Interview with Juliana Vail, Chief Customer Officer at AIUTA
AI creates lasting value when leaders redesign workflows, align teams and keep customer outcomes, human judgement and accountability at the centre.
For Juliana Vail, Chief Customer Officer at AIUTA and a retail technology leader with experience at FARFETCH, Gilt Groupe and Creative Force, successful AI transformation is less about adopting the latest tools and more about redesigning how organisations work. Her experience has shown that technology alone cannot deliver lasting value. Businesses must align people, processes and technology around measurable customer outcomes, establish clear ownership and preserve human judgement where it matters most.
Every major wave of technological change reshapes the way businesses operate. What was the turning point that changed the way you think about leading organisations through transformation?
I started my career as a creative, not a technologist. My first management role was in post-production at Gilt, where our team went from retouching hundreds of images a day to processing thousands, essentially overnight.
That’s when I started to understand. Technology is a critical enabler, but transformation is ultimately about redesigning how people work around that technology. You cannot increase creative output tenfold simply by asking people to work harder. You have to reconsider the entire operating model: the workflow, the division of responsibilities, the quality controls and the way technology fits into each stage of the process.
Adopting software is relatively easy. Creating an organisation that can use it effectively, responsibly and at scale is much harder.
That experience taught me to work at the intersection of creativity, technology and operations, rather than treating them as separate functions. It has been the through line of my career ever since, from building creative operations at Farfetch to the work we are doing today at AIUTA.
The turning point was not the arrival of a particular tool. It was recognising that the leaders who succeed through periods of technological change are the ones who redesign the workflow and consider the people within it. Adopting software is relatively easy. Creating an organisation that can use it effectively, responsibly and at scale is much harder.
AI has moved beyond experimentation, yet scaling it across an organisation remains a challenge. What separates businesses that create lasting value from those that struggle to move beyond the pilot stage?
Organisations need a clear vision and cross-functional alignment before they begin a pilot. It is very easy to think, “AI can do everything,” and start window-shopping for tools without first identifying the specific value you are trying to create.
Innovation in general, and AI in particular, generates excitement across an organisation. That is positive, but excitement without structure often results in duplicated experiments, fragmented ownership and pilots that are interesting but disconnected from a meaningful business priority.
Farfetch was an extremely technology-forward company, and people at every level were constantly exploring new ideas. What I learnt from that environment was that curiosity needs a framework. A business must be able to say: this is the problem we are solving, this is the customer or commercial outcome we expect, this is how we will measure it, and this is who owns the result from beginning to end.
The organisations that create lasting value also design pilots with scale in mind. They do not treat the pilot merely as a test of whether the technology works. They use it to test the entire operating system around the technology: data, integrations, workflows, quality standards, governance, ownership and employee adoption.
A successful pilot should provide evidence that something is technically possible, and that it can become a repeatable part of the business. Otherwise, it’s not just a demonstration, not a transformation.
AI projects often fall short even when the technology itself works as intended. What do you believe are the biggest reasons these initiatives fail to deliver meaningful business impact?
Many organisations buy compute and expect an outcome. But it doesn’t work that way.
The technology can perform exactly as designed and the initiative can still fail because no one has taken responsibility for the quality of the final deliverable placed in front of the customer. A model may generate an image, a recommendation or a piece of content according to its instructions, but that does not automatically mean the result is accurate, appropriate or consistent with the brand.
I recently saw a retailer publicly criticised for a visibly distorted product image. The issue appeared to have occurred during an automated reformatting or production process without adequate human review. Whether or not the system completed its technical task correctly was irrelevant to the customer. The outcome was still wrong, and it was the retailer’s reputation at risk, not the technology provider’s.
That is the gap many initiatives overlook: the last mile between a technically valid output and a customer-ready experience.
At AIUTA, we quality-check 100 per cent of the imagery we produce and rigorously validate virtual try-on experiences before they go live. We are not simply providing access to raw compute; we are responsible for helping deliver a finished experience that meets an agreed quality standard.
The initiatives that fail tend to treat AI as an “easy button”. The ones that deliver meaningful value retain clear controls, keep human judgement where it matters and establish ownership of the outcome. Brand, quality and customer trust all live in that last mile.
Buying AI has become easier than deploying it at scale. What organisational barriers are leaders still underestimating, and how can they overcome them?
The opportunities presented by AI are easy to understand and easy to become excited about. Its limitations are far less well understood, and that gap is where many deployments quietly fail.
Leaders can often articulate what they hope AI will do for the business. Still, they may not be able to explain where the technology is likely to break, where it could produce inconsistent results or where human intervention will remain necessary. As a result, they build business cases around an idealised version of the capability rather than its real-world performance.
At AIUTA, we incorporate education into the sales and onboarding process. We work closely with clients to define the ideal outcome, but we also engineer the “first mile” and “last mile” processes around the technology. That means understanding the quality and structure of the inputs, as well as establishing the review, refinement and delivery steps required to ensure that the final output exceeds expectations.
