What is a marketing engineer? What the role does and how to tell a real one.

01What is a marketing engineer?

A marketing engineer is a marketer who builds and runs the software systems, now mostly AI agents, that produce a company's marketing, and who keeps the judgement over what those systems ship. The role combines two halves: the technical skill to build and fix agents, and the marketing taste to decide whether their output is any good.

In practice, a marketing engineer replaces manual production with systems: a content engine that drafts from the positioning brief, an outbound agent that researches prospects, a search agent that fixes how the business appears in AI answers. The engineer builds each one, reviews what it produces and rewrites the instructions when output drifts. If the term AI marketing agent is new to you, start with that definition, because agents are most of what a marketing engineer works on.

AI marketing engineer means the same thing; almost every system in the role now runs on a language model.

02Why is the marketing engineer title suddenly everywhere?

The title spread because AI agents made it possible for one marketer to build systems that used to need a developer, and companies started advertising for that combination. The growth is real but starts from a small base: by the most specific count available, fewer than two hundred people took the title in the past year.

The Wall Street Journal covered the shift on 30 September 2026, in a report by Katie Deighton on marketers rebranding as engineers. The Journal reported that LinkedIn job listings using "marketing engineer" rose 29% in 2025. It also cited Live Data Technologies data showing 173 people started jobs titled marketing engineer in the tech and marketing sectors in the 12 months to August 2026, up from 98 a year earlier. That is close to double, and still tiny.

Some of the push comes from vendors. The AI search company Profound runs a marketing engineer job board and podcast on Profound's own site¹, and the Journal reported that its co-founder, James Cadwallader, described the role at Profound's April 2026 conference as part builder, someone who can build, deploy and tinker with agents, and part marketer, someone with taste and judgement. Figma advertised a marketing engineer role in 2026 and, per the Journal, said success would be measured by agents deployed, the time those agents save people, and the revenue they drive.

Adoption numbers show the gap. Salesforce's tenth State of Marketing report, surveying 4,450 marketers in late 2025, found only 13% using agentic AI². Many marketing teams use AI tools; few have anyone who builds agentic systems and keeps them running.

03What does a marketing engineer do in a week?

A marketing engineer's week splits between running the systems already live and building or fixing the next one. Most of the running time goes on review: reading what the agents produced, approving or rejecting it, and turning rejection reasons into better instructions. Building happens in focused blocks once the weekly review is done.

A realistic week for one engineer running a small company's marketing looks like this:

• Read last week's numbers: enquiries, pipeline from outbound, which pages AI assistants cited, what each agent shipped.
• Review the content engine's drafts against the positioning brief, edit or reject, and log why.
• Approve the outbound agent's first messages for the week and take over any thread where a prospect replied.
• Check how ChatGPT, Perplexity and Google's AI Overviews describe the business, then fix what is wrong at the source: an outdated price, a service page that is mostly images with no descriptive text.
• Build or extend one system, for example an agent that summarises competitor pricing pages every Monday, or a paid media agent that drafts ad variants for approval.
• Fix whatever broke, such as a workflow that stopped firing after an API change.
• Write a one-page report for the owner or marketing lead: what shipped, what it produced, what to stop.

The Journal's report lists the same kind of work: AI systems that automate competitor analysis, paid media buying and personalised ads, and fixes that change how a company appears in chatbot answers, such as rewriting image-heavy pages into descriptive text and correcting outdated pricing that AI keeps repeating. That last job is answer engine optimisation, and it has become a standard part of the role.

04Marketing engineer vs GTM engineer vs growth engineer: what is the difference?

The five roles below overlap in tools but differ in what they build and what they answer for. A marketing engineer builds the systems that produce marketing. A GTM engineer builds the systems that feed sales. A growth engineer changes the product to lift sign-ups. Marketing operations and SEO managers run narrower parts of the same machine.

