DIBI - 2025 - Conference
Introduction
Last week I attended DIBI in Edinburgh on the 5th and 6th of March 2025 which was a conference for designers and developers, with the second day being the Conference itself after the first day of the Workshop, which was being held at Dynamic Earth at the foot of Arthur's Seat and just over the road from Edinburgh Marriott Hotel Holyrood where I had stayed the night and ended up being the first to arrive at the conference!
Dynamic Earth Globe
Dynamic Earth Entrance
DIBI Arrival
DIBI Stage
Conference
#DIBI2025 Welcome Address - Herb Kim
Herb welcomed everyone and had the honour to lead those there through the day and will be a fun packed and fascinating day. Herb is from Brooklyn in the United States and came over to the UK to do a semester at London Business School and came to Oxford to set up the internet bookshop for Blackwell's as part of their team and went to work to O2 for a while then came to North East to set up a company to help with growth of the creative sector there. They got a chance invite to the TED conference which was $4,000 but there was a bursary programme though Adobe and got to go to TED 2006 and they had an amazing life-changing experience but when they came back had a big idea to do a thought leadership conference in the North East which was Thinking Digital, TEDx Manchester and they helped fund the team that founded DIBI back in 2010 and is fantastic to see how it has continued and is good to come back as a host for this year.
The idea of DIBI was bringing together developers and designers to North East England and is great to see that in Edinburgh today. A great way to do this is to speak to Andy Greener, the organiser of DIBI. Andy graduated in the late 90s where there was no job for a young designer so had to create a job and business which they did and later founded another which was Komodo. Komodo is a digital product studio which designs and builds products that make sense of really complicated data. Entrepreneurial people create things that change markets, and they deal with small businesses to larger businesses and everything in-between.
The company who founded DIBI folded Andy later took hold of the conference which seemed like it was going to disappear but they didn't want this to happen so they decided to jump in and see if they could keep it going and their first year was 2022 which was the day after the lock down restrictions were lifted in Scotland and this is the third DIBI under their guidance. Herb has been doing conferences since 2008, and acknowledges they are difficult.
A lot of those difficulties are real, and you have to work hard to get past that and there's a mindset of a growth mindset, it all sounds really good but if you're not in that mindset how do you get into that. These events help with that if coming with a mindset of issues and leave with a mindset of possibilities. One of the deciding factors for Andy is that he has to come to at least one conference and is a positive experience that moves him forward in some way, there is always the urgent not important thing that distracts them.
#DIBI2025 Welcome Address
Herb Kim & Andy Greener
The Next Frontier of Audience Intel - Matt Garbutt
Matt was originally with Green Light and went to Brave Bison when they were acquired and is now Director of AI & Creative and has also been a previous speaker at DIBI and has used AI as an early adopter and is interested in using AI to help understand technology better and argue LLMs help take a step forward to understand that better.
Brave Bison capitalise on complexity from trend to spend and with engage they have a digital media network, sport and entertainment marketing, along with social and influencer marketing and Brave Bison do paid performance, organic performance and technology and experience. They work with a wide range of clients with new platforms, behaviours, audiences, trends, technologies and regulations. We've been through the internet boom but what is happening now is much faster than that.
If we could turn back time we used to know our audiences so well we had lots of first and third party data and data management platforms run by data scientists who could pull out the people you need with the granular data with extreme levels of data but then along came four evils which was GDPR, Cookies / No Cookies, Platform black boxes which is when build audiences on platforms this is done by proxy which leads to fragmentation so audiences have become silhouettes.
Market research and focus groups are costly and lengthy but we still have plenty of data, but it's just been hard to use it. It takes weapon's grade data platforms and data scientists to do this who are in short supply as they are being sucked up by AI. But now there has been a paper in 2022 which was before ChatGPT launched said about language models to simulate human samples which is they show that algorithmic bias in within one tool, the GPT-3 model is present and that models with sufficient algorithmic fidelity constitute a powerful tool to advance understanding of humans and society.
Can you build audience understanding using AI and can use this with AudienceGPT which is the only immediate way of bringing their client's audiences to life, a focus group at their fingerprints and phase one is to find our people in the data, this ideally uses anonymised first party data then this is run through their clustering model. Segmentation is an older way of doing audience demographic by labelling but clustering finds mindsets in people and clusters these together to find like-minds. Then this can surface most digitally engaged highest value clusters and then they can generate seed audiences for paid social channels to create look-a-like audiences and then can generate psychographic-enriched personas.
