Spark Intelligence #29: How to use AI without killing your creativity + AI's impact on hiring

Greetings earthlings,

Emma co-founder of Spark AI here with your fortnightly dose of creative AI insight for agency and brand leaders. This week we are looking at how to use AI without killing your creativity (spoiler: it involves embracing a bit of chaos), more news on AI and employment, and some great news for MS office users that will genuinely change how work gets done, rather than just making the same old tasks slightly faster.

Table of Contents

• 1. AI's complex impact on hiring

• 2. Using AI without killing your creativity

• 3. Claude's productivity breakthrough

• 4. Recent legal developments worth watching

• 5. AI for Client Services

• 6. What we are up to this week

1. AI's complex impact on hiring

Multiple studies are starting to paint a clearer picture of what’s happening with employment across industries exposed to AI. So far studies cover industries like software engineering, customer service and accounting. Findings are confirming what people are starting to feel - that where firms are adopting AI, junior employment is reducing due to slower hiring. For instance, globally there has been a 13% relative decline in employment for workers aged 22-25 in AI-exposed sectors since 2022 (PWC) and a whopping 50% decline in new graduate hires across major tech firms in the US (2019-2024) (SignalFire).

However the same data reveals significant global job creation and value enhancement, for instance workers with AI skills earn a 56% wage premium compared to those without (PWC).

What’s missing is research on what’s happening in our own industry - so The Industry Club and our team at Spark AI have teamed up to fill that gap and find out what’s happening out there in the agency and brand hiring landscape.

Our survey has just launched and if you are an agency or brand leader reading this we’d love your contribution to this important piece of research. In return you will get:

• Early access to the results (benchmark against peers)

• Invite to a November leadership breakfast to discuss findings

• Entry into a draw for a 90-minute 1:1 AI coaching session for your brand or agency, with our very own Jules Love (my Spark AI co-founder, and lecturer on AI in Business at Oxford University)

Agency or brand leader?

If you are worried about junior roles in your team right now - here’s some practical actions you can take:

1. Redefine what junior roles look like - focus on AI fluency and supporting the creation of AI enabled workflows rather than production grunt work

2. Invest in hybrid skills - the winning combination is creative thinking plus AI fluency, not one or the other

3. Double down on distinctly human capabilities - relationship building, curiosity, complex problem-solving, cultural intelligence, and creative judgement become more valuable, not less

Bit of historical perspective here: 60% of current US jobs didn't exist in 1940. Technology creates more jobs than it destroys, but the transition requires intentional skill development. We've been here before, just not quite this quickly.

2. Using AI without killing your creativity

Waaayy back in August, when it still felt like summer, I filmed an AI Summer School episode with our talented AI coach, creative director and creative AI PhD Matthew Maxwell - where we tackled the question we are all thinking: how do you use AI's power without losing your edge?

Matthew's got the perfect background for this - fine art training, creative direction experience, and he's been deep in our AI Accelerator programme where he leads the creative workstream. This is what he said:

Embrace the chaos (seriously)

Rather than trying to control AI's unpredictability, Matthew advocates using it as a creative catalyst.

"We spend so much time talking about controlling AI," he explained, "but sometimes what we need is for it to take us somewhere we wouldn't have gone ourselves."

This sounds terrifying if you're used to being in complete control of your creative process. But stick with me here.

Three ways creativity works with AI

Drawing from philosopher Margaret Bowden's framework (yes, I know, philosophy in a newsletter about AI - but it's actually useful):

Exploratory creativity: Working within familiar guidelines but finding infinite variation - think 12-bar blues with endless possibilities

Combinational creativity: Bringing together diverse influences to create something surprising - the "what if we mashed this with that" approach

Transformational creativity: Removing rules altogether to create entirely new approaches - the "sod it, let's try something completely different" method

The generator-discriminator relationship

Matthew positioned this like generative adversarial networks: AI generates ideas tirelessly, whilst humans provide critical discrimination - filtering, refining, directing toward strategic goals.

"Everyone's a creative director now," he noted. "You need both vision to brief effectively and judgement to know what's worth pursuing."

Which is both exciting and a bit yikes, depending on how comfortable you are with that level of responsibility.

Real examples from agencies actually doing this

The session featured examples from agencies using this approach in practice in our AI Accelerator programme:

• A food brand that used AI to explore noir aesthetics for plant-based products instead of the usual "clean eating" approach (brilliant, and completely unexpected)

• 8-bit animation concepts that emerged from AI suggesting unusual stylistic combinations

• Complete creative pivots where AI proposed entirely different product positioning that the team would never have considered

Catch up the 30-minute episode

3. Claude's productivity breakthrough
(and Microsoft's response)

Two major AI developments dropped this week between Anthropic and Microsoft that signal a genuine shift in how we'll work with these tools.

First, Anthropic (the company behind the Claude large language model) launched something that might genuinely change daily workflows: Claude can now create and edit Excel spreadsheets, documents, PowerPoint presentations, and PDFs directly inside Claude.ai.

