Greetings earthlings,

Emma here, co-founder of Spark AI. To kick off this distinctly back-to-school feeling week I’m unpacking the new evidence on AI's impact on creative jobs, sharing what you need to know about Google's breakthrough image model that got everyone talking in August, and showing you some brilliantly simple but impactful AI tips in client services.

Actually, before all that - a quick interlude.

I'd love your take on the format of this newsletter. Are these fortnightly long reads working for your schedule (and attention span!), or would more frequent bite-sized updates suit you better?

Thanks for that! Now back to it.

Here’s what I have in store for you today:

• 1. A new narrative about creative job threats 🙌

• 2. Nano Banana (we couldn't not)

• 3. AI for faster and smarter client onboarding

• 4. Claude moves to opt-out

• 5. Essential reading

1. A new narrative about creative job threats 🙌

New research from the World Economic Forum gives us the evidence we have been waiting for: that AI isn't ‘replacing’ creative jobs in the way that many LinkedIn doomsayers expected. In fact the WEF report has shown that AI is transforming jobs in ways that favour human creativity and strategic thinking. Here are the findings you need to know for you and your team:

A. Creative roles are growing, not shrinking

Jobs requiring creative problem-solving, emotional intelligence, and strategic thinking are seeing increased demand. The research shows these skills become more valuable when paired with AI capabilities. Now, I’ve been harping on about this for a while - the trend of creatives building stuff to be more creative (see my interview with Craig Ward as just one example, or read about the team behind Springboards, or eight&four building Platform12, to new emergent tools like FirstConcepts). In fact Y Combinator are encouraging more designers to become founders. Exciting stuff for agencies and freelancers alike.

• Our takeaway 👉 Curious creatives are uniquely positioned to capitalise on AI - but in team or businesses this takes setting the right vision, purposefully encouraging an innovation culture and intentionally making time and thinking space to identify your unique opportunities (we built AI Accelerator to show agencies how to do this)

B. An AI cliche, but it really is about augmentation, not replacement 

The most in-demand people right now are those who can combine AI efficiency with human insight. Think strategists who use AI for rapid research synthesis, or designers who blend AI-generated concepts with their own creative direction.

• Our takeaway 👉 Investing in AI literacy for your people is an important component of stopping flight risk, like investing in any other upskilling.

C. New hybrid roles are emerging

The data reveals entirely new job categories: AI creative directors, prompt engineers for brands, and creative technologists.

• Our takeaway 👉 Hold this thought as we are co-developing a deep dive research project on this very topic - more very soon 👀.

D. The skills gap is real

While demand for AI-augmented creative roles is growing, there's a significant shortage of people who can bridge technology and creativity effectively.

• Our takeaway: 👉 This is the time to upskill your team and gain the competitive advantage for your business . This is another reason we launched our AI Accelerator at the start of the year - check out the difference it made to two of our clients. We have a backlog of client case studies to write up - its the eternal problem of finding a chance to write them!

Interestingly, this is in stark contrast to a previous WEF report published in January - which has been cited in nearly every AI talk I’ve been to - The Future of Jobs Report - where graphic design was just outside the top ten of the largest declining jobs 🤷🏻. I think it just goes to show not just how fast the tech is evolving but how quickly our curious creative minds are mining the opportunities AI presents.

Final thought on this:

Jobs might be evolving but this doesn’t mean you need to become a technologist. It's about having curiosity and applying creative direction to the incredibly powerful tools we now have at our disposal. The opportunity is enormous. But to exploit it requires intentional skill development, an innovation focussed culture and good old fashioned change management.

Read the full WEF report.

2. Will Google's Nano Banana spell the end for Flux LoRAs?

Google has released what it's calling Gemini 2.5 Flash Image (codenamed "Nano Banana"), that it would have been hard not to notice over the last couple of weeks unless you went LinkedIn cold turkey over the holidays.

What makes Nano Banana different?

The key breakthrough is consistency and control. Unlike previous models where maintaining character consistency or visual style across different angles was hit-or-miss, Nano Banana tackles this head-on. You can combine different images into new ones, edit with simple language prompts, and maintain visual coherence across a series.

Creative strategist Ali Cheikali at Google calls it "the Veo 3 moment for creative control", which was endorsed by co-founder of AI-first creative agency Aigency Amsterdam, Chrissie Cremers who describes it as finally solving "consistency, accuracy and speed" in AI image creation.

Why this matters for you

Until now, we’ve relied on Midjourney’s character reference or more recently their omni-reference feature to get consistent characters and objects across scenes. And for teams taking it to the next level it meant training Flux LoRAs (Low-Rank Adaptations) in ComfyUI to achieve consistent character looks and brand elements across campaigns. Definitely a technical and time consuming challenge for most teams.

Nano Banana eliminates much of this complexity. Instead of training custom LoRAs for characters or objects, you can achieve similar results through intelligent prompting and simple image referencing - “have the person in image 1 hold the beer can in image 2”.

