Spark Intelligence #30: Where does your agency stand in the great AI divide?

Greetings earthlings,

Emma co-founder of Spark AI here with your fortnightly dose of creative AI insight for agency and brand leaders.

We are celebrating at Spark HQ as today is the day we publish The Spark Report, our latest six-monthly benchmark of creative agencies and their AI adoption - and the findings reveal our industry at a critical turning point.

While 80% of agencies now using AI, we’re starting to see a stark divide. While most remain trapped in the experimental stage, a small group are pulling decisively ahead and beginning to fundamentally reshape how they deliver value and generate income. Scroll down to find out how.

Plus, we're celebrating some rather lovely recognition from the National AI Awards and are building up to some exciting events in October you won’t want to miss.

Table of Contents

• 1. The Spark Report: The traits of agencies leading with AI

• 2. Spark is a National AI Awards finalist! 🏆

• 3. Gemini Gems are finally sharable with your team

• 4. Build practical AI skills in October at AI Changemakers

• 5. What's happened to our podcast? 👀

1. The Spark Report A/W 2025 drops: benchmark your agency against the AI leaders

The numbers tell the story:

• 5% (and growing) are really pulling ahead - having built proprietary workflows, new IP around AI and new AI-powered services.

• The rest, even though they use AI more frequently (regular users jumped from 20% to 35% in six months), remain stuck in "the experimentation trap" - enthusiastic individuals still ‘experimenting’ but without a clear strategy.

The difference between success and stagnation?

Agencies succeeding with AI have moved from scattered experiments to systematic skill and capability development. While most continue playing with isolated tools, the leaders are building structured approaches that transform their entire service model.

What separates the innovators from the rest?

The minority who are pulling ahead aren't focussed on using AI for efficiency alone, they are being far more radical. They are systematically rebuilding their service models around AI capabilities. These agencies are building proprietary workflows, creating new IP, and delivering previously impossible client experiences. We found that these innovators share three critical approaches:

1. Leadership treats AI as business transformation, not just efficiency. They're not asking "How do we do the same work faster?" but "How do we deliver value that wasn't possible before?"

2. They invest in structured training tailored to their business. Agencies with proper AI training are 3x more likely to use AI strategically. The differentiator isn't which tools you choose - it's giving your people the skills, time and confidence to use them well.

3. They build prototypes that quickly move into testing and rollout. And rather than avoiding AI conversations with clients, they lead them with clear policies around data, IP, and creative ownership.

Innovation creates differentiation

The agencies reaching maturity are building capabilities competitors can't replicate:

Brave Bison's AudienceGPT creates synthetic audiences with 90th percentile accuracy, replacing expensive focus groups with on-demand insights. Eight&Four's Platform12 includes custom AI tools from image generation to social listening. R/GA's Smart Story Engines generate personalised narratives at scale, while OLIVER's Seance achieves 85% correlation with human focus groups at 10% of the cost.

As one leader told us: "AI hasn't made our work cheaper, but is has made it better. Now we are exploring more territories, testing more ideas, and then refining them more."

How have they achieved this? What can we learn?

The agencies reaching AI maturity share a systematic approach to innovation. Canva's model is the ultimate example of this: they gave 5,000 employees a full week to experiment with AI tools - no meetings, no output pressure, just structured exploration with role-based sessions and cross-team prototype building. The results were immediate: sales teams built custom assistants, designers halved production time, product teams generated synthetic data for testing.

Leading agencies often pair subject matter experts with technologists - having creatives and strategists lead tech development rather than leaving it to IT departments without subject matter input. You wouldn’t develop a brand strategy without engaging your client. They also provide dedicated experimentation time with clear frameworks for testing, quick go/no-go decisions, and structured pathways from prototype to product. It’s intentional experimentation, and not just left to organically evolve out of unstructured play.

As the report details further, agencies succeeding with AI have moved from "scattered experiments to systematic skill development," embedding AI specialists into creative squads and redesigning how briefs are shaped and iterated. Our AI Accelerator is designed to support you through exactly this process - and clients tell us the programme pays for itself before it even finishes.

Where this leads

The research reveals three competitive categories emerging: volume agencies using AI to scale content production, innovation leaders creating previously impossible experiences, and craft specialists emphasising human creativity. The middle ground faces the greatest pressure.

The window for competitive advantage is open now, but it won't stay open forever. When we benchmark again in six months, we expect to see the Innovators category, currently 5%, grow fast - with agencies building new services, SaaS tools, evolving business models, and proprietary IP that competitors can't match.

Our AI Accelerator programme is designed to guide you and your team to do exactly this - to move from the “experimentation trap” to structured adoption and era defining possibilities for your business.

Download The Spark Report to see the full AI Maturity Model, work out where you are, and understand your next steps with practical frameworks for moving beyond experimentation to strategic advantage.

2. Spark is a National AI Awards finalist! 🏆

We're pinching ourselves that we are finalists for the AI Award for Education at the National AI Awards.

