Designing for Agent Experience

By solle on 04/04/2013 — 10 mins read

The future isn’t in hiding from machines; the future is in trying to be recognised by them. @blaine

For hundreds of years we the designers have been hunched over our benches designing objects and services for other people. Sometimes we have thought carefully about these other people, but often we have just worked away with blithe disregard. In recent years many of us who have chosen to care have been referred to as User Experience Designers and/or Interaction Designers and our approach to work variations on User Centred Design. As designers, we have long poured ourselves into this work; heads down, confident of our methods, processes and engagements. Always ready to explain, listen and share our work and our best way of doing things and help guide and improve where we can and generally, as humans designing for humans, strive to make the world a better place. We are confident that all this listening, all this understanding, all this hard work has paid off. We can see it all around us. Acutely. But, are we paying enough attention to how this designed world of ours is changing and changing faster? As humans we have become symbiotic with technology. We have transformed ourselves and at the same time we have transformed the world. We always want to believe we are getting a handle on how things work, who and what our work is for, and the digital and physical landscapes that our work exists in, but what happens when we realise that all these accepted truths have begun slowly, almost imperceptibly, to shift and warp and all our commonly held norms, truths and rules have changed and everything is pointing to a far more complicated present and future.

The connected internet is filling up with many new voices all singing noisy, variable, messy and different songs. We look up from our benches and realise we aren’t just designing for people any longer; algorithms, bots, machines, other products, optimising software all are busy interacting and interfacing with our products and services. Products are people too. Matt Webb of Berg has previously spoken of how customers and users are possibly also the ‘agents’ of other products and services, devices and networks connecting to a product or service. The network is the new electricity and connecting products is the electrification of the 21st century. He asks how do we live alongside products, how do products act in the world and whether we might use these ideas to better inform our design decisions. Design can be easier when we acknowledge that products share our homes and malls and have wants and lives of their own. Also, if we are already seeing our ‘users’ (our humans) gradually becoming less than 100% human (increasingly complex pacemakers, robotic limbs) the potential, technically, in the not so distant future is that we won’t ever be designing just for userhumans. Add to this also when robotics continue to mimic and mirror human behaviours and actions, are we then not approaching nearside futures where human behaviour grows in its mimicry and mirroring of robotic, device, computational actions and behaviours? The financial markets are already ‘socio-technical‘ systems. The omnipresence of Siri and its voice recognition cousins is building machine language and machine understanding through continued mass use. Systems combining both human and robot participants. Understanding and designing these systems helps to manage the behaviour within them, even when it is undesirable. As Kevin Kelly has written, if technology wants something then we can design for it. And as Kevin Slavin has said of trading software, we are now writing code that we cannot read, with implications we can’t control. Kelly (in the same book) has also written of how “(our) planetary electronic membrane is already comparable to the complexity of a human brain” with three billion artificial eyes (phones and webcams) plugged in. And “when computer scientists dissect the massive rivers of traffic flowing through it, they cannot account for the source of all the bits”.

It’s not that these multifarious connected things (at present anyway) have actual ‘experiences’ as we understand it but that their very existence alters our experiences and how we design experiences. For example, as a designer understanding how social applications – such as Twitter – have been adapted beyond their intended use. Successful software doesn’t stand still. When it was launched in 2006, even the early adopters stumbled over their initial ‘Is this thing on’ and ‘I’m having breakfast’ tweets before everyone worked out how it might work. The majority definitely would not have guessed at some of the current uses. Timelines are now full of bots and bot shadows and we are building relationships with them. As a reaction, maybe; as part of the progress of software and technology, yes definitely. Their existence, their language and their patterns interest us and alter our experiences and expectations. This might seem small and inconsequential (and some marketeers may think they are weakening Twitter’s legitimacy) but I think my timeline is only interesting because of the bots. What comes next with these services? We can only guess but guess we should.

