We Want Better AI Ads
Marketing Unfiltered #51 → AI Ads Have To Be Better - Here Are Some Ideas
Happy Friday leaders, I think it’s safe to say it’s Autumn.
This week I am diving into how AI Ads have to be better for us and our customers and not just optimisations of what we have currently!
Here’s A Teaser For Next Week’s MU’s First Birthday With A Mini Brand Masterclass By Harry (grilled by me)
I Have Some Unsolicited Advice For LLM And AI Companies.
AI Ads Have To Be Better!
Not optimised, BETTER!
Better look, better feel, a better experience for customer and brand side (both sides of the marketplace).
This week, I jumped with both feet into the future of advertising for the AI world, while recording my AI Moment podcast and how AI companies can and should rebuild the ad unit.
Or enjoy NotebookLM’s take on my article
Let’s Shift To Better
In recent conversations with Marketing, Growth and business leaders, I am pushing for leaders to start thinking differently as we are on the brink of the ad unit to shift.
Many have become reliant on the current but lets face it basic ad offering, relying on Google, Instagram, Facebook and co to drive potential traffic and our product to sell or upset potential customers.
Choice & Clicks To Personalised Solutions
Ads have to be re-engineered and move away from keyword-based ads (Google style) and what appears as the Meta overly creepy intent-driven ads towards what I view as a highly personalised “solution-driven ad experience”.
My core vision for AI ads is for a hybrid advertising model that blends the best of search, social, and traditional media to create something entirely new.
“You could buy a product, you could buy an outfit, you could do something all in one click, which is very, very different to how Google works, how LinkedIn works and how traditional ads have worked.”
- Me on AI moment Podcast
Fancy words Danny. Explain it a little more
Ads Are Intent-Based: AI ads’ intent is clear from the prompts you ask, it is clear from the set of prompts you have entered and likely iterated a number of times and then for most guessing the output you are looking for.
Not new but different: Google has to infer your intent from the keywords you search for, to the language you use and what they do is combine a number of searches to guess what you are looking for or the action you haven’t taken yet.
I believe the new ad types present an enormous opportunity for businesses to deliver genuine value to their customers; however, it also requires a fundamental rethink of success metrics and brand safety.
I have clearly thought a lot about ads and AI, (from predicting it back in December 2024 in my ChatGPT ads), there should be options, here are 4:
Solution Ads: As the brand, we can decide to serve a dedicated ad, an ad that is workflow driven and offers a solution - these likely would be most expensive
Option Ads: We could decide to serve a cheaper ads that is more here are options - these ads would likely be cost per click or cost per answer
Generic Ads: Or we could serve a generic TV/cinema style ad that’s designed for awareness - this could be based on views, clicks or something new.
Invite Ads: Imagine tagging in your brand onto the prompt and inviting you to answer the prompt or inquiry - this could be highly priced but highly convertible
If Google is shifting its Search from links to answers, Search is now having to become more conversational, proactive, and creative; helping users solve issues, learn different and new approaches, and make decisions, therefore now is the time to improve the ad experiences across tech, making answers and solutions displayed more creatively and in the format our users want.
Agentic Ads: we could/should also think about how to shape new ads moving towards the agentic web, where a potential customer (or bot) can go from prompt to purchase or creation in a single, seamless interaction or allow its agent (our bots) to do this on our behalf.
I’ll leave how agentic ads could work for now…but I have ideas…
Please vote below
My Predictions / Unsolicited Advice For OpenAI, Anthropic, Google & Co…
The Risks
Incorrect Answers: What if we misread the question or misunderstand the prompt, this is already often the case, we will mismatch ads to keywords, we will show up on sites we didn’t know we were, we often actually appear in feeds where someone has shown interest in something and we appear with a product they will never want or never can afford.
Rethinking Success Metrics: Something we should consider is the need for new key performance indicators. If an ad delivers a fully customised asset, its value is exponentially higher than a click on a generic link.
I have to urge you to prep your business leaders and owners to start thinking beyond traditional metrics like CPC and ROI and consider how to measure the deeper value and engagement these new ad formats will generate.
I could propose: cost per customised answer, cost per solution, cost per usage, cost per resolve (think how the AI CS Agents like Intercom Fin and Sierra AI).
Attribution is always going to be tricky for the providers but here are ways they could offer fair attributions: how about we build multi surface tracking (based on user details), we do not charge based on the models that are used/or re-prompted
Navigating Brand Safety and Privacy: While I sound AI Pilled, aka an AI optimist, I want to ensure to I ground my future in reality. There are (potential) critical risks of AI hallucinations, and we, as the brand leaders, have to control which prompts and terms they are associated with. This is no small feat.