This recognition reflects our work helping creative teams turn AI curiosity into genuine capability. Over the past year, we've worked with agencies of all shapes and sizes, from independent to in-house, helping them navigate the shift from scattered experiments to systematic skills and business model development.

The nomination feels particularly meaningful because it recognises what we've always believed: the best AI transformation happens when you invest in people, not just technology.

Awards aside, what matters most is seeing our clients build confidence, reshape their services, ignite excitement across teams and lead their markets with AI. That's our real measure of success.

But I did get to go to an incredible finalists drinks patry last week at the top of 22 Bishopsgate, on the most beautiful evening…

3. Google closes its biggest gap: Gemini Gems are finally shareable for teams

Google Gemini Gems (the equivalent of Custom GPTs or Copilot Agents) are now shareable - and this changes everything for team workflows. Until now, Gems were only accessible to individual users, which was the biggest drawback of Gemini compared to ChatGPT or Copilot for collaborative work.

This limitation has been the subject of countless discussions with clients and deliberations over which LLM to subscribe to. Many teams who were already Google users have been going for other models like ChatGPT because this feature to share AI Assistants across the organisation wasn’t previously possible.

That limitation just disappeared. Last week Google made sharing Gems as simple as sharing a Google Doc, finally giving teams the collaborative AI capabilities they've been missing - and potentially reshuffling the agency AI decision-making process.

What's changed?

If you built a brilliant Gem for writing brand copy in a specific tone, reviewing creative briefs, or generating social content ideas, it previously lived in your account alone. Your teammates had to recreate similar versions from scratch. Sure, they could copy your instructions, but then when you made a tweak to improve performance they would have to do it all over again. The result - too much friction. Teams just never did it, and we saw agencies that had gone for ChatGPT or Copilot rolling out fleets of AI Assistants much more quickly.

Now, sharing a Gem works exactly like sharing any Google Workspace file. Hit the share button, set permissions for view or edit access, and send the link. Recipients see shared Gems in a dedicated "Shared with me" section, and can either use them directly or make personal copies for customisation.

Practical applications to try:

• Brand voice consistency: Share a Gem trained on specific client guidelines across account teams, or your own ToV for marketing

• Brief standardisation: Build review templates that ensure consistency and completeness, and suggest questions to ask clients to clarify what’s ambiguous

• Client-specific research: Create industry or sector-focused research assistants that everyone can consult

• Pitch deck optimisation: Share presentation review tools across new business teams, built pitch scripts and test where they’re weak

• Sales or stakeholder follow-ups: turn meeting transcripts into follow up notes in a consistent format and tone

The strategic implications

Combined with recent upgrades to Gemini like Deep Research for citation-backed reports and tighter Workspace integration, Google wants Gemini to be where teams get work done, not just where individuals experiment.

What to do now

If your agency uses Gemini, start thinking systematically about which AI assistants could benefit multiple team members. If you’ve been wavering between ChatGPT ($30 per user each month) and Gemini (free within Google Workspace), that choice just got easier.

Check out Gemini

4. Join us at AI Changemakers in October to build practical AI skills

The full day covers:

• Building your own AI fluency and confidence

• Identifying valuable AI opportunities in your company or industry

• Best practice around AI Ethics from Ofcom and how to embed AI across marketing teams from Virgin Media O2 

• Strategic frameworks for career development in an AI-shaped world

AI Changemakers was launched in Sydney at Canva's HQ and received an overwhelming positive response. Attendee's will walk away with an AI Action plan and unique insights into real-world transformation success. It’s the last few days for early bird tickets - get yours!

5. What’s happened to the ‘What’s New in Creative AI’ podcast? 👀

Maybe you haven’t given it a second thought, but I’ve been feeling guilty that our last episode of What’s New in Creative AI was way back in June!

Why’s that? Well this summer real-life popped up reminded us it is unpredictable! You may or may not know that Jules Love and I, the co-founders of Spark AI also happen to be married. One of the downsides of being married to your business partner* is when life gives you lemons it affects you both at the same time! Elderly relative caring responsibilities and health scares all popped up at once to remind us we are not superhuman. So we doubled down on client delivery and things like the podcast had to go on ice for a while. But don’t worry we’ll be back later October. I’ll give you a heads up when it’s out!

Thank you to everyone who’s been so supportive of us over the summer - especially our incredible team, Alicia, EJ, Steve, Mette, Matthew, Simone, Jonas and Anya.

*you’ll be surprised that there are many upsides too - perhaps I’ll write about this one day!

If you're looking to develop a structured approach to AI adoption, we help you and your team build confidence and capability through hands-on learning you can immediately implement your work. It’s built specifically for agencies and brand teams.

 Book a slot in my diary if you’d like to have a chat.

That's all for today. My big takeaway from this edition is that AI success isn't really about the tools you choose. It's about building the confidence, mindset and clarity to use them strategically across the business.

See you next time!

Emma

Co-founder of Spark AI

About Spark AI

We help creative and brand leaders turn AI experimentation into confidence - through structured adoption, upskilling and applied AI workflow building. Trusted by Oxford University Saïd Business School and backed by Innovate UK.

👉 Find out more 

Not a subscriber yet?