Yes. AI now drafts, sorts, targets and reports faster than any person, but it cannot decide what a brand should stand for, judge whether a claim is true and appropriate, or take responsibility for what is published. Marketing that works uses AI for speed and volume, and people for judgement. The more of the routine work machines take on, the more those human parts decide whether the result is any good.
What does AI do well in marketing?
A lot of marketing is repetitive, and that is where AI earns its place:
- First drafts and variations. Ad headlines, product descriptions, email subject lines and social captions in many versions, quickly.
- Summarising data. Turning campaign numbers into a readable weekly report.
- Bidding and targeting. Ad platforms already use machine learning to set bids and find audiences.
- Answering routine questions. Opening hours, prices, order status and booking requests, at any hour.
- Sorting and qualifying leads. Asking the first questions and routing an enquiry to the right person.
Our article on what actually works with AI in marketing goes through these in more detail, including the uses that sound good in a demo and rarely survive real customers.
Where does marketing still need a person?
Knowing what is true
AI writing tools can state wrong prices, invent features, misquote regulations and describe services you do not offer, all in fluent, confident prose. Someone who knows the business has to check every factual claim before it goes out. A mistake in an advert or a price list is your mistake, not the software's.
First-hand experience
AI has not used your product, visited your site or spoken to your customers. Google's guidance on AI-generated content says its systems aim to reward original, high-quality content demonstrating expertise, experience, authoritativeness and trustworthiness, "however it is produced". Experience is the part a machine cannot supply: the photo from the job, the detail only the person who did the work would know, the honest view of when your product is not the right choice.
Taste and positioning
Deciding what a brand stands for, what it will never say and which idea is worth backing is a judgement about people and markets. AI tends towards the average of what it has seen, which is useful for a first draft and poor for standing out.
Context and tone
A joke that works in one market offends in another. A cheerful automated reply to an angry complaint makes things worse. Festivals, local language, current events and the mood of a particular customer all need someone paying attention.
Accountability and disclosure
Platforms increasingly expect people to be open about AI use. YouTube, for example, requires creators to disclose content that is generated or meaningfully altered with AI when it appears realistic. A person has to own those decisions and the consequences.
How should people and AI share the work?
| Task | AI is good for | A person must |
|---|---|---|
| Blog posts and articles | Outlines, first drafts, summarising sources | Add first-hand knowledge, check facts, decide what is worth publishing |
| Ads | Headline and description variations, bid adjustments | Set the offer, budget and limits; approve claims |
| Social media | Caption options, scheduling, reply suggestions | Choose the stories, handle complaints, judge timing and tone |
| Customer chat | Routine questions, collecting details | Take over anything unusual, emotional or high-value |
| Reporting | Pulling numbers, spotting changes | Explain why it happened and decide what to do next |
What goes wrong when the person is removed?
- Sameness. Content generated without direction reads like everyone else's, because it is built from everyone else's.
- Confident errors. A wrong price or invented claim published at scale is harder to undo than one written by hand.
- Search penalties. Google's guidance says that using automation, including AI, to generate content with the primary purpose of manipulating search rankings violates its spam policies.
- Customers trapped in loops. A chatbot with no route to a person turns a simple problem into a lost customer.
How do you keep the human touch while using AI?
- Name an editor. Every piece of content has one person responsible for approving it.
- Write a short brand voice guide. Words you use, words you avoid, how formal you are, so drafts start closer to your voice.
- Keep a fact sheet. Current prices, services, locations and policies that every draft is checked against.
- Add what only you know. Real photographs, customer stories with permission, opinions you would defend.
- Give every bot an exit. Our AI chatbots are designed to hand a conversation to a person when it needs one; whichever you use, test that the handover works.
- Disclose where it matters. Follow each platform's rules on AI-generated media, and do not present synthetic people or scenes as real.
If you are deciding which parts of your marketing to automate, our AI automation team can help you separate the routine work from the judgement calls. For a closer look at how search itself is changing, see will AI replace SEO?
Questions people ask about AI and the human touch
Will AI replace marketers?
It is replacing some tasks, such as first drafts, routine reporting and basic bid management, rather than the job. Strategy, positioning, fact-checking and customer relationships still need people, and they matter more as routine content becomes cheap.
Does Google penalise AI-written content?
Google says it focuses on the quality of content rather than how it is produced. Using automation mainly to manipulate search rankings breaks its spam policies, while helpful, original content can do well however it was made.
Should businesses tell customers when they use AI?
Where customers would reasonably expect to know, yes. Some platforms require it: YouTube asks creators to disclose realistic content made or meaningfully altered with AI. A chatbot should also be clear that it is a bot and offer a way to reach a person.
Which marketing tasks should stay human?
Deciding what the brand stands for, approving factual claims, handling complaints and sensitive conversations, and judging tone in context. These are the points where a mistake costs trust, and where accountability has to sit with a person.
Further reading
Official guidance on the subject, from the platforms themselves: