On the Horizon:
The Next Wave of AI Tech 

Oliver Feldwick
Chief Innovation Officer
T&P   

It’s a fascinating moment because, for all the hype, most of what we're seeing isn't revolutionary yet; it's evolutionary. 

We're still in the very early days, using powerful new tools to do old things a little bit better. 

This abundance of intelligent capability needs the right scaffolding and thinking to be truly transformative. As the computer scientist Grady Booch famously said,

"A fool with a tool is still a fool." 

The most impactful development right now isn't any single tool, but the way that we’re all experimenting and applying these tools to our processes. 

We must ask what we want to achieve before we start prompting. The foundation models themselves are already incredibly powerful partners in this. With their ability to reason, refer to specific documents, and conduct their own research, they can help across the entire strategic process; from synthesizing vast amounts of research and expanding our ideation into new territories, to crafting narratives and even acting as a critical mirror to reflect on the quality of our own thinking. 

These knowledge tools already have proven potential to deliver value – but we’re in the messy moment of understanding how to use them consistently and meaningfully in our processes. They have the potential to allow us to operate at a greater velocity and incorporate ever more variance in the avenues we can explore. 

Beyond language, the latest image and video models -whether you're using Nanobanana, Flux, Qwen, or Stability - are starting to move beyond just accelerating concept visualisation into creating final ‘production quality’ assets. 

This isn't a simple efficiency gain; it's a new technique that opens opportunities for greater speed, experimentation, and creative optionality. Like the early days of computer animation, these tools have their own distinct aesthetic and limitations that we are learning to work with and around. This leads to the true frontier: agentic tools and workflows. Whether it's building simple automations in n8n or exploring the agentic frameworks from OpenAI, Google, and Anthropic using protocols like MCP, we are only at the tip of the iceberg of their potential. 

All these developments - reasoning, generative content, and agents - will deliver genuine improvements by allowing us to experiment with greater velocity and variance, covering more ground and creating in compelling new ways. 

The limiting factor in the next 12-18 months won't be what's possible with the tools; it will be how we coordinate ourselves and our organisations to get the most out of them. 

Gerard Crichlow
SVP global social strategy director
McCann

At McCann, AI has genuinely supercharged how we think, strategise and get to more interesting work.  As a strategy team, we’re using AI in three ways:  

Truth Hunting: We have a proprietary AI tool that has supercharged how we Truth hunt using our Truth to Impact process.  We can accelerate our cultural intelligence by scanning culture in real-time, not just trends, but emerging communities, shifting cultural norms and identifying moments where brands can authentically participate and add value. This helps us get to a singular truth to guide our creative teams, inform our connections plans and inform our influencer approaches.   

Social Audit: In the past doing a social audit was a tedious process that required hours of searching feeds. AI has shortened how we audit a brand’s social presence, its competitors and its brand values to develop a unique social voice that is as distinctive as its logo, sonic sound and brand colours. We have a better handle on which cultural conversations a brand should participate in, the communities that matter and a point of view on the type of content that resonates.   

Sparring Partner: If we’re doing our jobs, we’re pushing boundaries. Getting clients to come along the journey is part of the art of selling an idea in. As a team we’ve experimented with creating AI agents to help us anticipate what a client’s feedback might be ahead of a tissue session or prior to reviewing creative work, to help us sell in better ideas and rid ourselves of level 1 ideas faster and get to level three ideas that matter.  These agents are not perfect, but they help us see around corners we couldn’t see ourselves because better questions lead to better answers.   

Kahmen Lai
Senior Vice President, Integrated Media Strategy
Weber Shandwick UK 

We tend to repeat the reassuring mantra that AI is for efficiency and humans are for creativity, but in reality, the boundaries are blurry.

The relationship between human and machine has already shifted from directive to collaborative and is quickly moving toward supervisory.  

AI’s biggest contribution so far is democratising both data and early creativity; people who once relied on specialists to extract insights, build concepts, or visualise a first draft can now do much of this themselves, and do it quickly. That speed is meaningful. It compresses timelines, accelerates approvals, and gives teams more time to refine the quality of ideas rather than contend with process.  

It isn’t always perfect, and sometimes it’s not even very good. But it’s undeniably faster and easier to get started than it used to be.  

Today, we utilise a combination of AI tools to accelerate workflows, including our proprietary platform, HALO. HALO is a secure GenAI workspace built on Google’s infrastructure and incorporates data partners like Blackbird, Infegy, Memo, Morning Consult, Peak Metrics, and Talkwalker. It is uniquely differentiated by being trained with the "Weber brain", Weber Shandwick’s proprietary frameworks, methodologies and strategic knowledge, making it tailored to the specific needs of our agency, employees and clients. This exclusive setup sets HALO apart from external platforms. 