The second phase can use clustered first party and third-party data and then use AudienceGPT which can use combination of prompt engineering and input data and can have multiple use cases and workflows including Creative Preflight. You can use this to generate use cases such as a premium patron, banking buyer and QSR questor. You can ask them about the advertisement that can be provided using the AI which can include engagement and takeaways and avoids having conversations about whether clients like what they are seeing. Custom journey mapping workflow can allow the AI to be asked to workshop something such as taking something through inspiration to consideration to actually buying something and get an empathy map as an output from the AI and can get as much as they want. Quant and Qual is another workflow and you can ask the a structured survey and then get answers to those questions and get some analysis and output from this and can be as many as needed.
Responses from language model are 90% of what has been seen in real-life, how do you do this with a ChatGPT account you can create a custom GPT and create this in a simple process where you put in a system prompt about what you want it to do but this takes a bit of prompt engineering along with providing knowledge for the data and you give it a specific set of data which is an audience dataset you are working with and then you have you custom GPT to work on that data. They've learnt that the methodology is accurate and moves so fast it is a window to future trends and good inputs equal good outputs along with reaching nice audiences but issues can be not to over rely on it as human rigour is critical and it can miss non-linguistic cues and bias must be prompted out and there is a risk of “leading the witness”, they don't mention the brand they're working with and questions are open-ended.
How it has been useful where they wanted to identify new audiences to drive more effective performance creative and they had reliance of a single customer personal that was aging out and the response was clustering based on their data which revealed five audiences with two who high digital engagement and the conversion ads are driving the best performance they have seen.
The Next Frontier of Audience Intel
Matt Garbutt & Herb Kim
Building Scalable and Adaptive Products in a Fast-Evolving Landscape - Ayo Owolabi
Ayo is originally from Nigeria and came over to the North East and started a company, MiniPod and his partner will be studying here in the UK, Ayo started as a designer and taught himself how to build product.
Things are moving fast in the world in AI, in September 2024 OpenAI o1 model was launched which is their thinking model. Claude was launched in October 2024 which could use a browser and access the internet then Amazon launched in Nova in December 2024 which could monitor communications and Llama 3.3 was launched in December 2024 which had computational efficiency and a relatively lightweight model and Gemini 2.0 was launched in 2024 which was quite a capable AI and Deepseek in December 2024 featured even more computational efficiency and Grok 3 in February 2025 which has real-time knowledge from X / Twitter where billion dollar companies are trying to outdo each other.
Ayo is an indie hacker, and it is quite an exciting time to build things and the time to prototype to getting something is much quicker. There are patterns where it seems you can have a product or an idea and then just sprinkle some AI. AI for something or AI for some other thing with just a simple connection to ChatGPT where the business model is hoping people don't find out about ChatGPT.
It is exciting and delicate times as now it is easier than ever to build something that becomes obsolete with AI. When thinking of ways to innovate you may be looking at different technologies that work together so have the AI stack to provide opportunities for builders with semiconductors from Nvidia, AMD and intel and Cloud Infrastructure such as AWS, Google Cloud and Microsoft Azure then have foundational models from organisations such as OpenAI, Anthropic and Meta and where most of the opportunities lie is applications where you can innovate on the application layer but also below this layer you can have AI agents.
So how to you build for a fast-changing world and how to build something that remains relevant for years, you can look at companies that have survived similar technological changes to help improve your chances by applying three basic principles, the main thing is you should get a product to market as quickly as possible.
The first principle is scalability, which at its core is the product's ability to grow and you will start with a resource that your service depends upon, and you will reach those limits so you can scale, the limitation is how far can you do this without interrupting the service. Scalability in applications is long-term efficiency to make sure that costs of service don't outweigh profits of the business. Instead of doing blind API calls you should do fine-tuning of models along with RAG which can be a lot cheaper and better for a lot of use-cases and it needs to be modular so it is not tightly coupled with an AI service so you can swap out models easily and be able to plug and unplug easily and separate business & AI logic.
