At Pulse eCommerce Summit ‘25, we hosted a session titled The Future of Commerce - Emerging Trends, Technologies & the State of AI, exploring where commerce is heading and what this evolution means for brands, platforms, and consumers alike. With a strong focus on AI, the session covered everything from the rise of low- and no-code tools and shifts in technical architecture to changing consumer expectations and the broader impact of AI on retail.

Below, we’ll outline the key takeaways from the session, including the rise of generative content, the shift toward agentic AI, the growing importance of LLM visibility, and the early impact of tools like Model Context Protocol - along with some predictions for the year ahead and why brands should adapt.

AI in 2025: From novelty to necessity

Over the past year, AI has progressed from interesting Large Language Models (LLMs) and prompting experiments, with some automated Linkedin posts, into production-ready tools used across industries. We’ve begun moving beyond the hype - and towards tangible business impact.

AI is now maturing across distinct pillars, each evolving at different speeds and creating new commercial opportunities. We’re going to cover a few of the areas changing most rapidly - and those likely to have the biggest impact on eCommerce.

Machine learning and data analysis

Machine learning isn’t new. It's been around for decades, steadily improving and powering many of the tech systems we use - from Google and Facebook Ads, Klevu, Dynamic Yield, Rebuy, and more. These tools analyse behavioral patterns and associated data to optimise ads and product placements, ensuring the right content is served at the right times.

What’s changing now:

• The introduction of LLM commentary and interaction is democratising data, lowering the barrier to people that can interrogate and gain insights from data sets.
• Brands (and individuals within them that are not data scientists) can now gain insights conversationally, rather than relying on deep SQL expertise.
• We’re seeing these applications in platforms like Census, transforming areas such as inventory management, forecasting, benchmarking and international trend analysis.

Generative AI

The past 12 - 18 months have seen explosive growth in generative AI capabilities. Tools that once simply produced decent blog posts can now create production-grade, on-brand written content, images, videos, and even development code.

• Text-based content:
  
  Specialised Generative AI tools like ChatGPT, Gemini, Claude, Jasper, Copy.ai, and EMFAS are being adopted by marketing teams to scale and localise on-site content production and generate marketing copy.
  
  Brands such as Represent and Current Body have used AI to manage content translations and Seraphine have leveraged EMFAS to generate localised product descriptions at scale, while maintaining a newly defined tone-of-voice.

• Visual content:
  
  Midjourney, Adobe Firefly, Canva AI, and Greenroom are unlocking product imagery for eCommerce brands. Marketers can now create lifestyle imagery for paid media and social, A/B test personalised model shots for gender, race, or size, and brands can generate variant colour, pattern, and style changes from product flatlays. This not only saves on expensive photoshoot budgets, but also reduces post-production lead times.
  
  Brands like Hugo Boss, Mango, and Levi’s have been early adopters of AI-generated imagery. H&M, in particular, has pioneered the use of AI-generated ‘digital twins’ of models for marketing campaigns - creating AI replicas of 30 models to generate alternative image content without the need for traditional photoshoots.