So. Should we be designing for these algorithms and bots? And how are we to approach it? Well, we think we might be able to see what they are seeing and there is certainly a new aesthetic, but will they or are they already capable of autonomous experiences? Well, these are strange and incredible times and I don’t think anyone really knows – even the most digitally literate amongst us. But, if we use the metaphor of ‘experience’ to describe the behaviour of software we probably can. But is ‘experience’ something that will always be uniquely human? Or is it something that we can create in machines? Where does that take us? As Ben Bashford has pointed out, we don’t necessarily have cultural reference points aside from the skeuomorphic legacy of say the internet fridge or Terminator. It is our responsibility to create these reference points. There is no escaping that designing future ‘user’ experiences will need to consider and at times include designing for ‘agent’ experiences (there will come a time when we will really need to think beyond the actors in our networks). Especially when it is expected that 700,000 to 1 million new jobs will be created by robots in the next five years we really will need to sit up and take notice. But we do not know what those experiences will be, at present they can not tell us but they are using the same services and products and influencing anything from price to perception – and sometimes with very very unpleasant consequences. Designers at Amazon surely cannot ignore these agents. They already account for such a vast amount of its transactions and traffic. It is on the world stock markets where this is being played out most acutely where sometimes they seem to control themselves. The table vividly shows the incredible increase in high speed algorithmic trading of which the dynamics and characteristics are poorly understood. The quite possible reality is that future trading bots will be able to adapt and learn with little human involvement in their design and far fewer human traders will be required in the financial markets of the future. Already, high frequency computer-based trading (HFT) has grown in recent years to represent about 30% of equity trading in the UK and possibly over 60% in the US. We need to imagine how these agents might perceive products and services if we are going to be able to accommodate them, design for them and integrate them into our planned experiences.

How do we design and incorporate all this into our workflow? What do we have to change? Currently, I don’t think you have to change very much. It’s more of a cultural shift than having to completely change any way of working. Today, it is just important for designers and developers to be aware, but at the same time they need to stop thinking that they are only ever going to be designing for other people. We need to stir up this settled culture of only people. The activities and behaviour of these agents, nodes and products in the network are going to pose designers new challenges. How they connect, where they connect, when they connect and when they disconnect. And this connecting/connected and disconnecting/disconnected are the ‘beautiful seams’ of the network. This is where the designer and developer should focus most of their concentration. This is where the opportunities for design and experience will develop. The reality is these beautiful seams and this is where designers are needed. This is where great experiences are needed. The digital utopias of seamless digital experiences will always only ever be utopias. The design and development communities also need to be learning how to talk about agent-things. Discussion of terminologies and basic understanding need to be in the open and everyone should learn how to talk about them clearly and unequivocally. We don’t want to get saddled with any unnecessary confusion. We need to be introducing agents into our design thinking, our planning and the culture of our organisations and the production of our products. Ignoring them as irrelevant and believing that technology will always remain technology as we understand it now will place our work at a serious disadvantage.

So, where to get started.

There are no definitive steps or lists or ways to know but certainly designers and developers should start scrutinising logs and quizzing the analytics of the products and services they are working on. You need to start asking new questions. Can you recognise agent-things in your data? How do you recognise different types of agent-things? How are they different from people and does their existence and activity influence people’s experience? Is this a problem? How does it impact your design and development thinking and approach? For example: personas, user testing, interviews, feedback, support. How much is it altered? From this can you further understand or get information on the umwelt and goals of the agent-thing? What tools can you employ? The right answers are hard to come by – especially when most of us aren’t really sure what we are looking for – but by actually looking you are in a better position than if you didn’t. Certainly designers and developers working on seriously big data, information, activity and applications at companies such as Google, Amazon, Facebook and world stock markets are already interfacing with the aforementioned agents and so should be in a good position to investigate further and consider the relevance.

In conclusion, I think it is worth reminding ourselves of Lev Manovich’s theory of the database vs the narrative where each are competing for the same territory of human culture, each claiming the exclusive right to make meaning out of the world. Have databases become so intelligent that they now jump over the fence and begin writing narratives? Even though we may struggle to always be able to ‘read’ these multiplying things as the connected internet contracts and expands, rather than ignore their presence and influence we will be better off (and the products and services we design) if we optimise our experiences for both.

This kind of discussion has only just started but the responsibilities are already there. Chris Heathcote has a well documented quote that, as of yet, I don’t think has been bettered:

“The world is going to get magical and strange, and people will be confused and fearful. Designers have to do what they do best, helping people to understand the world and the way they live in it – and make the tools that people can use to shape their own lives.”

More designers need to be thinking about it, mentioning it and discussing it. I think we easily forget that technology is pre-human. It dates back to the primates. At present, I think this extra effort thinking and considering is still only at the edges of most people’s scope and as Matt Jones has said of this concept of the robot-readable world “it is pre-Cambrian at the moment, but perhaps in a blink of an eye it will be all around us”.


//somewhere in england


I am a user centred design leader working across senior design leadership teams and user centred design operations. I work across a wide range of digital services within the public, private and not for profit sectors.