Second principal is adaptability but at some point, there will be some point where you can no longer scale or landscape changes fundamentally. The company has to realise the landscape has changed to keep growing and this does happen but with AI the market is changing more quickly than ever so you have to inherently build adaptability into your product when integrating AI. AI is just a set of tools and generative AI is just one part, and it was built on the back of supervised learning using the resources on the internet and you can end up with something like ChatGPT. Supervised learning is done with large amounts of data, but you can use it to improve the adaptability of your product or could be user supervised learning to build a system to understand your users.
Third principal is human-centricity, there was a man who was a railway worker who had a metal pole go through their brain and damaged part of their brain and found it difficult to make simple decisions and those with similar damage found it hard integrating emotion into decision making so they found it difficult to make simple decisions but could analyse problems and discuss rationality. Emotions drive decision making, people may say there are logical, but it is always emotional. Your product / AI feature will require users to decide to use your product or to keep using. You need to connect emotionally to your user and this helps builds loyalty it is more than just about the features this will help with long-term adoption. Are you connecting your business to your users emotionally?
Building Scalable and Adaptive Products in a Fast-Evolving Landscape
Ayo Owolabi & Herb Kim
Shaping Product Strategy as an Individual Contributor - Anglela Fleming
Angela works with FreeAgent as their VP of product which helps sort out and organise financial life, Microsoft was one of the first places for people to go into a product function which needs a whole separate discipline. Angela reflected on hearing about the confusion and frustration on the topic of strategy. They have been with FreeAgent fir five years and have had over twenty years in tech and most of that time in the product management space. The role of Individual Contributors in product strategy is everybody's role and everyone should be involved. Satya Nadella, CEO of Microsoft said that strategy should be everyone's business because the best ideas don't just come from the top – they come from everywhere.
Claim a seat at the table as an Individual Contributor, you won't get an invite you need to be proactive to share your insights and get involved in the conversation. Individual Contributors can influence product strategy is with data-driven insights such as code efficiency or performance driven metrics to identify insights you should be sharing these as management and leadership may not always see this so you should get into the mindset that people want to hear from you.
Customer understanding could be being involved in beta programmes or focus groups and with technical leadership whether you are in design or development you can bring your insight and knowledge forward to enhance your user experience as a designer or reduce tech debt to achieve this as a developer. Cross-functional collaboration is about working well across departments, but it is also outwith other teams and understand their pain points and what their priorities are and influence direction moving forwards.
Angela was given some advice and someone told them to share the things they were saying and to remember influence isn't about authority, but it is about knowledge and expertise and how you share that. Expanding your sphere of influence as an Independent Contributor could be within your team, organisation and beyond. You want to be recognised as that go-to person, you can share your knowledge, and you can start to build respect and have people come to you for advice and your recommendations and opinions then carry more weight and you can expand your sphere of influence and get out more into an organisation.
Stepping up and how Independent Contributors can be more strategic so think about features or bugs then think about how this aligns with the business strategy and if you struggle to answer that then you can ask questions of your own leadership and demonstrates you are thinking of the bigger picture. Also, you can become a subject expert and think of things you need to be mindful of and stay informed along with sharing that information. So, think beyond code, stay informed and communicate proactively and make sure people understand the direction you are going in. Mentor and lead, and how you can shape and support others, what skills and expertise you could share with others, if you have an idea or opinion get more people involved.
How leadership can empower Independent Contributors, this can be to create clarity so if you don't understand your company's vision you will struggle to align what you do with that. Leadership needs to create an environment to empower you all to enable you to contribute to conversations about product strategy. There can also be companywide and department meetings and ask more open questions for a truly two-way process. A good strategy consists of three stages, diagnose the current situation, have guiding principles and take coherent actions.
Empowered and purpose driven teams that answer five questions on why the team exists, how do they align with company strategy, how will they succeed, how do they know they are successful and what will they to. This needs to be a connected strategy and coordinated input guided by the knowledge opportunities and measures defined by the teams. Leadership and Independent Collaborators are a truly collaborative process and is important to work closer together to form a great strategy. They have a seat at the table, they can influence strategy, expand their sphere of influence and step up more and translate your work and share it across the business. Everyone can contribute and build a strategy which helps the produce achieve the goals that have been set.
