Your next buyer may be an algorithm but, please, don’t write advertising for one

15 July

Tech brands now have two audiences: humans who need to care, and machines that need to understand. The problem is that most advertising manages only to bore the first and confuse the second.

Take a moment to do a scan of tech advertising and you’ll probably be struck by how remarkably generic it is. There’s a stock image of several attractive professionals staring thoughtfully at a screen and a headline containing some alliterative combination of words like “Secure. Scalable. Seamless.”

Then there will be three paragraphs about “solutions” and a blue button marked “Learn More.”

If that’s what your advertising looks like, well congratulations. You have created digital wallpaper.

The problem with such advertising is not a shortage of information. Tech companies are magnificent at producing product sheets, webinars, whitepapers, capability decks, explainers, emails, comparison tables and enough LinkedIn content to dam a medium-sized river. The problem is that remarkably little of this information gets noticed, remembered or associated with the right brand.

And as if getting some cut through wasn’t already hard enough, artificial intelligence has now arrived and made the problem even more difficult to solve.

Your advertising now has two audiences

Marketing research body WARC has just released its Creative Impact Unpacked 2026 report, which identifies what it calls the “two audience problem”.

The report says that while advertising must still influence human beings it must also, increasingly, survive the ranking, retrieval and recommendation systems used by large language models and AI agents. And these two audiences do not respond to creative work in the same way.

Research presented in the report found no correlation between how humans and large language models judged 480 Creative Effectiveness Lions entries. A separate Kantar study of 800 advertisements found only a 0.10 relationship between consumer and AI assessments.

While humans respond to story, humour, emotion, tension, surprise and recognisable characters, machines look for description, context, hierarchy, authority and corroborating evidence. In other words, while one wants to feel something, the other wants to classify something.

This doesn’t mean your advertising should be turned into machine food. Goodness knows, the internet already contains quite enough keyword porridge. What it does means is that your creative and your broader brand system must now have different but connected jobs. The advertising must make the brand worth remembering and the brand system must make the business easy to understand, verify and recommend.

Humans are not procurement spreadsheets

Many B2B technology companies behave as though human emotion turns off the moment someone receives a corporate email address. It does not.

Even the most pragmatic and rational CTO might be stressed by the task of comparing and evaluating technical capabilities, integration requirements and compliance standards of competing cybersecurity software. They may be worried about choosing badly, having to defend their decision internally, disrupting operations or becoming the unfortunate protagonist in next quarter’s breach notification. In other words, while trying to make a rational choice, the CTO may be burdened by fear or lack of confidence, and concern about their professional identity, social standing and self-preservation.

Yet much of the category continues to advertise as if the buyer were a spreadsheet that had recently become sentient.

In contrast, WARC’s report highlights the importance of surprise, dramatic tension and a perceptible “jolt” in earning and focusing a potential customer’s attention. Its analysis also associates product-centricity, flatness and obtrusive words with weaker attention and emotional response.

Responding to this WARC’s insight requires tech brands to produce advertising that contains an idea, has a point of view and an enemy worth naming. It requires advertising that describes a problem from an original angle, not one the customer has already seen a thousand times before.

The IGNITE principle is straightforward: information alone rarely creates a memory. Distinctive, relevant and emotionally engaging expression gives that information somewhere to stick.

Machines do not feel, they infer

Just in case you’ve bought a ticket on the AI hype train, let’s be clear: an AI model cannot feel tension, relief or surprise in the way a human buyer can. It builds associations from patterns.

The AI model looks for the things the company repeatedly say it does, the problems it is associated with and the industries and buying situations that appear around it. It will look for evidence that supports the company’s claims, whether independent sources describe it in similar terms, and whether the company’s language remains coherent across its website, case studies, product pages, media coverage, customer reviews and executive commentary.

This means that brand building is a larger task than simply advertising.

While paid media may introduce the promise, the website explains it, customer experience tests it, and case studies substantiate it. Earned media amplifies the promise and independent commentary either supports it (or pokes it with a sharp stick).