We’ve also partnered with Google to create a custom, agentic AI platform for our teams, further enhancing our capabilities. HALO is integral to workflows that include: 

• Insights generation: industry overviews, social listening, competitive mapping, audience segments, and market positioning 
• Message development and testing: personas, message frameworks, supporting proof, and reaction modelling 
• Creative ideation: initial concepts, reference images, mock-ups, and AI-augmented design tools 
• Content development: first drafts, audience-specific iterations, and localisation 
• Issues preparation: scenario planning and response prep 
• Admin: meeting notes, summaries, and action lists 

By integrating HALO with our strategic knowledge and proprietary methods, we deliver tailored, impactful solutions that elevate both the agency and client outcomes far beyond the capabilities of general AI platforms. 

To use GenAI effectively, you must understand the strengths of each tool and how to move between them to achieve your objective. This has become a skill in its own right. And while that orchestration matters, human intelligence remains the first and final word.  

Looking ahead to the next 12-18 months, the shift to agentic AI will reshape workflows, autonomously monitoring the environment, sequencing tasks, and optimising outputs with minimal prompting. Much of the heavy lift from workflows like media monitoring, crisis communications and performance optimisation will be agent-led. And that will leave humans with more time to spend on the distinctly human work of relationships, discernment, negotiation, and nuance. 

Ben Edwards
Executive Creative Director
AMV BBDO  

We’ve never had so little time to make things. Briefs are tighter, budgets thinner, competition fiercer, yet expectations for brilliance have never been higher. It’s the perfect storm for mediocrity, unless you find ways to move fast without flattening the work. 

AI shouldn’t be seen as a replacement for creativity, rather, a protector of it.

Tools like Sora and Veo are already redefining what’s possible in motion. Seedream and Nano Banana are pushing image generation into entirely new visual languages. Higgsfield is reimagining performance and 3D realism. And most exciting of all, Aaru is starting to shape how we work from the very beginning – helping at the strategy and brief stage, and changing how we research and defend great ideas before they’re even made. 

Because speed alone is useless. The goal is to combine speed with craft to give us more time for beauty, rigour, and the human instinct that turns clever into unforgettable. 

AI isn’t here to do our jobs, rather give us back the time to do them properly. Let’s just hope it doesn’t accelerate expectations too much – unlike the machine, we still need sleep. 

Jasleen Carroll
Director of AI and Product Alex Calder
Head of Education and Innovation
Anything Is Possible 

The most genuine enhancements from AI right now are less about generating a single award-winning video and more about systematically improving efficiency. For our teams, the biggest gains come from AI-powered media planning and programmatic buying tools that handle the heavy lifting of data analysis and real-time optimisation.  

This frees up strategists to focus on the bigger picture. On the creative side, we treat AI as a co-pilot; it’s brilliant for generating dozens of ad variations for testing, but the core brand storytelling remains a human skill.  

Most people engage with applications like ChatGPT, Gemini and Claude so what we are seeing is more AI generated content that is common and generic.

To really stand out and enhance your AI outputs you need a combination of LLM (AI), Prompts (well formatted questions) and proprietary data. That is the formula for embedding AI successfully.  

We’re now moving beyond prompt-based tools to truly agentic systems - AI that can browse, reason, and act. OpenAI’s new browser (ChatGPT Atlas) capabilities, native Stripe integration, and agent workflows, alongside platforms like Perplexity’s browser, Comet, signal a shift to AI that can research, plan, and even execute campaign tasks in real time. 

This evolution will transform creativity from linear to living ideas that update dynamically based on live data or audience feedback. In practice, that means faster iteration, more personalised storytelling, and more time freed up for human judgment and strategy. Over the next year, we’ll see AI become less of a tool and more of a creative collaborator - orchestrating media, crafting copy, and surfacing insights instantly. The future of creativity isn’t AI replacing people; it’s AI releasing people to think bigger and move faster 

Major platforms are aiming to have AI generate entire campaigns from a simple product image and a budget by 2026. This makes the most critical job for leaders defining the human:AI ratio: deciding where human creativity and strategic oversight can make the most impact. 

Liz Duff
Head of Commercial and Operations
Mediaplus UK  

The AI revolution presents both unprecedented opportunity and existential risk for media agencies around the world.  

Rather than pursuing every emerging tool, we're taking a more focused approach: how to strategically integrate AI to enhance human creativity while preparing for fundamental industry transformation. We're experimenting with AI-assisted brainstorming and strategic ideation, exploring how these tools can serve as intelligent collaborators to expand creative thinking and to deliver results in terms of both efficiency gains and enhanced creative exploration.  