Shaping Product Strategy as an Individual Contributor
Anglela Fleming & Herb Kim
From AR Filters to Product Success: Lessons in Agility, Data, and Momentum - Miriam Khenissi
Miriam is originally from Tunisia and came to Scotland for University and began their path towards digital and design and they created an AR filter that went viral with over a billion impressions. Miriam works as a product designer at Perkbox and they have worked on AR filters. AR filters started on SnapChat back in 2015 and filters are now a popular feature on most social media platforms and AR filters are now a tool to increase engagement and improve brand awareness organically and AI is now introduced in filters.
Miriam works as a product designer at Perkbox and they have worked on AR filters. AR filters started on SnapChat back in 2015 and filters are now a popular feature on most social media platforms and AR filters are now a tool to increase engagement and improve brand awareness organically and AI is now introduced in filters.
You can create AR filters with lens studio from Snapchat and Effect House from TikTok and they built one to find places to eat in Edinburgh, there is no coding it is all nodes and logic and there can be triggers for the filter such as nodding your head or starting recording. The filter they had is Guess the Country which showed clues to a country. Once of the crucial lessons they learned was agility and iteration and they realised perfection isn't a prerequisite for success they shipped the game with just 10 countries but they got feedback with more countries and clues they embraced the concept of improvement and they also lengthened the time.
MVPs in product is where you start small and iterate and can ship value to your customers and users fast, learn from user behaviour and fail fast. With things like Cursor, it has never been easier to experiment with ideas. Data driven decision making was monitoring how people were using and reacting to the filter and they realised outrageous clues meant higher engagement and feeling good when getting a clue right and get a country correctly they get a little dopamine hit so needed to balance hard far-fetched clues with easier ones and last insight was sharing outside of Instagram which is where they failed as people were recording themselves and posting it to TikTok so they should have recreated it there but didn't.
User research at PerkBox involves interviews with admins and users, group discussions with advantages, multiple surveys, design testing over Maze or Zoom and customer events observation and by using custom GPTs they can enquire about anything they need. Maintaining momentum should have meant putting the filter on TikTok and someone else did it first and capitalised on their own idea, the lesson is not to sleep on something too much as in social media things move on really quickly and would be quite stale by today's standards as things have moved on so much. Momentum breeds momentum, collaborative environment across multiple teams, continuous user excitement and feedback implementation and internal cross-team collaboration to maintain momentum. If you have an idea that you think is silly, just put it out there!
From AR Filters to Product Success: Lessons in Agility, Data, and Momentum
Miriam Khenissi & Herb Kim
Can you build Algorithmic Justice? - Sidrah Hassan
Sidrah is an AI ethicist working for AND Digital and works with the BBC and is a large company working with large brands. Sidrah is talking about the ethics of AI and how this applies to technologists. What is AI and the foundational understanding of what it is, what algorithms are and how are they committing injustice. Her background is in AI governance and product and is an AI ethics content creator for BBC Scotland.
Sidrah is talking about the ethics of AI and how this applies to technologists. What is AI and the foundational understanding of what it is, what algorithms are and how are they committing injustice. Her background is in AI governance and product and is an AI ethics content creator for BBC Scotland.
The Experiment - filling in the blanks of a nursery rhyme row row row your… boat, humpty dumpty sat on a wall, humpty dumpty… had a great fall. An AI will take in content and understand this to fill in the blanks to get the answer you have asked for. AI refers to machines that can independently process information and execute tasks based on that information with sub fields including machine learning, natural language processing, computer vision and robotics.
Algorithms are essentially a set of instructions on how to utilise data in a meaningful way allowing data to be extracted as needed and deliver on the outcome by putting things together. They sound pretty objective so where's the injustice - it is with bias which is a natural thing for humans and it keeps us safe but there is unfortunately negative bias but this is an indicator of our flawed society and we strive for better but the AI doesn't know this and they pick up the patterns and think this is the optimum way of the world and is harder to correct and turn off compared to humans.
78% of North America and 74% of Europe is using the Internet but only 20% of sub-Saharan Africa so there is a Northern skew to the data. For example, asking for a performance review with same prompt except for name of John and Jane resulted in different outcomes for performance rating based on biased data.
