AI for Creative Ideation and Concept Development
How to use AI to expand creative thinking, develop stronger campaign concepts and explore more possibilities without losing originality.
Why this matters
AI has dramatically reduced the cost and time required to generate ideas, copy, images and variations.
The new challenge is no longer producing enough ideas. It is recognising which ideas are genuinely interesting, strategically relevant and worth developing.
Kantar’s 2026 marketing trends argue that AI can strengthen creative development and optimisation, while stressing that human insight remains essential to authentic brand storytelling.
Key point: AI can increase creative capacity. Marketers still need to provide creative judgement.
Session aims
By the end, attendees should be able to:
Use AI to explore creative territories rather than generate generic lists
Apply proven creative-thinking models with AI
Move from audience insight to campaign concept
Generate genuinely different alternatives
Evaluate ideas systematically
Protect brand distinctiveness and originality
Use AI responsibly in creative development
What role should AI play?
Think of AI as a combination of: Brainstorming partner, Research assistant, Provocateur, Alternative perspective, Rapid prototype tool, Critic, Variation engine
Avoid treating it as: Creative director, Customer, Final decision-maker, Automatic source of original insight
Key point: AI expands the creative workspace. Humans still define what deserves to survive.
Creativity needs divergence and convergence
Creative work involves two different modes:
Divergence
Create possibilities, explore unusual connections and broaden thinking.
Convergence
Evaluate possibilities, apply constraints and choose what is worth developing.
The Design Council’s Double Diamond deliberately moves between divergent and convergent thinking throughout the innovation process.
AI strength: divergence at speed.
Human strength: context, judgement and purposeful convergence.
Use the Double Diamond with AI
The Design Council model gives us four stages:
Discover – Explore the audience, problem and context.
Define – Decide what problem is genuinely worth solving.
Develop – Generate and test multiple possibilities.
Deliver – Select, refine and implement the strongest solution.
Key point: Do not start by asking AI for campaign ideas. Start by understanding the problem.
Stage 1: Discover before generating
Give AI evidence to work with:
Customer research
Reviews
Search behaviour
Sales feedback
Competitor claims
Campaign performance
Category trends
Brand positioning
Then ask:
What tensions appear repeatedly?
What frustrations are unresolved?
What language do customers actually use?
What assumptions should we challenge?
Strategic use: identify fertile creative territory.
Tactical use: find hooks, objections and language for execution.
Stage 2: define the creative problem
Weak creative challenge:
Give me ideas for promoting our new software.
Stronger challenge: How might we help overstretched marketing managers feel that adopting our software will reduce complexity rather than introduce another system to manage?
Key point: The quality of the creative problem strongly influences the quality of the ideas that follow.
Build from insight, not AI invention
A useful concept-development chain is:
Evidence → Insight → Tension → Proposition → Creative idea → Execution
Example:
Evidence: Customers avoid switching suppliers because migration feels risky.
Insight: Familiar inconvenience feels safer than uncertain improvement.
Tension: “I know this could be better, but changing feels dangerous.”
Proposition: Switching can be simpler than staying stuck.
Creative territory: The Cost of Staying Put
AI can help at every stage, but the evidence should anchor the thinking.
Stage 3: use AI for divergent thinking
Do not ask for “10 ideas”.
Ask for deliberately different territories: Emotional, Rational, Humorous, Contrarian, Customer-story led, Product-led, Cultural, Visual, Behavioural, Unexpected analogy
Practical instruction: “Generate eight fundamentally different campaign territories. Do not produce variations of the same underlying idea.”
Use SCAMPER with AI
SCAMPER provides structured creative provocation:
Substitute – What could be replaced?
Combine – What could be brought together?
Adapt – What could we borrow from another category?
Modify – What could we exaggerate or reduce?
Put to another use – Where else could this idea work?
Eliminate – What happens if we remove something?
Reverse – What if we invert the assumption?
Practical use: Run one promising proposition through all seven lenses.
Force creative distance
Generic prompts encourage predictable ideas.
Create distance by asking AI to:
Borrow a mechanism from another industry
Combine two unrelated cultural references
Reverse the category convention
Explain the idea to a child
Imagine how a luxury brand would solve it
Imagine the campaign with no words
Remove the product from the advertising
Turn the problem into a physical metaphor
Key point: AI becomes more useful when you make sameness difficult.
The originality warning
A 2024 Science Advances experiment found that access to generative AI ideas increased individual creative performance, particularly for less creative participants. But AI-assisted outputs also became more similar to one another.
Marketing implication:
AI can help an individual produce stronger work while simultaneously making the category look more homogeneous.
Reflection: Could your competitor produce essentially the same idea from the same prompt?
Use the anti-sameness test
Before developing an AI-assisted idea, ask:
Does this sound like our category?
Could a competitor run it unchanged?
Is the language suspiciously familiar?
Does it use obvious AI metaphors or clichés?
Is there a genuine human tension underneath it?
Is the brand necessary to the idea?
Key point: An idea becomes stronger when removing the brand makes it harder to imagine anyone else owning it.
