Leveraging Generative AI to Drive User-Centricity

About this guide

This guide is based on Elsewhen’s in-depth experience in helping organisations to become more user-centric. Combining market insights with real-world examples and industry research, these insights are relevant for people working on the ground in digital teams, right through to senior leadership. By reading this guide, we hope you will strengthen your ability to champion user-centricity in your organisation.

Interested in getting a PDF copy of our report?

Download PDF

In this guide, we will

Identify how GenAI can be leveraged to drive user-centricity
Recommend actions that can be taken immediately to improve user-centricity using GenAI and LLMs, including Personalisation, AI-augmented user feedback mechanisms and Employee experience
Offer strategic guidance for building AI-driven user-centric organisational culture and leadership

User-centricity: What is it and why is it important?

User centricity has taken a number of different guises over time, influencing ideas around customer-centricity, human-centricity and, even, life-centricity. But these ideas all draw upon the same principle: placing user perspectives, requirements and needs at the centre of an organisation’s culture, strategies and operations. Whilst the foundations of user-centricity lie in design, with Don Norman’s user-centred design principle, the philosophy now extends across and influences functions across organisations.

Yet even today, there’s still a gap between acknowledging user requirements and truly understanding user needs. Within organisations, user-centricity is often blocked by top-down decision-making, departmental silos and an emphasis on output-focused projects. In fact, new research shows that only 15% of organisations believe they have the data or organisational processes to meet the needs of their users and customers.

And now, artificial intelligence (AI), such as GenAI and Large Language Models (LLMs), is changing the user-centricity game. GenAI and LLMs will supercharge the user research toolkit, equipping organisations with an unprecedented understanding of the user. Elsewhere, GenAI, with its ability to deliver more personalised and seamless experiences, is set to widen the gap between CX leaders and CX laggards over time. And for employees, from sales to engineering or insurance to construction, next-gen AI will deliver more user-centric employee experiences; bringing with it massive productivity gains and greater job satisfaction. At Elsewhen, we believe that those brave enough to be AI early adopters, wanting to supercharge their services, will benefit the most.

Building your AI-augmented user feedback mechanism

To place users at the centre of an organisation, first it is key to develop efficient and effective user feedback mechanisms and loops. At the core of user-centricity best practice, there are three specific processes that need to be in place:

Capturing user feedback
Creating clear personas
Closing the feedback loop

Whilst your organisation may already have some experience in these areas, GenAI, once again, is revolutionising the tools and capabilities available to build a better understanding of user needs. In particular, Natural Language Processing (NLP) in combination with fine-tuned LLMs, can help organisations optimise and connect these activities to close the loop and turn insight into action.

1. Capturing user feedback

User feedback, whether quantitative or qualitative, is the only way that organisations can truly understand the requirements and needs of users. As such, organisations need an accessible and multi-method system in place for gathering and acting on user feedback. This means going beyond simply asking users what they want and digging deeper to understand the underlying problems each user is experiencing.

The keys to capturing user feedback are:

Establishing a streamlined process to talk to users. Which feedback collection methods, channels and tools? Do the necessary organisational processes exist?
Defining a cadence to capture feedback regularly. Too frequent and users don’t see impact, too infrequent and we lose insight.
Creating a repository to store these insights. Use a database, spreadsheet, Notion page or Miro board to store and analyse feedback.

GenAI and LLMs are already being utilised by organisations to supercharge user feedback processes in a number of ways. When compared to traditional chatbots, LLMs are better equipped to collect user complaints, feedback and recommendations that can then be plugged into the product and service development process. Through transformer architecture and vast training datasets, these models provide multilingual support, personalise interactions and boost customer service agent efficiency; all whilst continually improving and learning with each interaction. Furthermore, GenAI can analyse user feedback obtained from interviews, surveys and reviews, providing valuable insights into user needs, pain points and preferences. GenAI, such as Condens, Dovetail or ChatGPT Enterprise, can be used to analyse user insights and identify themes.

2. Creating clear user personas

User personas, built up by collecting user data and direct user feedback, enable you to address user feedback directly from your target audience. Without clear personas, the development of products and services may not meet the needs or expectations of their target audience, resulting in user frustration and loss of business. Personas are not just for the product or marketing team, it is important that everyone in the organisation understands what would resonate with each of the target personas.

To define user personas within your organisation, start by:

Identifying your target audience. Who are you trying to reach with your products or services? What are their needs and wants?
Gathering data about your target audience. This data can come from a variety of sources, such as surveys, interviews and focus groups.
Creating a profile for each persona. This profile should include information about their demographics, lifestyle choices, interests, behaviour, a name and a face.