An AI hiring tool was ranking female candidates lower than male candidates and it was fed in CVs from successful candidates and the majority of these were male and taught the algorithm that men were preferred over women, they tried to anonymise the CVs but the models would pick up on other factors but they got rid of it in the end as could not guarantee it would be neutral. Someone was detained for stealing a luxury watch due to being misidentified by an AI facial recognition tool that is more likely to misidentify black people than white people, the police fed in a grainy CCTV image and got a match but didn't do anything else so wrongfully incarcerated him.
How can you build Algorithmic Justice? Build inclusive, ethical and transparent systems that acknowledge different people, have ethical boundaries that don't amplify inequalities and produce verifiable content and the process of producing the content is clear. AI governance can help where ethical AI systems and products are the love child of governance and innovation as we're building products for people - make sure you are building tech responsibly.
Thought Diversity is about hiring, retaining and empowering diverse teams allows you to build AI products for a diverse society and you need to let them thrive with psychological safety to allow your team to voice their opinions and most importantly their concerns. Ethical literacy means providing training around building AI systems with a focus on safety and responsibility which is the most important and underpinning point.
Turn AI systems into weapons of moral construction rather than weapons of mass destruction to build for a more economical world and a better environment starts with questioning and putting the user to the front of our minds, does the product help or hinder the user. Move fast and break things impacts marginalised communities the most.
Can you build Algorithmic Justice?
Sidrah Hassan & Herb Kim
Customer Feedback is Still King - Kiran Jawanda
Kiran is a senior product designer at Monzo, which was founded in 2015 with over eleven million customers. There is the challenge of turning insights into action and Monzo take this seriously by taking customer insights into actual action.
As individual contributors we start chasing the next big thing but understanding how customers interact with our products is fundamental. Kiran has worked at Monzo for a year but has been in fintech for over ten years. Research and forming a vision is understanding customer needs and validate them, understand pain points and opportunities and understanding how satisfied customers ore with their current way of doing things. Usability testing is product comprehension and ease of use allowing you to proceed with confidence to find issues both big and small to proceed to the next stage, which is usually launch, the MVP can often be very lightweight without all features or spend so much time looking at it you don't understand what it should be like.
In life feedback is using produce analytics to understand usage and with the value of feedback. They turned up to an airport and were excited about going on holiday, but they were in a queue that wasn't moving for economy and lots more for business that weren't being used, their metrics would show he checked in late but wouldn't show their actual experience about feeling small and insignificant. They had a positive experience where a dish had five pieces and were about to add one to make it six which made it feel special so feedback can come in different forms it can be positive or negative and big or small along with being immediate or long after the initial experience.
In fintech it is people's financial lives, and this stuff matters, and you need to listen to that feedback and it is important as we minimise on these interactions and miss out on unique experiences and views. Using it can help understand the real-world impact of your product and it can inform what you prioritise what to work on. Monzo use this feedback is they launch as a pilot and using produce analytics to understand usage and give customers space for feedback at the right time and running further rounds of research. They embed feedback throughout the application where they have a form at the bottom or for more verbatim feedback with an embedded form. They have feedback embedded in slack and are able to see customer feedback embedded in their workflow and they have a comprehensive design system which allows engineers to react quickly when customers have issues or reprioritise any work.
When Monzo launched credit insights they knew holistic credit help was needed so they launched with two but wanted to add a third one in and people wanted that third credit score with Experian and people were willing to pay so they needed to build that third score in. Customers cared about their score changes over time and why it is going up and down, but their feedback channel wasn't stating this but what they were saying but they noticed something on their credit score they didn't recognise so could embed this journey in Monzo as that is what the customers wanted to know. Credit score isn't impacted by credit insights as it is soft search but customers were thinking it impacted their score just to check it so they added a small message when asking to stop tracking a credit score to say that checking the score doesn't impact the score but were also able to iterate on this messaging quickly.
Validating that are doing the right things, they will have a meeting with a product manager, and they can get feedback during the meeting to see if they are doing the right thing and will get positive feedback and knowing you're making a difference helps you appreciate the work that you do. They could still improve by using AI to summarise customer feedback better or for keywords but also want to make sure they don't lose sight of verbatim customer feedback. Listening to this feedback will help your customers see a product change over time which could be a morning or a week and is a way to keep customer engagement high. You can do this by varying the way you speak to customers, access the options on how to get this feedback, embed the process where you work, fight for the space to do this even if it is really small and use your design system and 90% of your app should be using a design system so they know how things should behave and can leave 10% for special experiences to light up features where needed.