The commoditisation threat is becoming increasingly apparent. As AI capabilities expand to handle programmatic execution and campaign optimisation, we recognise that the agency value proposition must evolve beyond transactional efficiency. This realisation is driving our emphasis on "AI collaboration competencies" – developing the uniquely human skills that will become more valuable as AI handles routine tasks.  

We’re envisaging an evolution of our talent development strategy toward skill mobility frameworks, anticipating an agency landscape where traditional role boundaries become more fluid. We're passionate about preparing our teams for a future where success depends on seamlessly blending human judgment with AI capability - learning to frame better questions, model complex scenarios, and understand both AI's potential and limitations.  

Looking ahead 12-18 months, we expect significant advances in audience behavioural analysis and real-time campaign adaptation. However, we believe the agencies that will thrive are those positioning relationships as the ultimate differentiator. While others focus purely on tool acquisition, we're investing in building the next generation of media professionals: strategically minded, creatively empowered, and increasingly fluent in human-AI collaboration.  

Callum McCahon
Chief Strategy Officer
Born Social  

I see AI as leverage. It's a useful multiplier for good thinkers and creatives, not a replacement for them. It's most useful at the bookends of the creative process: the messy beginning and the executional end.

At the start, AI helps with research, synthesis, and exploration. We're using it to work through large amounts of data, generate initial ideas quickly, storyboard concepts, and test different audience perspectives. Training GPT models on specific audience insights lets us sense-check thinking before we commit resources. It's helpful for getting unstuck or exploring alternatives. 

At the finish, it's about scale and variation. Once you've got a winning creative idea, AI can help adapt it across different touchpoints. Social variants, format adaptations, personalised versions, all without losing the core concept. It adds breadth and depth that would otherwise require significantly more time and resources. 

Looking ahead 12 to 18 months, real-time adaptive creative and AI-driven media optimisation seem promising. The ability to generate and test creative variants based on performance data could be genuinely useful.

But the fundamental remains: AI supports good creativity. It doesn't replace it. 

Jim McGorty
Creative Director
onepointfive 

The tools we use such as Midjourney and ChatGPT, while useful, are only as powerful as the minds behind them. Experiential thrives on real-world connection and at onepointfive, we actively encourage our team to unplug, explore, and seek inspiration from unexpected places. Our creative processes are designed to draw inspiration from outside industries, promoting cross-sector thinking and unconventional sources to unlock fresh ideas. That’s how we keep our work unique from the inside out. 

As certain creative processes become more streamlined, I see space opening up for a quiet revival of analogue methods like sketching, journaling, and tech-free collaboration. These slower, tactile practices encourage deeper thinking and more original work, allowing creative teams to reconnect with the essence of ideation beyond screens and automatio

Secondly, as algorithms increasingly shape creative outputs, the value of human intuition is being reasserted. Drawing on experience, original thinking, gut instinct, and cultural nuance will become the real currency for generating ideas and steering machine-led processes. In fact, these human qualities now represent the only real intellectual property an agency can own. This shift signals a renewed appreciation for the creative mind as a compass in an automated landscape. 

Finally, as AI speeds up idea generation, the risk of creative copycatting is rising. Agencies will need sharper, more proactive IP protection, legally and operationally. Expect smarter frameworks, watermarking tools, and processes that safeguard originality and ensure creative ownership stays crystal clear. 

At the end of the day without instinct, protected time to think, and smarter IP safeguards, how can any agency truly compete or claim ownership of original work? 

 

Troy Farnworth
Executive Creative Director
Leith 

Most AI tools can enhance creativity if used in the correct way. Our approach is fairly platform agnostic. What is more important to us is why we are using it. There are two pieces about creativity that ring true to me during these disruptive years.  

The first is Alfred Hitchcock’s naivety of what was possible with a camera. As an early pioneer of invisible cuts, moving sets, and one-shots, he completely transformed cinema. But he attributes much of this to his imagination and the freedom of not knowing what was possible. As such, his approach was if he can imagine it, it surely can be possible. It led to him innovating cameras and techniques to create what he wanted to achieve. So, for us, does AI help us expand what is possible? Can we do things that the budget would not allow us to do? Can we do something that’s simply not possible? Our motives behind the use of AI is, can it improve or push us forward. As with Hitchcock, the most important thing is what you are trying to achieve. What is the vision?   