The use of GenAI to create user personas varies by organisation. For some, analysis of user or customer data using LLMs could be used to identify clear target personas within the user base. Alternatively, LLMs can be used to create synthetic personas, algorithmically generated users, that help test digital products and digital marketing initiatives. This more innovative approach utilises the biases and preferences that are baked into LLMs to generate personas and carry out user testing using ‘AI participants’. This approach not only eliminates the need for lengthy, expensive user testing but also leans upon the billions of parameters used to train LLMs as opposed to the creative limits of 2 or 3 product managers or marketers in a room.

3. Closing the feedback loop

To be truly user-centric, organisations must close the feedback loop; acknowledging feedback, engaging with users and actioning user feedback findings. At a time when loyalty is at an all-time low, leveraging user feedback effectively and efficiently will be a competitive advantage.

To start effectively closing your user feedback loop today, think about:

Developing clear lines of communication. Develop clear processes, channels or tools for closing the feedback loop.
Being open and honest with users. Clearly outline what you are trying to achieve, how you have used the feedback and when it might have an impact.
Actioning user feedback. Deliver product and service development that meets the demands of user feedback.

Arguably, turning feedback into action is where GenAI and LLMs is already having the biggest effect on organisations. Tools, such as ChatGPT, Bard, Figma, Midjourney and Canva, boast a plethora of capabilities that augments the ideation process using both artificial intelligence as well as user feedback. Furthermore, popular product management tools such as Notion, Figma and Miro, are increasingly leveraging GenAI that allow users to seamlessly uncover user needs and validate potential product ideas. Many organisations are already leveraging these tools to turn user feedback into product and service development, from copywriting, product features and UX/UI designs.

Whilst your organisation may already have some experience in these areas, GenAI, once again, is revolutionising the tools and capabilities available to build a better understanding of user needs.

Generative AI and mass hyper personalisation

GenAI and LLMs represent an enormous opportunity for organisations to shift the emphasis of personalisation from passive to proactive. LLMs’ ability to synthesise natural language, understand intent and decode unstructured data allows organisations to plan, launch and scale customer and user experiences according to customer data, preferences and historical behaviour in real-time. Here, LLM-driven mass personalisation will help deliver a new level of customer satisfaction, engagement and, ultimately, user-centricity.

Significantly, there are a whole host of examples of leveraging LLM-driven personalisation for user-centricity. We recommend exploring a number of use cases, including:

Data-driven experimentation: Organisations have understood the value of customer and user data for decades. Despite this, user data remains siloed and unstructured; making analysis expensive and inefficient. Now, enterprise-level Large Language Models, such as GPT-4 or PaLM 2, can analyse unstructured data from across an organisation, unlocking insights on user behaviours and preference. Using unprecedented, data-driven knowledge of users, teams can deliver hyper personalised experiences for individual customers at scale.
Conversational experiences: LLMs, with an ability to understand complex queries, leverage a superior knowledge base and generate and justify recommendations, massively outperform inefficient and inaccurate AI chatbots of the past. Furthermore, many of these models are now multi-modal, accepting written, spoken or visual prompts. These developments have laid the foundations for more personalised, conversational user experiences such as Klarna’s smooth shopping ChatGPT plugin or Walmart’s Voice Order. In this case, user needs and preferences are placed at the centre of the experience, removing both physical and digital barriers between the user and the product or service.
Personalised user engagement strategies: Using LLM-powered data analysis, brands can develop hyper-personalised reward or loyalty schemes. Brands, such as Tesco and Starbucks, are already deploying GenAI to shape customer rewards and promotions based on their behaviours. This benefit extends to how retailers communicate with customers, using GenAI/LLMs to create copy assets code in highly tailored creative campaigns.

How is GenAI reshaping the employee experience?

New research suggests annual labour productivity growth, attributed to GenAI alone, would average between 0.1 and 0.6% until 2040. And, by combining these tools with other workplace technologies, this figure could look more like 3% growth. In total, by automating between 60 and 70% of employees’ tasks, GenAI could add the equivalent of $2.6 trillion to $4.4 trillion annually to the global economy.

Critics here might suggest that, given the era of low employment rates and skill shortages, retention and recruitment will be key in delivering sustainable productivity gains and associated growth from GenAI. In fact, a recent study in the UK found that 67% of workers already using it are experiencing higher levels of job satisfaction, and overall better employee experience.

By freeing up employees from manual, time-consuming tasks, GenAI and LLMs have the potential to not only drive unprecedented levels of productivity but also create a more user-centric employee experience. Employees will no longer have to deal with tedious, monotonous tasks; instead focusing their time on value-add and rewarding activities.