Customer Feedback is Still King
Kiran Jawanda & Herb Kim
When Software Understands the User - Kevin Stewart
Kevin is head of design at Waracle and are based in Scotland and they have hit 200+ in terms of team size and their CEO was named CEO of the year for digital transformation. Kevin is a design leader at Waracle and their job is to find ways to delight customers of clients to get a competitive edge and the best way is to elevate their customer experience. Wartime is one of the periods where humanity is the most innovative and comes up with the most inventions and the second most common times for innovation and another are innovations during recession, which is when you think aren't going well but innovation can come when things are really tough so when you have a recession there is a necessity for a new way of doing things.
Customer experience leaders outperform the market against those who don't, over the past few year's things have got more expensive to manufacture and when this happens customer experience spend goes down and therefore the quality of it goes down. During recession and tough times, the barrier for entry gets lower. During Wall Street crash there was an issue with making movies, but Disney went into animation to save money. Amazon during the dot com bubble not having inventory helped them to keep costs low. There is also the gig economy which reshaped many things was a result from a recession and this reshaped the entire way the economy works such as Uber, Airbnb, Task-rabbit and Fiverr. This changed how we did business completely and made businesses understand that a digital product could be viable and coined the term of digital transformation. How people invested and how people invested completely changed and you could make it free and charge no money and you could get money, and you could get another company to buy that who would give you more money.
During recession we had Web 2.0 was when we had stable internet where you could not only pull information you could push information and do things on the go with smart phones, and this allowed people to do peer-to-peer things to transform the economy with digital products. Today we have near zero latency connectivity allowing things like remote surgery where can process any amount of data anywhere in the world. AI has been coming out and doing quite a lot of things. You've got to start with the customer experience and work backwards to the technology. Modern AI is very similar to humans, they are prediction machines to predict the next thing in sequence. Back in 2022 ChatGPT launched, the P means pre-trained, AI has been around a while but you had to train it but ChatGPT has a lot of context to start with and is capable of understanding most contexts. The interface was plain language rather than a programming language which means anyone can interact with it. All businesses are language businesses, all regulations are language regulations, all customer interactions are language interactions and the time to 100 million monthly users was 2 months. You have a pre-trained AI anyone can interact with and can embed anywhere in a business so need to do a similar thing now with intelligence transformation.
Intelligence transformation we need to think about how to embed intelligence into the business, you would have been asked questions from a person and that would help them to make a decision and could react in real time but that person can't do that all the time and the more people you have the more it costs and can turn to technology to help scale things. We can use design-time intelligence which is all the things that are needed such as questions and outcomes would be designed in. Real-time intelligence is expensive to scale but design-time intelligence ignores edge cases.
With AI businesses have a new type of intelligence to provide run-time digital intelligence with apps like Cleo which will help give advice on bad spending habits and the intelligence is in the running of the software rather than what was designed up front. This can be done in the music industry to help discover new music with pre-digital you have record shop, with Spotify you can get recommendations based on other plays from other people and Spotify DJ learns what you listen to and will play what you want with a DJ voice over. You can also have with vitamins you would get them from a shop or online but with intelligence and could take in inputs and see potential for optimisation. Could also have an AI see pension data that you could chat to but there was a bit of misbehaviour, but it would work quite well and when someone gets a pay rise it could notify them how much better off they would be putting some of that into a pension. If you had infinite human intelligence, what would you do differently in your business.
When Software Understands the User
Kevin Stewart & Herb Kim
User Research and GenAI: Better together - Rachel Walker
Rachel is from Silicon Valley and has worked for Apple, Google and Amazon and is principal UX researcher for invanti. User research can play a critical role in creating AI for good. Rachel only gets on a train when she knows where it is going except when the doors are about to close and gets on it if it is vaguely going in the right direction. Companies are looking at ways to put AI into their business. Why does it matter? It can be easy to dismiss AI as a gimmick but it can do good such as Cancer detection which is one of the most promising AI use cases that is showing promising results but only works well with those with lighter skin tones and it is possible to adjust the model to make them less biased but this is a conscious decision that needs to be made.