The second piece is an interview with Steve Jobs about the process of making a film at Pixar. He talks about how expensive it is to create animation, so they must know it’s right before they actually make it, unlike in film where you shoot lots of stuff and create the film in the edit. Storyboards were the way Pixar tested the story as they go. Animatics of what they would eventually animate for the main film. Steve openly says things they thought would work just didn’t, so they stopped the process and fixed it before they moved to the next scene. AI has allowed this process to be adopted within agencies at a much lower cost and much quicker than ever before. You can now board and block out adverts, test, tweak and go again until you hone both the message and the engagement. AI can’t write the story for you, but it can help you evaluate it.   

I feel we’re only just getting started with this technology.

Creatives and agencies that use AI to enhance their vision will be the ones that will win. The ones that think AI is the vision will fail.

A great camera or technique doesn’t make you Hitchcock. 

Adam Cleaver
Founding Partner and Executive Creative Director
Collective 

Obviously, everyone is obsessed with the next big tools, the next ‘this changes everything’. But let’s face it, they don’t. Gains are more incremental when used in the real world of advertising.  

We use many of the usual suspects: ChatGPT, Midjourney, Runway, Firefly, DaVinci, ElevenLabs, Templater. They help us explore, visualise ideas and take care of some of the donkey work in production and post, versioning, localisation, voice-over and polishing things that used to take days. But they haven’t rewritten creativity; they’ve just made parts of it faster.  

Where we’re seeing some real step-changes is in tools that allow tailoring, such as custom GPTs, AI assistants or trained models built for specific roles. Quietly they’ve changed the rhythm of how we work, allowing us to query vast amounts of data, test thinking faster and, in some cases, save days or even weeks.  

And tools like Intangible.ai start to hint at what’s next, the ability to move fluidly between idea, copy, imagery and motion in one place. It's all about more joined-up systems.  

That’s where the next 12 to 18 months will go. Yes, more tools, but more importantly better integration. Linking AI into proper workflows that deliver faster and better without losing quality.  

We’re moving into an orchestration phase: creating systems that bring structure to the chaos. Systems that deliver consistent outputs, reusable assets, connected data, and seamless handovers between teams and stages. That’s what we’re helping our clients do.  

As this happens, the question of governance and ethics will matter more, not less. As ever, the responsibility sits with us to make sure the work remains original, safe and human.  

Overall, we’ll see AI start to disappear into the process. Pretty much invisible but useful and part of the craft. Which is exactly where it belongs. 

Alistair Hornsby
Strategist
Craft Media London 

Craft Media is a media strategy consultancy. Our edge relies on the quality of our thinking. We use and build AI co-pilots to handle long, procedural tasks—freeing time for strategic thinking and collaboration.

We never use AI outputs as final recommendations; our team is trained to know how and when to use AI effectively. Broadly, we use it to explore, synthesise, analyse, and ideate.

Our work starts with deep research into category, competitor, consumer, and cultural contexts. Alongside this, we run explorative searches using ChatGPT’s Deep Research, Claude’s Extended Thinking, and Perplexity. These surface sources we might miss and speed up understanding of brand contexts. Meeting transcribers like TL;DV or Gemini capture long client conversations, adding valuable context. 

To spread new insights across the team, we synthesise them using custom LLMs—downloading studies from publishers like the IPA, running them through AI models to extract key insights, and comparing them with stored knowledge. We’re heavy users of audience tools, building custom AI systems to analyse data and uncover audience nuances quickly. For example, on a beauty brand project, we analysed 20 segments across five markets in under 72 hours—freeing more time for strategic recommendations. We also use chatbots trained on qualitative data to explore niche groups such as HNWIs and small business owners. 

As we develop hypotheses, we use LLMs for critical assessment—asking them to identify weaknesses in our ideas to make our work more rigorous. 

We see four themes shaping the next 12–18 months. 

1. Vibe-coding: tools letting non-specialists build digital products and creative work that feels cinematic, widening creative potential while keeping craft essential. 
2. Efficiency gains: AI enables creative and media teams to manage complex, multi-market campaigns with smaller teams, producing quality work faster. 
3. Search upheaval: as AI summaries replace search clicks, Generative Engine Optimisation (GEO) will surge. Good content and genuine community engagement will matter most. 

4. Analog nostalgia: as AI fakes reality better, people will value real experiences—events, human interaction, and crafted creative. 

Long term, AI may mirror farming’s mechanisation: higher yields but greater standardisation. Advertising could become more efficient yet more uniform—making standout creative both rarer and more valuable. 

At Craft, we see AI not as a replacement but as a multiplier. As planning, production, and measurement become easier, good strategy becomes even more important. 