Real-world application

A workflow with Nano Banana might combine Midjourney for base imagery with Nano Banana for product integration - the power is still in combining different tools to work together rather than one replacing the other entirely.

Nano Banana has also just been integrated into Weavy, which integrates AI models and editing tools together into repeatable workflows for your team. We are now using Weavy in our AI for Creatives workstream of our AI Accelerator programme (we have no affiliation - just loving this tool right now!).

The broader shift

Following the same train of all previous leaps forward in technology, the future is heading toward more intuitive, conversation-driven creation rather than requiring complex technical workflows. This makes advanced AI image creation more accessible to everyone, including marketing teams at clients. Remember, AI creates risks for your agency, as well as opportunities.

Google is rolling out Nano Banana for free across Gemini app, AI Studio, and Vertex AI - making enterprise-grade creative control accessible to small agencies and individual creators.

Try it yourself: 👉 The model is available now in Google AI Studio. The best way to understand its potential is to test it against your current workflows.

3. AI for faster and smarter client onboarding

A couple of weeks ago Mette Davis, our client services expert, delivered Episode 2 of AI Summer School with a brilliant demonstration of how AI can transform your client onboarding process. Mette showed us exactly how to go from chaotic project starts to structured, strategic client relationships - all in under 30 minutes.

My big takeaways for using AI in client onboarding:

From brief-takers to strategic partners. Mette's central message was more fundamental than AI workflows - it was client services moving from reactive brief-takers to true strategic partners. That's where real value lies - and where AI can help you get there faster.

Set up client projects like a pro. Using setting up a Gemini Gem as an example, Mette walked through creating a client 'Gem' (or CustomGPT/Claude Project/Co-Pilot Agent in other tools) that becomes your go-to knowledge base. Upload company overviews, stakeholder profiles, brand guidelines, and annual reports. The result? An AI assistant that actually understands your client's world.

Better prompts = better outcomes. Mette demonstrated the difference between lazy prompting ("give me an overview of the client context") and strategic briefing of AI. Spark’s "cheeky prompt" structure aka ‘CICI’: - Character, Instructions, Criteria, and Inputs - transforms vague requests into comprehensive, actionable client summaries.

Real-world time savings. Mette's research revealed that most people spend half a day to over a full day on client setup and onboarding documents for team members. With AI, that same comprehensive briefing document was created in minutes - complete with sector insights, risk assessment, and opportunity identification.

The human stays in control. Throughout the demo, Mette emphasised Spark’s "Think → AI → Think" approach. You set the strategic direction, use AI to accelerate the process, then apply your judgement to refine and improve the output.

👉 Catch up on Episode 2 of AI Summer School 

Next up: Summer School Episode 4 on spotting the AI opportunities and risks in your business

You might have spotted a sequencing problem here - where is episode 3? Well it aired last week (on using AI without killing your creativity with Matthew Maxwell) but I’m going to save it for a deeper review in the next edition of Spark Intelligence as there’s a recommended newsletter length and I always exceed it!

Meanwhile, this Wednesday at 11am, Emma Jackson (EJ), who we are proud to have on our incredible AI leadership coaching team, takes the spotlight to talk to me about identifying the opportunities and risks of AI in your agency. As well as an AI leadership coach EJ is an award winning prompt engineer and exec producer - she will help you think strategically about where AI can drive genuine value versus where it might create unexpected challenges. It’s a glimpse behind the scenes of how EJ guides our clients in the AI Leadership workstream of our AI Accelerator Programme.

✅ Perfect for agency leaders, department heads, and anyone responsible for shaping their organisation's AI strategy.

Haven't registered yet? You'll need to sign up to watch this free episode.

👉 Register for Episode 4 on AI Leadership

4. Claude moves to opt-out

At the end of September Anthropic is changing it’s T&Cs so that instead of opting-in to your chats with Claude training the model, you now need to opt out. You should get a pop us asking you to approve this change to privacy settings, just choose toggle off if you are using Claude for sensitive data (think for client work, HR data, company strategy - really anything work related).

If you don’t get the pop up then manually opt out in your Claude account’s privacy settings.

5. What Jules has been working on behind the scenes 📚

We are gearing up for Jules' book Shift: AI for Agencies to launches in 2 months. It's the practical playbook creative leaders have been asking for - tried and tested strategies for agencies ready to move beyond experimenting.

Jules has spent years helping creative leaders navigate technology and AI adoption, and this book distills everything: deploying AI across teams, building AI-forward culture, and talking to clients about it with confidence.

👉 Pre-register for Chapter 1 and get the opening chapter as soon as it's ready.

That's all for this week. We love to have your feedback - please hit a button below and let us know how you found this issue.

See you next time!

Emma
Co-founder of Spark AI

P.S. If you want to find out more about how Spark can partner with your team to accelerate AI adoption book a quick chat:

About Spark AI

We help creative and brand leaders turn AI curiosity into confidence - through structured AI guidance and upskilling for agencies and brand teams.Trusted by Oxford University Saïd Business School and backed by Innovate UK.

👉 Find out more 

Not a subscriber yet?