UXR can help GenAI, the role of the user researcher is ensuring that the AI is doing good and this plays into their core role, which is to advocate for user needs, challenge assumptions, insist on designing products well and needs are being met. In an ideal world you identify a user need and then do a tech advancement but in reality, it is a tech advancement then design an idea and then identify a user need, but those products don't do very well.
Important aspect of a researcher's role is to challenge assumptions but there is a lot of AI negativity particularly for chatbots which can save money, but a lot of end users aren't happy about them. Inclusivity and representation are important, not every user group is going to be represented fairly and could introduce a worse experience for certain user groups so need to make sure that data is inclusive and representative of all users. Researchers can make sure that user testing includes a diverse sample and then can make sure that performance and feedback is being monitored post-launch so if there are inclusivity issues they can be addressed.
AI powered features can have implications for privacy and regulations such as using things for a purpose they should not be used for. Legal department knows the most about GDPR but second is customer experience researchers, and they can have a role with ever evolving AI regulations. Accessibility and usability we can conduct research with people of varying needs and describe interactions that are accessible and usable for happy users, but reality is designing interaction without user testing then trying to make these accessible and usable means unhappy users. Researchers can identify and reduce negative impact such as putting warning about potentially misinformation to mitigate potential term.
How GenAI can help UXR by brainstorming augmentation by incorporating AI into your brainstorming sessions to generate more ideas or kick off a literature review by asking your chatbot about your topic. There is also data visualisation and upskilling where generative AI can create a chart or for help using unfamiliar quant platforms which reduces the barrier to entry. AI can also make interview transcription more reliable making analysis much easier. Summarisation at scale where can instantly summarise large-scale text data such as free-response survey answers or interview transcripts but then any insights can be checked. Persona image generation can be used to generate imagery form personas or presentations such as a photograph of a family using their smart display in their home to communicate. Content generation and refinement is a lot easier where can generate emails, surveys or adjust the tone of any content.
Cutting edge use-cases of AI include language translation which enables you to be able to interview users who are excluded due to language barriers. Better participatory design which allows participants to give feedback and bring their ideas to life. Pattern identification where AI can identify patterns and preferences over time and enables predative analytics to model how users will behave in the future. Simulated eye tracking studies which can predict how users will process an image for simulated A/B testing using previous data. AI-let interviews and surveys mean you can use voice-enabled chatbots to moderate interviews or conduct surveys including follow-up questions especially difficult subjects. You can also generate synthetic user data to supplement data for under-represented data or anonymise underlying data.
When Software Understands the User
Rachel Walker
The Strategic Mindset of a Designer Who Thinks Beyond Audience - Hardy Sidhu
Hardy worked with Akqa and is now founder and CEO at Format-3. He is not the usual person he would admit and will explain there is an invisible wall between designers and developers and don't breakdown to lead business, but they realise we need to stop talking about the same thing such as AI.
Will be approaching from a philosophical point of view and thinking beyond aesthetic approach. Without saying you're a designer-who are you? They are a questioner. Why limit ourselves based on our title of designer. They loved to solve challenges they love problems. Designers facing invisible wall about growth of product.
They said an actor shouldn't involve himself in politics, Deep Sidhu didn't involve himself he stood in and shook it up. As a designer can you not shape the future of business. What is design without business in the modern day?
Business allows designers to have reach and more and allows business to engage and interact with their customers to build moments of brand experience. There is the book A seat of the table, let's get a piece of the table not a seat, we're problem solvers as a designer.
How much does CEO care about empathy of design etc so designers need to widen their toolkit to think about financials and business model to get a piece of the table. Do you know how the thing you design is aiding revenue, so business could burn through money so if prove a return on investment and drive value how can they use it.
Have to understand cost per acquisition and is often the return on investment and if have objectives that need to be hit and the impact to reach that customers or prospects actually want this. Don't start with the process start with the questions. Three “W”s are What, Why and Wait for the idea, why we are listening and hold on a minute.
Approach for sports engagement platform who had funding and got in contact and had four months of money left and asked them to give them all of their capital runway they had left. Their problem was building out features and didn't have a single customer so stay focused and find out they wanted chat rooms and advertising rather than other things. By doing this they had big success after nearly pulling the plug after six years by really questioning what they are doing and why they are doing it.