WARC argues that reputation, earned attention and customer experience now contribute significantly to how brands are interpreted by AI systems. It also presents an analysis attributing a large share of LLM visibility to long-term brand equity (although the report does not provide enough methodological detail to treat the exact percentage as a universal law).

Despite that, the broader conclusion is still sound. Authority cannot be manufactured with metadata alone and an AI visibility strategy built entirely around technical optimisation will be only partially successful.

The worst response is to split the brand in two

One tempting response to the two-audience problem is to create emotional advertising for people and dense informational content for machines. That approach can, however, easily result in the development of two, unrelated, brands. While one brand is colourful, clever and full of cinematic ambition, the other lives on the website, where it wears a grey cardigan and speaks fluent committee.

This is where coherence matters. Coherence does not involve saying the same sentence everywhere until the market begs for mercy, it means that every expression of the brand contributes to one recognisable whole. While different messages can serve different audiences, products and buying stages, they still need to share the same strategic promise, category associations, distinctive assets, voice, proof system and view of the customer’s world.

WARC describes coherence as the best response to media fragmentation. It cites Les Binet’s distinction between mechanical repetition, and brand activity that collectively adds up to something larger.

This is particularly important for technology businesses, where marketing is often scattered across product teams, founders, sales decks, channel partners, performance agencies, content freelancers and enthusiastic employees who have recently discovered Canva. Without coherence, each new output starts from zero.

But with a coherent brand, each output strengthens what came before it.

An IGNITE response to the two-audience problem

For IT, SaaS and cybersecurity brands, effective advertising now requires six connected disciplines.

Start with real buying situations.
Link your brand to moments that cause potential buyers to enter the market. It might be an audit failure, rising cloud costs, a compliance deadline, an acquisition, a breach, an incumbent system reaching its limits or a board which is demanding evidence of risk reduction.
Create something worth noticing.
Use tension, surprise, humour, character or a sharp category observation. “Professional” is not a synonym for anaesthetic.
Make the brand unmistakable.
Build and repeat distinctive visual, verbal and sonic assets. Attention without correct brand attribution is a generous donation to the category.
Make the proposition explicit.
Both a buyer and a machine should be able to easily understand what the brand does, who it serves and when it is relevant.
Put proof beside the promise.
Use customer evidence, quantified results, certifications, demonstrations, technical validation and credible third-party support. “Trusted leader” is not proof, it is an adjective wearing a lanyard.
Build one, coherent, system.
This is the point where brand science, creative craft and machine readability meet. Advertising, website copy, product experience, sales materials, customer stories and public commentary should reinforce the same commercial meaning.

A giant pile of keywords, a beautiful campaign disconnected from the product and 200 pieces of content will collectively say nothing if there is no brand coherence.

A practical test before you advertise

Before approving the next campaign, ask:

Will a relevant buyer notice this without already caring about our category?
Is there an idea here, or merely information?
Can people recognise the brand before seeing the logo?
Is the advertising linked to a genuine buying situation?
Can our central claim be understood in one plain sentence?
Does credible proof sit close to that claim?
Would customers, employees and independent sources describe us in broadly compatible terms?
Are we giving the campaign enough reach and time to build memory?
Does this strengthen the brand system, or add another loose tile to the marketing mosaic?

A weak answer to several of these questions is not a media problem, it’s a strategy problem wearing an advertising budget.

Humans must want you. Machines must understand you.

Whether we like it or not, it seems inevitable that AI will influence more research, discovery and shortlisting. But how quickly and extensively this happens in complex enterprise buying remains uncertain.

People, committees, procurement processes, advisers and existing relationships are not about to evaporate in a puff of machine-generated efficiency, so it doesn’t make sense to abandon human creativity.

The best response to the rise of AI search and the two-audience problem it creates is to make your brand more coherent. You should create advertising with enough humanity to earn attention and lodge in memory, while also building information, evidence and reputation with enough clarity that machines can retrieve and interpret it accurately.

After all, while your next buyer may ask an algorithm who belongs on the shortlist, the customer still has a pulse.

Sign-up to WARC to download the full report “Creative Impact Unpacked - 10 Effectiveness Trends from Cannes LIONS 2026”

Grant Belcher