Rik Moore
Managing Partner, Strategy
The Kite Factory 

I’m at the demystifying/normalising stage. What I am trying to focus on is good prompt design, so I am refining how I ask LLMs to do tasks by building my knowledge around how to write really strong prompts that unlock better actions. 

Five things that unlocked AI for me: 

1. Finding writers I trust who give good insight and help you navigate and frame. For me, Zoe Scaman of Bodacious is phenomenal in this space, really insightful and generous with her output, and I find Bruce Daisley very insightful on the impact of AI on work. 
2. Thinking of an LLM as you would a human colleague, so that shapes your interactions to be more conversational rather than enhanced web searches. 
3. Moving beyond that, thinking of your LLM as an intern, which frames the level of detail and explaining you need to go into to guide them to do their best work. 
4. To support the above, using the Google prompt design 5 step framework of Task; Context; Reference; Evaluate; Iterate to get the best outputs from your prompts. 
5. Finally, seeing AI as an opportunity as opposed to a threat. Bruce Daisley highlighted a really powerful NYT article in one of his columns, that looked at the jobs AI will create (rather than the current narrative focus on jobs that will be taken away). This suggests three buckets of job-type: Trust; Integration; and Taste. Framing it like that shows where the strategist and the AI can collaborate to add real value and difference. To the old adage “strategy is about what you don't do”, in a world of endless AI possiblities, the strategist as gatekeeper of "taste", what to consider and what to leave out, is where we can add real value, focus and craft, giving the human a very clear and complementary role working with AI. 

David Morgan
Head of Creative and Production
What's Possible Group 

At The Specialist Works and What’s Possible Creative, we’re already seeing how these tools reshape the pace and ambition of the ideas we can bring to clients. But the truth is, we’re in what I’ve heard people call the “word processor era” of AI. When computers first arrived in offices, the big shift was simply moving from typewriters to word processors. It created immediate efficiencies, but it wasn’t revolutionary in the way the internet, email, or digital ecosystems later became. Those were the leaps that fundamentally reshaped how we work and think. 

AI is following a similar trajectory only the pace will be far faster, assuming the bubble doesn’t pop prematurely. The truly transformative phase hasn’t arrived yet. We’re still building the equivalent of the early tools. We haven’t invented the “internet moment” for AI. 

Right now, AI is making things quicker, smoother, and more accessible. These tools aren’t replacing creativity; they’re amplifying it, accelerating it, and giving teams capabilities that were previously impossible or unaffordable. But the real creative reinvention, the moment where the medium itself changes, is still ahead of us. 

Here’s what’s genuinely making a difference in the day-to-day creative process: 

Large Language Models: LLMs like ChatGPT have become idea accelerators. They help with early-stage strategy development, script explorations, tone-of-voice modelling, research and technical breakdowns. 

But the real power of LLMs isn’t that they behave like an assistant, it’s that they function as an extension of your own thinking. They give you instant range: the ability to explore multiple strategic directions at once, pressure-test ideas in real time, and move between creative, analytical and technical modes without friction. They don’t replace creative judgement, but they radically expand the surface area of what you can imagine, interrogate and refine. For us at What’s Possible Creative, LLMs have become a thinking partner that speeds up exploration without ever replacing the human judgement that defines our work. 

Visual Generative AI: Cloud tools like Midjourney, Runway and others or local models like Flux, Loras, Wan 2.2 are transforming how we visualise ideas. What used to require days of reference gathering, illustration, or storyboarding can now be prototyped in a single afternoon. 

It means more routes explored, better quality conversations with clients, faster alignment on tone and mood and more ambitious creative territories. 

Audio AI: Tools such as Suno or emerging music-generation systems are a massive leveller. Temp tracks, bespoke cues, quick mood pieces, all can be generated instantly for pitching or explorations. 

Upscaling and Enhancement: Upscalers, frame interpolation and image-to-video conversions are bridging the gap between “concept” and “something that feels real enough to judge creatively.” 

In general in the next 12–18 months I believe two big developments will stand out. 

AI-native pre-production or even pre-pitch! This will be a seismic shift. We’re heading towards animatics that look like finished films, storyboards that move, treatments with near-final imagery and dynamic scripts that update based on creative choices. 

Clients will see the idea far earlier. Ambitious concepts will be easier to sell. And agencies will be able to prototype four routes at once instead of one. 

The convergence of text, image, video and sound will also evolve. We’re on the brink of a single system where a script, a visual style, an edit, a soundtrack…all flow from the same prompt ecosystem.  It will collapse barriers between disciplines, making the creative process more fluid and more iterative, and yes, probably with smaller teams.