We will have to create AI ad policies, create a set of ad guardrails and then flag escalation paths. User journeys should be transparent, as the ad buyer we should have transparency of what our customers do, this is something traditional and digital providers will hide from us.
We should prepare ourselves for how “ad memory” would function, touching on the fine line between helpful personalisation and intrusive data collection. I’m sure this would need more thought around GDPR, PECR, CCPA, DSA, DMA etc but i’m sure this is in motion.
The Benefits
A New Hybrid Ad Model: My vision for AI ads as a powerful blend of Google, TikTok, LinkedIn, and the ability to leverage TV advertising. The goal is no longer just to answer a keyword-based query or serve ads based on what you’ve looked at or accidentally clicked on, but also to provide inspiration, aspiration, and most importantly, a personalised output, all within the LLM or AI interface. AI is the format to move the conversation beyond simple text ads to personalised solutions and outcomes.
The Rise of ‘Agentic’ Advertising
Ads should become more interactive and almost solely solution-oriented.
A few examples to tease you: a user could receive a highly customised whitepaper based on the prompt you search, they should have a bespoke report built for them for their role and their business, or could download a personalised infographic all built for you.
It could go even further, and, quite easily be created for you, it could be a pre-populated Google Sheet created on your behalf. We are seeing these organically complicated prompts, in their most advanced form, a user could upload a picture, the AI could show you what you would look like in the outfit, making recommendations, and in one click purchase an entire outfit or product directly from one prompt, a one-click process that is fundamentally different from any current ad system out there today.
Platform Specialisation
I want to offer a forward-looking perspective on how different AI providers might specialise their advertising offerings.
ChatGPT could focus on consumer-based ads and highly customise outputs and answers as above
Claude should cater to the developer and tech space with ads for tools, widgets, APIs and offering discounts from ad units, whether that’s on token usage or API calls etc. The potential here is huge.
Google, through Gemini could serve ads based on professional work and productivity tasks. Google sheets, Google slides etc based on your requirements, maybe Google jumped the gun on great tools like AI studio and NotebookLM… Google could also offer this through Google search or through browser search (Apple would love a rev split); imagine the SERPs offering great ads.
These brands have to be strategic in their ad offering, free accounts, or ad-supported accounts like Netflix and Amazon Prime Video won’t cover the cost of AI (infrastructure, costs of queries/token usage and demand on servers). I can imagine affiliate revenue being shared or a new rev share package created.
This could increase innovation at product level, browser side and ad side - ads have stagnated for a while now.
The Immediate Future
The current landscape has seen some baby steps in optimising campaigns and automating campaigns for you through Meta ACS and Google PMAX - these have seen increased costs and to a point become more of a black box.
You can see from my predictions that we have so many more opportunities to provide solutions than optimisations or links to a landing page with gated content…B2B ads could be exciting, it would mean B2C ads aren’t just pushing to landing pages but to experiences.
Not Just For Brands
Affiliates could be included in this, often being seen as the experts or in the know, or could be outsourced on behalf of the brand to answer questions or prompts.
I could see influencers and content creators being employed to offer customised videos, guides and solutions with brands. Brands would love this organic extension and could be demand based, people will wait like they do on Cameo video’s…
I predict many AI companies will take baby steps towards this type of ads, but then have to race towards the ad light at the end of the tunnel, it will hit like a truck for many Marketing leaders, who aren’t thinking far enough ahead.
We should prepare ourselves for an aggressive ramp-up as AI firms race to monetise AI tools and prove ROI, otherwise these mega rounds will stop.
2026 is the year of ROI in AI.
Not just for AI companies but for department leads and their team members, no CFO will want the uncertainty and unpredictability some have been given this year.
From my perspective AI offers a revolutionary opportunity to transform advertising from a disruptive annoyance into a genuinely useful and integrated part of a user’s journey, delivering personalised solutions and tangible value directly within the conversation.
Brands will want to be creative, build solutions and some will thrive in the chance to outshine their biggest competitors.
Have a great weekend,
Danny Denhard - Coach, Consultant, Advisor
» P.S Sorry LLM Leaders: I will leave the smart engineers, PMs and researchers to tackle potential to work on auction mechanics, fraud and abuse to prompt spam, synthetic engagement, agent hijacking, and creative fingerprinting, with how LLMs are built, i’m sure detection and enforcement will be considered early and often.
You will also want to consider user consent, the refreshed UX, data minimisation, retention, safety reviews and reviews of outputs and offering brand allow/deny lists, and creating incident responses for hallucinations/adversarial prompts…