Well, that was 2025.
Christmas tick, New Year tick, now the roller coaster starts up again. With that in mind, welcome back to ‘A Nerd’s Life’ my monthly exploration of the technologies, tools and ideas that are shaping our little corner of the blue green marble floating in the inky void*.
*Thank you Mr Sagan
So, what should every marketer be chatting about with their team/agency/personal AI chatbot? What should you be exploring and why does it actually matter?
Let's kick off 2026 with 3 big trends and ideas that will be impacting you, your business, and your bottom line.
1. "Hey AI, How Does Google Actually See My Brand?"
Here's a conversation starter that'll make you uncomfortable, ask your AI to summarise your brand based solely on publicly available information. Perhaps ask AI if you should buy from your brand, what is your reputation.
Not what you want it to say. What it actually says.
Because, your customers aren't visiting your beautifully crafted About page anymore. They're asking ChatGPT, Perplexity, or Google's AI Overviews: "What's [insert your brand] good for?"
And whatever answer pops up? That's your brand now.
The Reality Check
Your brand used to be what you painted on the wall. Now it's what AI models scrape, parse, and regurgitate. You've gone from being a presenter to being interpreted. From broadcasting to being translated.
Think about it:
1. Your customers don’t see your carefully considered colour palette, until you are an option selected by AI
2. Your tone-of-voice guidelines? Irrelevant to a language model, they matter once the user has been presented your brand.
3. That award-winning campaign? Great for attention, but to an AI compressed into a single sentence
Your brand has become an interface, not an identity. And interfaces need to be machine-readable, not just human-beautiful.
What to Actually Do
Start this conversation with your AI right now:
- Analyse my brand's online presence and tell me: if you had to explain what we do and why it matters in 3 sentences to someone who's never heard of us, what would you say?
- Then compare that to your brand positioning document. The gap between those two things? That's your 2026 homework.
Practical actions:
1. Audit your structured data - Schema markup isn't sexy, but it's how machines understand you. If your website doesn't speak machine, you don't exist to AI.
2. Write for summarisation and clarity - Stop writing clever, meandering copy. AI models extract clear, simple statements. If your key messages can't survive being compressed, rewrite them.
3. Test the AI customer journey - Literally ask ChatGPT or Claude: "I need [your product category]. What are my options?" See if you show up. See what it says. Then fix what's broken.
4. Create an AI brand brief - Not for humans. For machines. What do you want AI to say about you? AI is already forming an opinion of your brand. Decide what you want AI systems to say about you, identify the inputs and user contexts that influence that view, and document it. Write it down. Then structure your content to make that happen.
Bottom line impact: If AI misrepresents your brand, you lose consideration before humans even know you exist. That's not a marketing problem, that's a revenue problem.
2. "AI, Help Me Build My Marketing Risk Register"
I know, I know. "Risk register" sounds about as fun as a compliance training video.
But stick with me, because this is where the grown-ups are separating from the kids.
The Shift Nobody's Talking About
Marketing used to report to the Chief Revenue Officer. In 2026, you're increasingly answering to the Chief Risk Officer, whether there's actually someone with that title or not.
Because modern marketing is drowning in risk:
- AI hallucination risk - Your chatbot just told a customer something completely wrong. Who's liable?
- Data privacy risk - You're using AI to personalise experiences. Whose data are you feeding it?
- Brand safety risk - Your AI-generated content just said something offensive. Screenshot, Twitter, crisis.
- Regulatory risk - The EU AI Act, state privacy laws, advertising standards - all changing faster than your legal team can read.
One viral screenshot can do more damage than a failed campaign ever could.
What to Actually Do
Have this conversation with your AI:
"Help me create a marketing risk assessment framework, for [insert your industry]. What are the top 10 risks in modern AI-assisted marketing, and what guardrails should I consider?"
Then take that framework to your actual team meeting.
Practical actions:
1. Create AI content guardrails before you create AI content - Don't wait until something goes wrong. Write down: What can AI create autonomously? What needs human review? What's off-limits entirely?
2. Treat prompts like policy documents - That casual prompt you wrote for your content AI? It's now a brand instruction. Document it. Version control it. Audit it.
3. Build a "circuit breaker" review process - Any AI-generated content that touches sensitive topics (health, finance, politics, legal) gets human review. No exceptions.
4. Run monthly AI output audits - Random sample your AI-generated content. Look for bias, errors, off-brand messaging. Treat it like you'd treat financial reconciliation.
5. Get friendly with Legal and Compliance - Seriously. Buy them coffee. Learn their language. Because they're about to become your closest allies (or your biggest blockers).
Bottom line impact: One AI screwup can cost you more than your entire marketing budget. Risk management isn't about killing creativity - it's about protecting it.
3. "Am I Talking to Real Customers or Just Expensive Mirrors?"
This one's sneaky. And it's already happening.
You're using AI to generate customer personas. To test messaging. To simulate focus groups. To stress-test your positioning.
It feels efficient. It feels smart. You're "data-driven" and "AI-native."
You're also potentially building a house of cards.
The Mirror Problem
Here's what's actually happening:
1. You train an AI on your existing customer data
2. You use that AI to generate synthetic customer feedback
3. You use that feedback to refine your strategy
4. You execute that strategy with AI-generated content
5. You measure success with AI-analysed data
See the problem? You've created a closed loop. You're validating your thinking with a machine trained on, your thinking.
That's not insight. That's algorithmic confirmation bias with extra steps.
What to Actually Do
Ask your AI this hard question:
"If I only use AI-generated personas and never talk to real customers, what risks am I taking? What blind spots am I creating?"
The answer should scare you a little. If it doesn't, ask harder questions.
Practical actions:
1. Use AI as a stress-test, not a replacement - Synthetic customers are brilliant for finding holes in your logic. Terrible for finding new opportunities.
2. Implement the "reality check" rule - For every 3 AI-driven insights, have 1 real customer conversation. Seriously. Calendar it. Make it sacred.
3. Diversify your data sources - If your AI is only trained on your existing customers, it can't tell you about the customers you're missing. Bring in external data. Market research. Competitor analysis. Cultural trends.
4. Test for "model collapse" - Compare this year's AI-generated insights to last year's. Getting more similar? That's model collapse. You're optimising toward mediocrity.
5. Keep a "surprise log" - Every time a real customer says something that contradicts your AI's predictions, write it down. Those surprises are where differentiation lives.
Bottom line impact: Brands that only listen to AI will start sounding identical. The companies that win in 2026 will be the ones that know when to ignore the algorithm.
The Real Conclusion: Marketing Is Now Infrastructure
These three trends aren't separate problems. They're all pointing to the same uncomfortable truth:
Marketing is no longer about making noise, shouting the loudest, spending the most. It's about building systems.
You're not just creating campaigns anymore. You're creating:
1. Interfaces (how machines understand you)
2. Guardrails (how you control what happens when you're not watching)
3. Feedback loops (how you stay connected to reality)
The brands that win in 2026 won't be the loudest. They'll be the ones that work even when no one's watching. The ones creating valuable content. The ones that understand their audiences. The ones that AI represents accurately. The ones that move fast without breaking things. The ones that stay human while operating at machine speed.
Because the marketers who thrive in 2026 won't be the ones who resist AI. They'll be the ones who understand its limits and build accordingly.
The future belongs to the brands that are both machine-fluent and human-grounded.
Welcome to 2026. The roller coaster is moving.
You ready?