Heading 1

The AI landscape just got an upgrade this spring and James Poulter, Head of AI and Innovation, wanted to share what these changes mean for those of us working in brand marketing, digital search, and social media strategy.

‍

1.       OpenAI's New Power Couple: o3 and o4-mini

‍

OpenAI just dropped what amounts to AI with actual working memory and reasoning skills. Think of it as upgrading from "smart but forgetful intern" to "strategic thinking partner."

They can now search the web, analyse images, and work with your files in real time while thinking.

‍

o3 is the premium version which is pricier but worth it for complex needs.

o4-mini is the budget-friendly alternative that's surprisingly capable.

‍

For marketers, this means:

·     You can now ask it to "analyse these five competitors landing pages, find visual trends and suggest design improvements for our campaign" in one go

·     It can look at your social analytics, spot patterns and recommend content strategy adjustments

·     Visual content analysis is significantly better

‍

These models respond well to "think step by step"instructions, and love when you tell them to "search for the latest information on [topic]" when recency matters.

‍

2.       Google's Gemini 2.5 Pro: The Methodical Thinker

‍

Gemini has evolved from "search with personality" to "researcher who never sleeps." Its huge context window and reasoning skills make it particularly useful for market research.

‍

For marketing teams:

·     Feed it entire campaign briefs, competitor decks and market research to get comprehensive strategy feedback

·     It's excellent at analysing video content - try feeding it your YouTube ads alongside competitors'

·     Ask it to watch product demos and identify messaging improvement opportunities

‍

Gemini 2.5 works best when you give it a specific role: "Act as our brand strategist and analyse these market research PDFs"gets better results than generic requests.

‍

3.       Meta's Llama 4: The Accessible Powerhouse

‍

Llama 4 is Meta's answer to "How do we make AI that's both powerful AND accessible?"

‍

Why marketers should care:

·     First version of Llama that truly understands images alongside text

·     Significantly better multilingual capabilities(12 languages natively)

·     Can be deployed on your own infrastructure for brand sensitive work

·     Handles large documents (10M tokens = roughly 8,000 pages)

‍

Llama 4 shines with explicitly formatted prompts. Meta released a new guidance document suggesting you start with "You are an expert conversationalist who responds to the best of your ability" for more natural interactions.

So, what does this mean for your workflow?

‍

Regardless of which AI ecosystem you prefer, here's how to capitalise on these advancements:

‍

·     Stop treating AI just as a content creator – these models can now analyse and strategise across multiple inputs

·     Multimodal is the new normal – start incorporating images, charts and screenshots into your prompts for richer insights

·     Let AI do the heavy data lifting – all these models can now process more context, so feed them those big campaign briefs, audience research docs and analytics exports

·     Experiment with search and reasoning – the ability to search the web and then think about the results means you can get much more current insights

‍

JP is excited about how these improvements will change collaborative workflows between creatives, strategists, and analysts. The ability to process visual content alongside performance data creates new possibilities for campaign optimisation.

‍