How to build an AI team: the roles retailers need to get AI right
With AI becoming a core part of digital commerce, it’s forcing brands into a new kind of question.
Not “what AI tools should we use?” but “who should own AI inside our business?”
We see two extremes when brands start this journey.
They either allocate a sizeable budget to set up a tight-knit AI task force of one or two people or delegate the challenge to a few interns.
Put simply, brands and retailers are wondering how to assemble a dedicated AI team and what the roles and processes should look like.
Before we dig into the detail, if you want AI to deliver commercial impact, you start with the team, not the tech.
The need for AI skills is quickly accelerating. The Future of Jobs Report, published by the World Economic Forum, predicts that by 2030, shifting global trends in technology and the economy will create 170 million new roles while displacing 92 million others.
Some of the fastest-growing positions are in technology, data, and AI, which makes the question of who owns AI inside your business more urgent than ever.
According to McKinsey, nearly half of organizations are already shifting business goals to capitalize on AI opportunities, and 62% are hiring AI specialists to strengthen operations.
Start with ownership: the AI product leader
Like most programs, it begins with a single point of accountability.
In large enterprises this may be the chief AI officer, the CTO, head of data, or head of digital product.
This leader defines why the business is using AI, where it creates leverage, and how it supports the commercial roadmap. This may include prioritizing search relevance, personalization, pricing optimization, fraud reduction, or content automation.
This role ensures AI connects to measurable outcomes: conversion rate, AOV, retention, operational efficiency, etc.
They also own governance. Ethical use, regulatory alignment, data responsibility, and model risk management all rest at the leadership level.
Without this role, AI projects will unravel. But with it, AI can become a product capability.
The AI product manager
While leadership sets direction, an AI product manager will deliver it, translating commercial needs into AI features.
This role decides what problems models should solve, how they appear in customer journeys, and how success is measured.
This is likely to include how LLMs learn intent, supports merchandising, and aids purchasing products.
They also manage in uncertainty, as not all AI projects will work and scope can change fast.
An AI product manager owns prioritization, the experimentation, backlogs, and outcomes. They make sure AI maps to revenue, CX, and other KPIs.
The data engineers
Before models can be trained, data needs to be reliable, structured, and accessible.
Data engineers build the pipelines that collect and prepare data from across the commerce stack (such as product catalogs, customer behavior, transactions, inventory, marketing systems).
They ensure data flows cleanly between platforms and is available in the right format for AI models to use.
In many organizations, this role determines whether AI projects move quickly or stall at the data preparation stage.
The technical team: data scientists & ML engineers
Once the direction is laid out and ownership in place, execution can begin.
Data scientists are tasked with exploring datasets, training algorithms, testing hypotheses, and fine-tuning performance. This could be product recommendations, demand forecasting, price optimization, etc.
While data scientists develop and train models, machine learning engineers ensure those models actually run inside the business, so they can operate reliably at scale by integrating them into the wider technology stack – commerce platforms, CDPs, recommendation engines, pricing tools, etc.
More specifically, this could include deploying recommendation engines into product discovery, personalization models into the storefront, demand forecasting, and embedding LLMs into customer service or merchandising tools.
The AI ethicists
While data engineers and ML teams make AI work, ethicists make sure it works responsibly.
Their role is to anticipate unintended consequences and assess the broader impact of AI on customers, employees, and the business.
They determine how to manage risks and opportunities of AI use: bias in models, privacy concerns, the fairness in algorithms, and regulatory compliance.
This may include topics like how personalization might unintentionally exclude certain customer segments, or how automated decisions could create reputational or legal risks.
AI ethicists also embed guardrails into AI workflows, advising on data usage policies and monitoring frameworks, helping to translate abstract ethical principles into clear business practices.
The structure of your team
The optimal AI team structure depends on the size of your business as well as its AI maturity.
For large organizations, data engineers, data scientists, and ML engineers are already essential. But as AI adoption grows, additional roles, such as AI ethicists, prompt engineers, and AI product managers, will become critical to manage complexity and maintain impact.
In the early stages, a single, cross-functional team under a product or AI lead allows you to move fast and iterate.
As the business scales, functional groups may be introduced to handle the volume and complexity, although care must be taken to prevent silos emerging.
Some enterprises have established a centralized Center of Excellence that collaborates across product lines, while others embed AI talent directly within each product team to keep expertise close to delivery.
Emerging best practices
With artificial intelligence evolving at pace, it’s difficult to define best practices that won’t be outdated within months or even weeks.
However, there are a set of emerging principles that leading organizations are adopting to build more effective AI teams. Here’s the core:
- Share AI knowledge
Upskill your teams on AI fundamentals. Not everyone needs the knowledge of a data scientist, but basic understanding of AI concepts and limitations helps to foster a common language and facilitate smoother collaboration across teams.
- Integrate good governance
AI should, fundamentally, be responsible. Go through potential biases and discuss all compliance requirements from the beginning. Establish checkpoints for reviews and tests. In regulated sectors, ensure someone monitors evolving AI standards.
- Continuous learning
Because AI’s evolving rapidly, ongoing learning is perhaps the most important best practice. Encourage teams and individuals to do training, courses, and research, and allocate time for experimentation so that teams can develop new frameworks and test algorithms.
Where to begin
Building a team for AI is all about clarity. Clarity of ownership, clarity of outcomes, and clarity on how AI fits into your commerce strategy.
The mistake many retailers are making is treating AI as an isolated capability or a side project.
Those that are beginning to see real impact are embedding AI across teams that are aligned around practical, measurable outcomes.
That doesn’t mean building a large and complete team from day one. Instead, it starts with the right ownership and use cases, before it can then scale.
If you’re evaluating how to structure your AI team or where AI can drive the most impact in your business, get in touch with Tryzens.
We work with retailers to define AI strategies and help deliver sustainable digital commerce growth.