The Journal noted that "go-to-market engineer" is now used for sales operations work, and "narrative engineer" has appeared too. When you read a listing, the "what they build" column tells you more than the title. In small companies, one person often covers both marketing engineer and GTM engineer work.

05What skills does a marketing engineer need?

A marketing engineer needs working technical skills to build agents and keep them running, and senior marketing judgement to decide what they should produce. Either half alone produces a different role: a developer who ships plausible marketing nobody wants, or a marketer whose ideas stay in a document because nobody builds them.

On the technical side:

• Building with language models: agent instructions, tool use, retrieval and testing output against examples.
• A scripting language, usually Python or TypeScript, plus APIs and webhooks.
• Workflow tools such as n8n or Zapier, and the judgement to know when code is the better choice.
• Marketing data: analytics events, CRM fields, attribution, and enough SQL or spreadsheet skill to check a number.
• The technical side of AI search, such as crawler access and structured data.

On the marketing side:

• Positioning: who the business sells to and what it claims, with the proof, written down so agents can read it.
• Editorial taste: spotting generic copy and unsupported claims before a customer does.
• Channel knowledge across search, content, email, outbound and paid.
• Compliance sense: consent rules for email and messages, and privacy rules for lead data.

Many arrive from marketing and learn the build side. The guide to AI for marketers covers that path.

06Where is the marketing engineer role heading in 2026?

The fastest-moving version of the role is AI deployment inside a marketing team: one builder who ships agents that the team's marketers use every week, keeps the prompts and skills behind them in version control, and is measured on adoption. The title varies by company, but the job descriptions now read very alike.

Figma's listing for a Marketing Engineer, AI Deployment³ asks the hire to own a GitHub-managed library of prompts, skills, context files and agents, and measures success by agents live in recurring use, hours saved and revenue driven. It lists a base range of USD 127,000 to 296,000 and treats SQL across a warehouse such as Snowflake or BigQuery as a plus. Plaid's AI Marketing Technologist Lead⁴ combines building agents for marketing workflows with training the team to work alongside them, at USD 170,400 to 223,200.

Two things follow for anyone hiring or applying. The portfolio matters more than the title, because a versioned skills library and agents in daily use are easy to inspect. And the change work counts: getting marketers to rely on a system is part of the job description now, alongside building it. For how this looks in New Zealand and Australia, where the exact title barely appears in job ads yet, see the 2026 data on marketing engineer and AI marketing jobs in NZ and Australia.

07Is marketing engineer a real role or a rebrand?

It can be either, which is why the title alone proves little. Some people with the title build and run agents every week. Others have relabelled existing marketing work. Both camps are in the Journal's reporting, and both have a fair point. The difference shows up in what the person can point to.

The sceptical case, from a brand consultant quoted in the Journal, is that the title can be a surface rebrand of work marketers already do, one more step in a long run of reworked job titles. For anyone hiring, that is a fair warning.

The other side argues the job content has changed. A Google marketer wrote on LinkedIn that marketing engineers are the hire of 2026. Profound's argument, per the Journal, is that software engineers are often the wrong fit because marketing needs discernment, and a human has to orchestrate the agents and judge what they build. Survey data agrees: HubSpot's 2026 State of Marketing, surveying 1,500+ marketers, found 73.4% see AI working in conjunction with marketers, assisting them⁵. Someone still has to decide what the assistance produces.

08How to tell a real marketing engineer

My view, as someone who does this job: the title proves nothing. The role is real when the person can show you a running system with its weekly output, plus the human judgement that decides what ships. The six checks below work for recruiters screening candidates and for owners choosing someone to build their marketing.

