI trend or technology are you watching closely and which do you think are more marketing fluff than substance?

We’re actively exploring foundation models that process structured and unstructured industrial data. A key initiative, called DataHub, is underway and aims to consolidate brewing, bottling, packing and energy data into a single platform.

Once fully implemented, engineers will soon be able to simulate “what-if” scenarios and spot inefficiencies before they materialise in production. These models are helping us move toward predictive quality assurance and automated root cause analysis.

As for what’s more hype than value? Many so-called “AI dashboards” and chatbots add-ons in software are simply rule-based alerts dressed in modern UI. True AI should learn, adapt and improve, not just display data.

What is the one AI breakthrough that you believe will be a game-changer for your industry?

Generative AI in product development stands out as a future game-changer. As we look ahead, one of the most transformative shifts on the horizon is the evolution of generative AI in product development. We are not just talking about generating content or streamlining workflows anymore. We are talking about AI that can help conceptualise entirely new beverage formulas, drawing from real-time consumer sentiment, emerging health trends and even regional taste preferences across markets like Europe, the Middle East, Asia and beyond.

Imagine a product line where flavours are co-created by AI and our customers, personalised by region, culturally relevant and built with agility from R&D to production. This shifts us from reactive to proactive innovation. Combined with AI-powered supply chains, we’re not just optimising processes; we’re redesigning how products are conceived and delivered.

This interview first appeared in the April edition of GOLD magazine. Click here to view it.