And there’s a bigger story too: Microsoft is integrating Anthropic's technology into Office 365, marking their biggest step away from exclusive reliance on OpenAI.

Why this matters beyond convenience

This represents a proper shift from copy-paste workflows with Claude to integrated processes that handle entire tasks. It's the difference between having a research assistant who can only talk to you versus one who can actually pick up a pen and write things down.

What Claude can now do:

• Turn data into insights: Feed it raw data, get back cleaned spreadsheets with analysis and charts

• Build complex spreadsheets: Financial models, project trackers, budget templates with proper formulas that actually work

• Cross-format wizardry: PDF reports become PowerPoint slides, meeting notes become formatted documents

Check out Ethan Mollick’s highly amusing test.

The Microsoft angle (and why it matters)

Here's where it gets interesting: Microsoft will now use Anthropic models for certain Office 365 Copilot functions, according to The Information.

Apparently Microsoft leaders involved in the effort reported that Anthropic's models outperformed OpenAI's at specific tasks - particularly financial functions in Excel and generating PowerPoint presentations. And often more aesthetically pleasing (which has been a long term bug bear!)

This represents a significant shift in AI partnerships. While OpenAI will continue powering some Office features, Anthropic will handle the more advanced tasks. Microsoft will pay Amazon Web Services to access Anthropic's models - a notable arrangement given Microsoft's deep financial ties to OpenAI.

For context: Microsoft's Office 365 Copilot serves part of their 100+ million Copilot user base and is estimated to be on pace for over £1 billion annually.

What we'll be testing this week

Here are the workflows we're going to test with Claude’s new functionality - let us know your verdict if you try the same:

• Client briefs → project tracker: No more starting from scratch

• Raw survey data → analysis: Charts, insights and written summary in one go

• Research notes → formatted reports: Messy notes to client-ready documents

The broader picture

Both developments point to the same trend: AI moving from conversation to action. Whether it's Claude creating your spreadsheets directly or Microsoft choosing the best-performing model for each task, we're seeing AI become genuinely integrated into work rather than bolted on top. So all the more important to know how to use it to get great results.

Currently available for Claude Pro, Team, and Enterprise users, rolling out over the coming weeks.

4. Recent legal developments worth watching

Anthropic settles with authors

Anthropic recently paid $1.5bn to settle a lawsuit brought by authors claiming Claude was trained on pirated copies of their books without permission. That’s right, $1.5 BILLION. This is the first instance we’re aware of where an AI lab has actually paid rights holders for copyright infringement. This is straight out of the Silicon Valley playbook - move fast, break things, then pay the lawyers to clean up the mess.

Since this was a settlement rather than a court ruling, it doesn't establish legal precedent about AI training on copyrighted material. However, it does signal that AI companies are willing to pay significant sums to avoid lengthy copyright litigation.

Warner Bros. takes on Midjourney

Warner Bros. is joining Disney and Universal Stusios in suing Midjourney to prevent AI-generated knockoffs of Batman and Scooby-Doo, marking another front in IP protection battles.

China requires AI content labelling

Chinese social media platforms are implementing new requirements to clearly label AI-generated content - a regulatory approach that could influence global content transparency standards.

What this means for creative teams:

The legal landscape around AI and copyright remains unsettled. These cases highlight the importance of understanding your AI providers' data sourcing practices and considering content attribution in your own AI workflows, even if not legally required.

5. AI for Client Services

Last month we launched our AI for Client Services course led by Mette Davis - a one day immersive workshop followed by 3 months of coaching to embed AI workflows. We are getting amazing feedback. This is what you get:

• A more strategic role in shaping client work and relationships

• Time saved on admin, writing, and research

• Confidence using AI to enhance human judgement

• Tools to show up as trusted partners, not just delivery teams

• Practical assets to apply immediately on live accounts

• Bring consistency and capability to your whole CS team

Here’s what Sarah thought:

“I found the Spark AI Client Services workshop invaluable. There are so many things I learned that I can instantly implement within my team to make us more efficient. I am definitely an advocate for our entire client team doing this training. The skills Mette & Jules showed us in the workshop are a game changer for any client services professional." - Sarah Earley, Client Director, big fish

Watch a 30 minute taster or download the brochure:

While I’m talking about people skills, do you know Trenton Moss from Team Sterka? We are big fans. He runs a client leadership programme (not about AI - about human skills!). Trenton's running a taster on 2nd/3rd October with a senior agency crowd. Read more and sign up at sterka.team/taster or email [email protected].

6. What we are up to this week

Catch me at eight&four’s Foresight event for marketing leaders navigating change, where I’ll be presenting our report on AI’s impact on the environment, unpacking the sustainability questions surrounding LLMs.

We wrote this report because one of the number one questions that came up in our training sessions was around the impact of our increased use on AI. The results surprised us - and they will probably surprise you too.

Not read the report yet? Download it on our website:

If you're looking to develop a structured approach to AI adoption, our AI Accelerator programme helps 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 this week. I love to have your feedback - please hit a button below and let me know how you found todays issue.

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?