There are a number of use-cases to drive user-centric employee experience through GenAI:

Streamlining business processes: LLMs, when connected to proprietary, internal data, have the ability to automate business processes at a scale, speed and accuracy never seen before. Walmart store associates have an AI assistant, Munich Re have introduced AI-augmented underwriting whilst LLMs could be used to analyse and iterate business processes themselves, through large process models (LPMs).
Automate tedious tasks: Copilots, by combining LLMs with workspace datasets e.g. documents, emails, chats and meeting, are increasingly rewriting the capabilities of productivity tools. For example, Microsoft 365 Copilot fosters and encourages innovation by surfacing insights and trends as well as automating recurring and often frustrating tasks e.g. product email updates or scheduling meetings.
Augmenting creativity and experimentation: AI can generate initial drafts of written content and designs, which can serve as a starting point for further refinement and customisation by human designers and product managers. For example, ChatGPT and Bard draft decks and write or reword copy. Additionally, Canva, Midjourney and Figma offer LLM tools to automate the formatting of design components, ensuring consistency and saving designers’ time.

Strategic advice

Establishing efficient and repeatable processes to deliver user-centricity from GenAI and LLMs is just one part of the complex puzzle. A 15-year study found that healthy organisational culture, a function of ‘leadership’, ‘accountability’, ‘user focus’ amongst other indicators, has a correlation with long-term success. To keep up with the speed of competitors, organisations need to develop a user-centric organisational culture.

For some organisations, the path to a user-centric culture begins at the top. A Harvard Business School study found that CEOs, on average, spend just 6% of their time with ‘rank-and-file’ employees and only 3% with customers.

When those leaders driving the narrative and processes around user-centricity have never interacted with users themselves, it is perhaps no surprise that these initiatives fail.

Ahead of setting Airbnb’s 2023 strategy, CEO Brian Chesky adopted an ethos of “I can’t make products just for 41-year-old tech founders” and developed an initiative to understand user needs and pain points. Along with 50 major product upgrades, Chesky himself spent 6 months living in Airbnb properties to immerse himself in the variations between different hosts. This experiment shaped the company’s 2023 strategy, focusing more on renting an individual room rather than an entire house.

Leading by example is important, but it’s also critical that the whole company operates under a shared user-centric vision; to understand and meet user needs. For example, the medical device business, Medtronic, required all the company’s engineers and designers to attend and observe at least one surgical procedure a year – to capture user feedback from the surgeons. Over the last 10 years, the organisation has outgrown competitors to become the largest medical device manufacturer in the world.

Looking at AI in more detail, achieving user-centricity with GenAI and LLMs brings with it a number of unique requirements. These technologies and initiatives require an agile mindset and methodology, allowing for organisations to constantly fail, iterate and test different approaches. Strong and engaged leadership is even more important with the implementation of GenAI: empower employees, encourage experimentation, demystify the hype and ensure safety. These tools also require training and new skillsets. Organisations must make AI implementation accessible and fair for all employees, redeploying and retraining wherever required. Finally, organisations must not underestimate the security and ethical concerns of AI; adopting responsible AI principles to prevent negative consequences for employees, the organisation and society as a whole.

Conclusion

GenAI and LLMs have massive, untapped potential for user centricity. From driving a more efficient user feedback loop right down to augmenting the creativity process itself, AI tools are the key to bridging the gap between acknowledging user requirements and truly understanding user needs. But to achieve this, organisations need to rethink their ways of working, toolkits and skill sets.

At Elsewhen, we believe that those brave enough to be AI early adopters, wanting to supercharge their services, will benefit the most. To realise the user centricity advantage of AI, we recommend taking a number of steps:

Be bold and creative with GenAI and LLMs experiments that can drive user-centricity across your organisation now. ChatGPT, Bard or other tools could help uncover user insight from publicly available data, at speed and no extra cost. For instance, at Elsewhen, when delivering a time-sensitive research project, we recently scrapped app store reviews, identified the top 10 user sentiments using ChatGPT and baked this into our recommendations.

Engage with teams from across your organisation to start exploring and brainstorming potential enterprise-level use cases for GenAI. This might include engaging with the marketing and customer experience teams to understand where LLMs might fit into ongoing personalisation efforts. Alternatively, your data team might want to start centralising, cleaning, and preparing unstructured data for use in LLMs.

Define the longer-term strategy for GenAI adoption around user centricity, asking questions such as:
Do we have clearly defined policies and governance on using GenAI and user feedback?
Do we have the right skills, structures and processes to unlock AI potential at our organisation?
Are leaders engaged around leveraging AI to achieve a user centricity advantage?

We hope the above insights and examples resonated with you, and that you can take some of the advice given throughout this document to become more user-centric in the age of AI. If you have any questions or challenges to discuss, please reach out to Elsewhen at www.elsewhen.com/contact.