Variation is not invention
AI is exceptionally good at variation.
It can quickly change: Headline, Tone, Visual style, Format, CTA, Audience emphasis
But thirty variations may still represent one underlying idea.
Practical test: Strip away the wording and visuals.
If the strategic thought underneath is unchanged, you have variations rather than new concepts.
Stage 4: converge with an idea scorecard
Score each creative territory from 1–5 against: Audience relevance, Strategic fit, Originality, Brand distinctiveness, Simplicity, Emotional potential, Proof, Channel stretch, Commercial potential, Practical feasibility
Key point: Replace “I like this one” with a reasoned creative decision.
Use AI as critic, not just creator
Once you have several ideas, change the AI’s role.
Ask it to challenge them as: The target customer, A competitor, Sales, Finance, Legal, A sceptical journalist, A brand strategist, A creative director
Then ask: Which criticism exposes a genuine weakness rather than merely offering another opinion?
Strategic use: pressure-test investment.
Tactical use: catch weak execution before production.
Build a campaign idea, not just an advert
A strong concept should stretch across channels without losing its core.
Kantar reports that coherent cross-channel ideas have become substantially more important to campaign success, and warns against isolated executions that fail to connect clearly to the brand.
Test, can:
Paid media use it?
PR create a story from it?
Email develop it?
Social participate?
Sales use it?
It work in six seconds and six minutes?
Keep the idea, change the execution
Protect: Customer tension, Proposition, Promise, Proof, Creative platform
Adapt: Opening, Length, Visual treatment, Format, Channel behaviour, CTA
Example: The same campaign idea may become a six-second visual joke on social, an evidence-led article on the website and a customer case study in sales.
AI for visual concept development
Use visual AI early to explore: Mood, Composition, Metaphor, Settings, Character, Photography direction, Colour territory, Storyboards, Packaging or display concepts
Practical approach: Generate rough visual territories before polishing production assets.
Key point: Prototype cheaply before investing heavily.
Use creative iteration deliberately
A useful loop: Generate → Compare → Challenge → Combine → Refine → Test
Avoid: Generate → Like → Publish
The Design Council explicitly describes innovation as iterative, with ideas tested, rejected and improved rather than treated as finished first attempts.
AI should support brand distinctiveness
Kantar’s 2026 guidance argues that AI-enabled innovation should remain brand-led and rooted in consumer motivations and meaningful difference, rather than being technology-led.
Before approving an idea, ask:
Does this express something distinctive about us?
Does it strengthen recognisable brand memory?
Is it culturally appropriate?
Does it feel human?
Does it reinforce our positioning?
Responsible creative development
AI-generated creative can raise questions around: Copyright and provenance, Bias, Representation, Likeness, Product accuracy, Synthetic people, Misleading imagery, Disclosure
WFA has developed responsible AI principles specifically to guide brands’ use of generative AI.
Key point: Creative freedom still operates inside professional responsibility.
Transparency when AI changes reality
In April 2026, ICAS and WFA published global best-practice guidance on transparency in AI-generated marketing creative, covering people and likeness, product images, audio, background visuals and marketing copy.
Hilary Souter, President of ICAS, said AI offers “significant opportunities for innovation and effectiveness” while raising important questions around transparency, trust and consumer protection.
Practical question: Would a reasonable customer feel misled if they knew how this creative was produced?
Different organisations, different workflows
Small business – AI can act as additional creative resource.
Keep the process simple: evidence → ideas → shortlist → create → test.
SME – Use AI to expand internal capability while creating shared brand prompts, approval criteria and asset libraries.
Enterprise – Use structured governance, approved tools, agency rules, provenance processes and formal creative testing.
Key point: Scale the governance with the risk and complexity.
Common mistakes to avoid
Asking AI for ideas before understanding the audience
Mistaking quantity for originality
Accepting the first attractive output
Generating executions before finding the proposition
Copying category conventions
Losing brand voice through generic prompting
Using AI to manufacture customer insight
Automating creative judgement
Ignoring disclosure, copyright or representation issues
A practical AI creative workflow
Use this sequence on a real campaign:
Define the commercial objective
Gather customer evidence
Identify the tension
Write the proposition
Define the creative challenge
Generate contrasting territories
Apply provocations such as SCAMPER
Run the anti-sameness test
Score and shortlist ideas
Pressure-test the shortlist
Prototype executions
Test, refine and approve
Practical activity
Choose one current marketing brief and complete:
The business objective is…
The audience tension is…
The proposition is…
The creative challenge is…
Three genuinely different territories are…
The most distinctive is…
The main weakness is…
AI could help refine it by…
Human judgement is essential because…
The first prototype will be…
Key takeaways
AI can give marketers extraordinary creative range and speed.
The advantage will not come from generating the most ideas. Everyone can now do that.
It will come from asking sharper questions, finding better customer tensions, forcing greater creative distance, recognising genuine originality, protecting brand distinctiveness and knowing which ideas deserve human investment.
AI can widen the possibilities.
Marketers still have to choose what is worth making.
More webinars like this at Cambridge Marketing College