I run a whole marketing function this way, one person plus an agent stack I built: the systems do the production and I do the judgement. When I look at someone else's work, I ask these questions:

1. Shipped systems you can inspect. Ask to see a repository, a skill library, a live workflow or a published case study. A screenshot of a chat with a model does not count.
2. A measured weekly number. Time saved, articles shipped, meetings booked, enquiries from a page. The person should know the number without looking it up. Salesforce reports agents free marketers 6 to 7 hours a week² (surveyed late 2025); a real engineer can tell you what their own systems save.
3. Ownership of AI search visibility. They should know how ChatGPT and Google's AI answers describe the business today, and what they changed to fix it.
4. They write the brief the agents read. Positioning, voice rules, banned claims and examples, in a document every agent loads. If nobody owns that brief, the agents drift.
5. They know when to leave something manual. Ask what they chose not to automate and why. Gartner predicts over 40% of agentic AI projects will be cancelled by the end of 2027 due to escalating costs, unclear business value or inadequate risk controls⁶. Knowing what to skip keeps a project out of that group.
6. They review every send. A named person approves what reaches a customer. KPMG and the University of Melbourne found 51% of New Zealand employees had relied on AI output at work without evaluating its accuracy⁷ (surveyed over summer 2024-25). The human in the loop is the part of the role you are paying for.

The honest limit: hiring numbers for the role are still small, and most small businesses will meet a marketing engineer as a service or a part-time contractor before they hire one full-time. That is a sensible order. Buy the systems, see them run, then decide whether the work needs a permanent seat.

09How to hire a marketing engineer: a job description outline

A good marketing engineer job description names the systems the person will build in their first 90 days and the numbers they will answer for. Generic lists of AI tools attract generic applicants. The outline below can be lifted into a listing and adjusted to your business.

1. Role summary. One paragraph: you will build and run the AI systems that produce our marketing, and own the judgement over what they ship.
2. First 90 days. Name two or three systems, such as a content engine briefed from our positioning, an outbound agent with human takeover on reply, and fixes to how AI assistants describe us.
3. Ongoing responsibilities. Weekly review of every agent's output, maintenance when tools change, the positioning brief, a monthly report.
4. Must-have skills. Built and deployed at least one agent or automated system in production; a scripting language; marketing data; written positioning or editorial work.
5. Useful extras. AI search optimisation, paid media.
6. How success is measured. Systems running each week, hours saved, enquiries or pipeline influenced, accuracy of the business in AI answers.
7. Who they report to. Usually the founder or marketing lead, who signs off positioning.
8. Working arrangement. The work suits remote arrangements. State the overlap hours you need for reviews.
9. Interview task. Ask candidates to walk through a system they built: what it reads, who reviews it, what broke and how they fixed it. A short paid build task beats a whiteboard exercise.

Check pay against a current salary guide for your market. The title covers senior marketers with new skills and developers moving into marketing, so ranges vary widely.

10Does a small business need a marketing engineer?

A business of 5 to 50 people needs what a marketing engineer produces, a set of systems that run the marketing every week, more than it needs the full-time seat. For most, the right first step is buying the systems built on its own accounts, then hiring once several run weekly and need constant tuning.

A full-time hire makes sense when several systems run and someone senior can sign off positioning. Before that, a part-time engineer or a service can build the first marketing agents and train a person inside to review them. If the question is who plans the work above the agents, the guide to the AI CMO covers that layer.

Either way, every system and prompt should live on accounts the business owns.

11Where Flow AI fits

Flow AI is an AI marketing transformation practice: the strategy, the systems and the agents that run a marketing function, engineered end to end. Amber Lan is an AI marketing engineer in Auckland who builds those systems and reviews everything they send. The skill library is public, so you can inspect the work before you talk to anyone.

For teams, the written scope names the first system to install. For recruiters and hiring teams with a marketing engineer or GTM engineer role, Amber is open to conversations about senior roles, remote from anywhere.

12The short version

A marketing engineer builds and runs the AI systems that produce a company's marketing and decides what those systems ship. The title grew quickly in 2025 and 2026, from a small base, and it overlaps with GTM engineer, growth engineer and marketing operations.

Judge the person by what is running. Ask for a system you can inspect and the number it moves each week, then ask who reviews every send. For a small business, buying the systems first and hiring later is usually the better order.

To find out which system your marketing needs first, talk to Amber.