Simon Kingsnorth published Marketing in Web 3.0 in 2024. The book mapped marketing’s future onto AI, the metaverse, blockchain and a decentralised internet. It also came with a new cast of specialists: AI, virtual reality, blockchain, communities, data and the Internet of Things.

Two years later, parts of that list already feel like a period piece. The AI specialist survived. The metaverse developer, not so much.

The jobs had been named after the technologies. That gave them roughly the same shelf life.

I could easily make the same mistake. Guessing which software companies will use in 2030 is mostly theatre. The better question is what decisions people will still have to make once the software can do more of the work.

More automation, more responsibility

In a Gartner survey of 402 chief marketing officers, respondents expected AI to automate 36% of marketing work by 2028, up from 16% in 2026.

Companies will produce more ads, copy, video and analysis in less time. Platforms will take over more campaign setup and optimisation. Models will read the data and suggest what to do next.

Some jobs will disappear. The responsibility will not.

If a company can make a hundred variants in an afternoon, someone still has to choose one. If software adjusts the campaigns, someone has to check the decisions. And if every model learnt from the same public internet, someone has to bring in knowledge the company gathered for itself.

Once almost anything is possible, knowing what not to do becomes part of the job.

Marketing lead

The marketing lead decides where the company puts its money, attention and name.

Advertising platforms will keep absorbing work that specialists now do by hand. They will adjust bids, move budgets and produce new ad variants. The marketing lead will spend less time inspecting settings and more time deciding what the company is trying to achieve.

That means questions a platform cannot answer:

• Which market should the company enter?
• What can it credibly promise customers?
• How much can it invest in marketing?
• Which activities should it stop?
• How should marketing support sales?

Saying no will matter as much as choosing well. Software will offer more options than any company can use sensibly. The marketing lead has to narrow them down and answer for the choice.

Editor

The editor decides what deserves to be published.

A model can already produce an article, ad, image or video script in minutes. By 2030, it will do so faster and for less. Companies will have no shortage of material.

That does not mean they will have more to say.

The editor chooses the subjects worth covering, checks the facts, improves the presentation and removes the material that merely echoes everyone else.

The job starts to look like an editor-in-chief’s:

• choosing subjects that matter to customers
• checking the facts and the meaning of the message
• keeping the company’s voice recognisable
• stopping copy and campaigns that have no clear purpose

A model can write another article. The editor decides whether the company needs one.

Marketing engineer

For want of a better name, I call this person a marketing engineer. They own the technical setup on which the rest of the work depends.

They connect the CRM, analytics, advertising platforms, customer data and AI tools. They decide where the software gets its information, what it can do on its own and which changes still need human approval.

They also catch faults that produce no obvious warning.

Some faults announce themselves. The shop stops taking orders or a campaign never starts. Others sit unnoticed for months. Revenue may be attributed to the wrong source, an old offer may stay in circulation or a recommendation system may push products the company cannot supply.

The marketing engineer:

• connects the systems the company uses
• sets up automated workflows
• checks the data and the output
• looks for faults that routine technical checks do not reveal
• removes software that costs money without doing enough useful work

The job is not to spend all day trying new AI apps. It is to choose a few useful tools, fit them into normal work and keep them reliable.

Researcher

The researcher finds out why customers buy, reject the offer or choose a competitor.

Models will know the same public articles, studies and reviews as everyone else. That knowledge will be easy to buy and hard to turn into an advantage.

A company can still learn something its competitors do not know. Someone has to speak to customers regularly and ask better questions.

The researcher:

• interviews customers
• talks to sales and customer support
• tracks why deals are lost
• compares what customers say with what they actually do
• passes the findings to the people responsible for the offer, communications and sales

AI can transcribe and summarise the interviews. A person still has to ask the question that exposes the real problem. Then the company has to do something with the answer.

So, exactly four people?

No. These are four responsibilities, not a compulsory headcount.

A large company may have a department for each one. In a smaller company, four people may cover the lot. One person may also hold two roles.

The marketing lead might also choose subjects and approve content. The marketing engineer might own both analytics and automation. The researcher might spend part of the week improving the offer.

The number matters less than knowing who is responsible for each decision.

The greatest pressure will fall on jobs built around one channel or a repeated task. Parts of junior design, social media management, PPC and routine reporting all fit that description.

The work will not vanish overnight. One experienced person will simply be able to do much more of it with software.

This is already visible. I meet PPC specialists who also manage analytics, edit copy, test AI tools and work with sales on what happens to a lead. Marketing managers coordinate an internal team, agencies, a website and several outside suppliers. Their job title covers only part of the job.

By 2030, there will be more people like them. Deep expertise will still matter for difficult work. Routine work will need fewer people, each with a wider view.

Four questions for today

There is no need to wait for 2030 or redraw the organisation chart. Start with four questions:

• Who decides the direction of marketing and the budget?
• Who approves what the company publishes?
• Who is responsible for data, automation and the connections between the systems in use?
• Who speaks to customers regularly and shares what they learn with the rest of the team?

The same person may answer more than one. That is normal in a smaller company.

A blank answer is more revealing. That is where work disappears between management, the team, the agency and the software. Everyone completes a task, but no one owns the result.

Once the gap is visible, the company can give the responsibility to someone already there, hire for it or bring in outside help.

A marketing team ready for 2030 does not need exactly four people. It needs an answer to all four questions.