If You Want to Use AI to Drive Productivity, You Need to Start with Trust
By Marvin Heffernan & Cathy Fennell
The 2025 Landscape of AI in Customer and Employee Experience
When ChatGPT burst into the public consciousness in late 2022, it stirred a mixture of wonder and unease. Its ability to generate intelligent, contextually relevant responses on demand had many questioning whether artificial intelligence could realistically replace, or at least significantly augment, human labour. Fast forward to 2025, and the conversation has matured. AI is no longer a novelty; it is now embedded across core functions of how organisations work, serve customers, and engage employees.
According to Stanford’s 2025 AI Index Report, large language models are doubling in capability every six to ten months. Some experts believe artificial general intelligence could be within reach in as little as two to three years. AI is already transforming how customer service is delivered, how staff are managed, and how strategic decisions are made. However, if trust is not prioritised alongside performance, the gains of today may quickly give way to the crises of tomorrow.
Where AI is Delivering Value
Across customer and employee experience, the promise of AI is compelling. In customer service, companies like Parloa are deploying conversational AI agents that now manage millions of customer interactions. These virtual agents are reducing wait times, lowering operating costs, and increasing personalisation, especially in high-volume industries like telecoms and utilities. According to McKinsey, generative AI has helped improve marketing and sales productivity by 15 to 30%, with some organisations now aiming to automate up to 80% of customer-facing tasks.
In retail, AI co-pilots support customers through product recommendations, tailored discounts, and seamless return processes. In banking, tools like Bank of America’s Erica are enabling customers to manage transactions, set savings goals, and receive fraud alerts without needing to speak to a human agent.
On the employee experience side, AI is rapidly becoming a silent co-worker. Tools like Microsoft Copilot and Salesforce Einstein are helping knowledge workers draft emails, summarise meetings, and automate tedious tasks. Sales platforms such as Gong.io offer real-time coaching by analysing customer conversations and providing feedback on tone, timing, and delivery. In HR, platforms like Eightfold.ai and Workday use AI to predict attrition, optimise recruitment, and personalise engagement strategies.
Yet for all the promise, there are growing signs that the deployment of AI is beginning to fray the edges of trust.
Where AI Has Gone Wrong
AI’s failures are not just theoretical. They are increasingly visible, and increasingly public.
In the customer experience space, over-automation has led to mounting frustration. A 2024 report from CCW Digital found that only 7% of customers felt AI had improved their service experience. More than half reported that it had made things worse. 48% attempted to switch providers due to unresolved issues, and 40% left negative reviews online. This dissatisfaction stems from an over-reliance on chatbots, particularly when they are unable to escalate issues or demonstrate empathy.
Equally concerning is the erosion of human touch. When AI replaces too many frontline interactions, customers begin to feel unseen. The same CCW study reported that customers were less likely to remain loyal to brands they felt had automated away any genuine sense of care.
In the employee experience domain, the picture is equally complex. A 2025 Fortune report highlighted that workers who regularly use AI tools reported 45% higher rates of burnout compared to non-users. The rapid proliferation of digital assistants, unclear guidelines for use, and increased cognitive load are making some roles feel more exhausting, not less.
Academic research adds further nuance. A 2025 study published in the International Journal of Contemporary Hospitality Management found that AI failures in hotels, such as robotic concierges malfunctioning or automated check-ins breaking down, led to inconsistent service and growing employee resentment. Where employees felt excluded from decision-making, some disengaged or procrastinated. Others overcompensated, leading to stress and burnout.
And then there are more serious incidents. In one high-profile case, an insurance firm was found to have used an AI system that denied policies to certain applicants based on flawed data interpretations. Another case involved the use of AI to analyse sickness and attendance records within a workforce. Inaccurate conclusions led to HR decisions that negatively affected employees, eroding trust and morale.
These are not just technical glitches. They are human failures, resulting from poor governance.
The EU AI Act: A Shift in the Rules
Despite growing concerns around a lack of a regulatory framework for AI, the EU AI Act signifies a positive step in the right direction. The EU AI Act is the first comprehensive regulation of its kind and shifts accountability away from vendors and onto the organisations that deploy AI. Crucially, it requires AI governance to happen at the use-case level rather than at the tool or software level. This means that companies must list and assess where, on what and how AI is being applied, and manage it based on the risk each use case poses.
AI used in hiring, employee monitoring, or credit scoring is now classed as high-risk and must pass formal conformity assessments. Companies are required to inform users when they are interacting with AI, and critical decisions must always involve a human reviewer. Transparency, explainability, audit trails, and human oversight are no longer optional. They are now legal obligations.
This regulatory shift matters. It reinforces what many customers and employees have already come to expect: that accountability is necessary to drive trust.
Building Trust Through Responsible AI
If AI is to become a lasting competitive advantage rather than a reputational risk, companies need to embed responsibility into the design and deployment of every AI solution. That starts with governance.
Legal, HR, ethics, IT, and customer experience teams must be involved in assessing impact, reviewing outputs, and identifying risks before deploying any AI solutions.
As AI seeps into the background of multiple new tools available, enthusiastic staff or departments can unwittingly expose company confidential or personal data as they try out new tools and ideas. This also needs guidance, controls and a keen eye.
Transparency and explainability are also key. Customers should be told when they are dealing with AI and given a clear route to human support when needed. Tools that offer explainable AI (XAI) functionality will play a central role in bridging the trust gap.
Next comes ethical design. AI systems must be tested rigorously, not just for technical performance, but for fairness, inclusivity, and bias. Unusual data or situational cases should not be afterthoughts – they should be part of every deployment roadmap.
And finally, organisations must close the loop. Employees should be involved in feedback processes, empowered to shape the tools they work with, and supported to collaborate with AI rather than feel replaced by it. Establishing cross-functional governance councils, regular audits, and clear escalation processes will be essential for long-term success.
Final Thought: AI Should Not Just Help You Do More. It Should Help You Do Better.
The story of AI in 2025 is one of duality. On one side, it offers extraordinary potential to enhance efficiency, personalisation, and performance. On the other, it raises difficult questions about fairness, transparency, and control. The companies that win will not be those that automate the most, but those that understand where automation ends and human connection begins.
As customer and employee expectations rise and shift, the pressure is on to make your customer or employee experience simpler, faster, better and more efficient. The use of AI (in all its forms) can and will help with that mission.
Just remember – Trust is so powerful that building it can secure your revenues, whilst breaking it can instantly wipe out all the good work you may have done.
So ask yourself the following – are you jumping on the AI bandwagon just to grab efficiencies? Or are you making sure that your use of AI specifically improves your ability to build and maintain trust?