The second major barrier is organisational structure. Responsibility for how a product is visualised, marketed and managed often sits across creative, e-commerce, product, technology, data and legal teams. In many companies, it is unclear who owns the exploration of AI across the shopping experience.
That lack of ownership is becoming more problematic as the traditional customer funnel gives way to a more fluid experience. Discovery, inspiration, product evaluation, personalisation and purchase no longer take place in neatly separated stages.
Overcoming this does not necessarily require creating another department. It requires appointing an accountable owner with the authority to work across functions, establishing clear decision rights and agreeing on shared measures of success. AI cannot scale when every team controls one small part of the process but no one owns the outcome.
Innovation is often measured by how quicklynew technologies are adopted. In your view, what should leaders prioritise to ensure AI creates better customer experiences rather than simply adding another layer of technology?
Internal adoption is the wrong primary metric. The first question should be: what matters to the customer, and which customer behaviours have the greatest impact on the business?
Retailers need to connect AI investment to meaningful outcomes, such as stronger engagement, higher conversion, increased basket size, improved product confidence, lower returns, faster time to market or more efficient content production. The appropriate measure will depend on the use case, but it should always be connected to customer value and, ultimately, the profit and loss statement.
The lesson wasn’t that creative craft no longer mattered. It was that internal assumptions should not be confused with evidence of what customers actually value.
At Farfetch, we conducted user testing comparing manually created imagery with AI-assisted imagery. Customers couldn’t reliably identify which images had involved AI, and when they were told, it didn’t change their opinion. The lesson wasn’t that creative craft no longer mattered. It was that internal assumptions should not be confused with evidence of what customers actually value.
Internal teams can understandably become protective of existing processes, particularly when new technology challenges established roles or definitions of quality. Those concerns deserve to be taken seriously, but decisions should still be informed by customer testing rather than internal preference alone.
The best AI experiences are often the ones in which the technology itself becomes almost invisible. The customer does not need another feature labelled “AI”. They need a faster, more relevant, more useful or more inspiring experience.
Leaders should therefore prioritise the experience they want to create and work backwards into the technology. When technology becomes the starting point, businesses tend to add complexity. When the customer is the starting point, AI can remove it.
Periods of transformation rarely come with clear answers. When faced with difficult decisions, what principles guide your leadership, and what lessons have shaped the way you approach change?
Put simply, employees are smarter and more resilient than leaders sometimes give them credit for.
Every organisation needs to get ahead of the conversation about how AI is likely to change its operations. It is easy for leaders to agonise privately over difficult decisions, particularly during major organisational shifts or when dealing with a technology as controversial as AI. But delaying the conversation does not protect employees. In most cases, it creates uncertainty and erodes trust.
People can handle difficult information. What they find much harder to handle is silence, inconsistency or the sense that important decisions are being made behind closed doors.
My principle is to communicate what we know, what we do not yet know, what we are testing and how decisions will be made. Leaders do not need to pretend to have every answer, but they do need to be honest about the direction of travel.
It is also essential to involve the people closest to the work. They understand the practical exceptions, quality risks and customer requirements that are easy to miss from a leadership position. They are often the first to identify where a new process will fail, but they can also be the people who see its greatest potential.
The best transformations are done with employees rather than to them. That means giving people a role in designing new workflows, being direct about how responsibilities may change and investing in the skills they will need next.
My experience has taught me that human judgement does not disappear when technology improves. It moves. The responsibility of leadership is to help people understand where it is moving and how they can continue to create value within the new operating model.
As AI becomes part of everyday business operations, how do you see leadership evolving over the next five years, and what capabilities will distinguish the organisations that thrive from those that fall behind?
The most successful leaders and organisations I see are approaching AI with curiosity and discussing that exploration openly.
One of the worst things an organisation can do is investigate AI in secret and then surprise its customers or employees once decisions have already been made. That approach turns what could have been a constructive process of learning into a question of trust.
Over the next five years, AI literacy will become a core leadership capability. Leaders will not necessarily need to be machine-learning specialists, but they will need enough understanding to challenge where AI can optimize their organization. They must be able to ask where the technology is reliable, what data it depends on, how quality will be measured, where human judgement is essential and who is accountable when something goes wrong.
Leadership will also become increasingly focused on orchestrating hybrid systems of people and technology. The challenge will not simply be deciding which tasks can be automated. It will be redesigning roles, workflows and decision-making so that machines handle what they do well while people remain focused on judgement, creativity, relationships and accountability.
Access to AI will not be the lasting differentiator, because the underlying tools will become widely available. The differentiator will be an organisation’s ability to integrate those tools into its operations, culture and customer proposition.
The businesses that thrive will create safe environments for experimentation, share what they learn, establish clear boundaries and act quickly when evidence supports change. They will combine curiosity with governance, speed with judgement and technological ambition with a clear understanding of the experience they want to create.
Ultimately, the leaders who thrive will not be those with all the answers about AI, but those who have the courage to ask the right questions, bring people with them and redesign their organisations around what technology makes possible.
