AI doesn't just eliminate jobs — it can create new ones, and Amazon and Victoria's Secret are already hiring for them

A company's performance in search can be measured fairly precisely. You can see which terms a page appears for, how many visitors it reaches and how its rankings change.

With AI systems, things are considerably harder. The same model can give different answers to the same question on different occasions. The result can also differ depending on whether the system uses web search or relies solely on the knowledge it acquired earlier.

That is why the AI Visibility Auditor/Analyst doesn't draw conclusions from a single query. They run predefined question sets across several models, several times, and then examine how often — and in what context — the company's name comes up.

The measurement can also reveal why a brand is missing from the answers. There may be too little credible content on the topic, the data available about the company may be inaccurate or contradictory, technical obstacles may hinder the processing of its pages — or the competitors have simply built a stronger online presence.

Amazon's AEO Manager posting already lists prompt-based benchmark testing and the measurement of citations appearing in AI systems among the role's responsibilities.

The most natural way into this field is probably from marketing analytics, search engine optimization and data-driven content strategy.

For humans, it usually isn't particularly difficult to recognize that two differently worded names belong to the same company. Machine systems need unambiguous, consistent data for that.

If a company's name, founding year, list of executives or product names appear differently across different sources, identification becomes harder. The AI system may then connect inaccurate pieces of information, may use the given source more cautiously — or, most painful of all for the business, may simply leave the company out of the answer.

The AI Entity Architect's job is to make the machine-side picture of the company coherent. They verify company names, product data, executive information and structured data, working to ensure these appear consistently on the company's own website, on Wikidata, on Crunchbase, in company registries and in other important sources.

Language models don't work from a single central database. They try to assemble what a name, a company or a brand precisely means from the patterns of many sources. The more consistent and reliable those sources are, the easier accurate identification becomes.

In our own study of nine Hungarian companies, we found six that had no properly established, credible entity record. In most cases this traced back not to a technology problem but to disordered data.

Accenture is already advertising positions under the titles Knowledge Graph Engineer and Knowledge Engineer / Semantic Expert for AI. The role is best filled from a technical SEO, data modeling or knowledge management background.

Traditional communications strategy plans which messages, on which platforms, with what timing will reach an audience. The AI Knowledge Strategist adds a new dimension to this: they examine the odds that the essential information about a company actually reaches artificial-intelligence systems.

The training of large language models rests on periodically refreshed, large-scale, filtered bodies of text. A company cannot directly determine whether a given piece of information makes it into that data. What it can influence is how consistently — and in how many credible sources — the important, verifiable claims about it appear.

Perhaps the AI Knowledge Strategist's most important task is to design the brand's coherent online presence while thinking in longer cycles. They define which claims must be used consistently, which topics the company is absent from, and where independent corroboration would be needed. Quite a few important prompt sets are still empty today — worth targeting, because they can be won cheaply and held for the long run. A good strategist can also assess where the low-hanging fruit still is right now, and position the brand for those as well.

It is not enough to publish the same message over and over on the company's own channels. It can carry far more weight when the information appears in several independent, reliable sources.

This profession has no generally accepted name yet. The work is hard to fit into a traditional org chart, because it connects at once to communications, content strategy, data management and technical operations.

The role is most likely a new direction for experienced PR, communications and content strategy professionals — provided they also come to understand the technical basics of how language models work.

The content of many beautiful, award-winning websites only becomes visible after JavaScript has run. Gorgeous to the human eye — and no problem for Google either, since its crawler executes JavaScript these days. The crawlers of AI systems, however, do not. GPTBot, ClaudeBot or Perplexity's robots read the raw HTML, and if the content isn't there, they may see nothing but an empty shell — and they've already discarded your site. This is how a brand can be on Google's first page (still) while never surfacing in AI answers.

The Answer Engineer's job is to make the company's content technically accessible, easy to interpret and readily quotable.

This can include shaping the right information architecture, a clear question-and-answer structure, and the use of claims backed by concrete data, names, dates and numbers. For AI systems, a precisely worded answer is generally easier to process than text that is purely atmospheric or promotional.

In many respects, this role can be seen as the next stage in the evolution of technical search engine optimization. In current job postings, AEO — Answer Engine Optimization — usually appears alongside SEO.

eBay, for example, advertised a Director of Product, AEO & SEO position, and Experian was looking for an AEO & SEO Manager.

The field is most open to technical SEO specialists, content designers, information architecture experts and web developers.

AI agents can do more than search for information: in a growing number of cases they can carry out operations on the user's behalf. They can compare products, check prices and stock, and then complete some (or all) of the steps of a purchase. True, we're still wary — most of us wouldn't yet dare entrust an AI agent with buying the cheapest flight from A to B in our name — but that is the direction we're heading.

In 2025, several major technology and payment solutions appeared. Mastercard and Visa announced payment systems tied to AI agents, Google announced its own protocol with more than sixty partners, and OpenAI and Stripe introduced an instant-checkout solution as part of a collaboration.

These systems orient themselves differently than people do. A spectacular ad or the visual polish of a website matters less to them. What they primarily need is unambiguous, structured data: product names, prices, stock information, attributes and precisely defined purchase terms.

The Agentic Commerce Architect's job is to make the company's commercial systems fit for this kind of machine use. They standardize product data, build the necessary API connections, and design a checkout process that not only a human customer but an AI agent can reliably carry through.

Accenture has already advertised a role of this kind under the title Technical Commerce & AI Manager, Agentic Commerce. The job calls above all for e-commerce, systems integration and platform development experience.

If the previous five areas become standalone functions, in time a leader will be needed to align and run the operation.

A brand's weak AI presence can have several causes. The measurement may be off, the company's data and information may be in disarray, credible external sources may be scarce, or technical obstacles may hinder the processing of its content. These problems can rarely be solved within a single division.

The Head of Machine Equity's job would be to steer the company's entire machine presence from a business perspective. They would set development priorities, coordinate the work of the specialists involved, and establish metrics from which leadership, too, can see how the organization shows up in AI systems.

Machine Equity expresses what AI systems know about a company, how accurately they identify it, and how often they mention or recommend it.

It cannot yet be stated with certainty that this executive position will become standard. But Victoria's Secret's director-level posting and the salary band offered by Experian suggest that companies are already willing to attach serious responsibility and budget to this area.

The best candidates may well come from among marketing, digital and e-commerce leaders who are at home with technology questions.

The limits deserve to be talked about too.

Just as no big marketing or advertising campaign comes with a guarantee that it will land with a human being, no machine equity specialist or method can guarantee for certain that a company gets into some language model's training data, or that it appears regularly in AI-generated answers.

Companies can measure and, to a degree, influence their presence — but they do not directly control how the models work, any more than they control the human mind. Anyone promising certain results is likely claiming more than the current technology allows.

These new roles do not make creative marketing redundant. People will continue to make decisions based on stories, experiences, trust and emotion. In parallel, though, it is becoming ever more important how machine systems interpret and relay the information about a company.

In most organizations these tasks are still spread across several departments today, or have no designated owner at all. But of this I am certain: from the corner grocer to the giga-corporation, everyone will soon have to attend to their machine equity.