Expand vocabulary to have the right conversation for desired effects and reframe how we talk about opportunities on why need a design system, why test their solutions and why need more designers and reframe from point of designer to a business leader. Use that reframed vocabulary to embed into three areas of discussions.
The Strategic Mindset of a Designer Who Thinks Beyond Audience
Hardy Sidhu & Herb Kim
Designing Gamification for Good - Adrian Hon
Adrian is a game designer and former CEO and co-founder if SIX to Start. He managed to get a spot talking on stage at TED and co-creator of Zombies Run which wasn't just about making money but delivering a healthy.
Gamification is about using mechanics from games with other things or things people may not want to do. Zombies Run was launched from a Kickstarter and been around for six years with over ten million players and launched Marvel Move and worked with others to do gamification well.
If you have Peleton then you see a leader board and that is a powerful motivator with competitive mechanics. You can get achievements for how long and much you have exercised by linking game mechanics. Duolingo has missions and double XP which works for them and gets them engaged with the app.
You can use gamification for anything such as Rock Smith Plus which makes learning the guitar much better although you can't make it easier. You can see your progress and gives you more feedback about your progress. Lots of other apps have gamification built in and many will have a target to beat your precious target but that isn't a good idea long term for exercise.
Uber drivers get incentivised with quests to get more rides, and this can disguise people's compensation. Amazon fulfilment centres have based on activities it will progress you through games you can choose from this such as the more you work the more virtual pets you collect, and people get attached to them and they stay on longer.
Principles of effective gamification should be fun, the vast majority doesn't do this, but you can achieve this by making it more interesting and more different. Approachability is important it needs to be easy to get into. Longevity where there is an idea of novelty effect and if you change something it doesn't matter what it is people will pay attention and engage more but if there's not enough depth or is enjoyable it isn't worth doing. Try to design something that works solo or in groups, often can be based on competition and sometimes that is motivating for some people but if there's not a large user base or demotivated by competition it doesn't work. Scaffolding new behaviours or training wheels for users where you don't want people completely dependent on your app to do the thing but want to make something that is fun anyway even if they don't need it. Is your thing still fun if you remove all points, badges, leaderships or achievements.
Risks of gamification include overuse or spending too much time using the app which is good for a business you need to remember why you are doing what you are doing. Catering to core players too much especially if have forums and this can happen a lot with new game designers, and you end up changing it too much and people don't care. If multiplayer, then can have a toxic community but can design cooperative games this feeds into community. If multiplayer cheating can be an issue but what would be the point of that and if people are playing for love of it. Focusing only on gamification features is also a risk.
Be specific, so people take badges and achievements and attach this to your existing website, but you can do this for anything but for activities it needs to be very specific. They weren't the first to gamify running. Nike split London up into segments and there were teams or there was another one that was location based, and you could collect squares and capture territory, but it just took someone to do a bit more than you to beat you. People run the same route usually who run regularly and don't want to look at their phone too much.
Zombies Run was designed from perspective of meeting people where they are, people like running their own routes and not where they don't want to run or pull out their phone every few minutes. People are listening to music or podcasts already, so it is not too much to ask to require audio. Put people in an audio post apocalypse and you run and collect supplies and sometimes you get followed by a zombie and you have to speed up so don't tell them where to turn as that is unsafe. They have had people for years who have subscribed and got fit or lost weight.
Beat Saber wasn't designed for fitness you get fit as a result of playing it and you are cutting blocks coming at you and sometimes there are walls you have to avoid and playing, they are getting fit and you're not asking them to do more than they would otherwise. When thinking about gamification try and make something they think highly of.
Designing Gamification for Good
Adrian Hon
Conclusion
Overall, it was a fantastic day at the conference which included catching up with a few familiar faces but also meeting and speaking to so many people I didn't know which was great! The conference was followed by a Party at Le Monde Hotel & Bar which was brought forward so was able to enjoy more of it and meet up with more people there too before heading to the station for the train from Edinburgh back to Newcastle and then home to Gateshead.
DIBI Closing
Herb Kim & Andy Greener
Leaving Dynamic Earth
Edinburgh Waverley Station