Designing Organisations That Work With The Grain Of Human Cognition

Last week, I gave a quick talk about the relationship between organisational design and human cognition at the Mind in Practice event held at Google Campus in London. The idea was to consider how we can design organisational structures that work with the grain of human cognition rather than against it. Here are the ten points I made at the event.

We have seen lots of competing metaphors for organisational design in recent years that contrast machine-like structures with organic metaphors from nature, but I personally believe we need organisations to become more machine-like and automated and also more human and creative at the same time, and this will require new thinking about collective intelligence and distributed cognition.

 

We have evolved with great spatial and structural awareness and pattern spotting abilities. We can even locate memories and information in spatial constructs. But our organisational structures are poor, low-resolution environments compared to the Forests and Savannahs we trained on.

Our organisations are not making the most of our abilities. In fact, they are making us stupid. Our org structures were designed to command and control reluctant factory workers in an age when we had very primitive coordination and communication tools, not to enhance cognitive work.

 

This is primarily a structural problem, and only secondarily a cultural problem, so we need to tackle the system that shapes behaviour and culture, rather than just ask people to change, which implies they are the problem.

We know now that good and bad behaviours are contagious, thanks to work like Nikolas Christakis’ and James Fowler’s book Connected, based on an analysis of on the Framingham Heart study data, and Mark Earls’ (et al) book “I’ll have What She’s Having” – more so if bad behaviours come from the top.

Therefore, we should try to design evolutionary structures that allow positive, human connections to spread and grow like neural pathways, rather than continue to use structures based on dividing people.

 

By way of example, one of the worst artefacts of the traditional organisational model is the way that so much communication takes place in private emails and phone calls that move up and down the hierarchy. This seems to incentivise negative, political behaviour, and perpetuates information asymmetries and knowledge hoarding.

Comparing internal corporate email between people in positions of differential power with how people speak in groups is very interesting. Moving communication and work coordination to more open channels, such as Slack for example, incentivises better behaviour (people don’t want to look like jerks in front of their peers), but it also creates better ambient awareness among teams.

 

One reason ambient information environments matter is the fact that we can process much more information than we realise. Information overload is a problem of delivery and filters, not volume. Sitting people in front of an email inbox, sequentially processing instructions like Pavlov’s dogs (ding ding!) uses just a tiny proportion of our pattern matching and sense-making skills.

Collective intelligence needs much greater ambient information, and not just in written form. I think the combination of IoT devices, data flows and lightweight personal AI inside organisations could throw up some very interesting new ways for us to maintain contextual awareness and receive alerts and updates.

 

With a decent level of agency and autonomy, people can achieve almost anything. If we strip them of agency and meaning and treat them like cogs in a machine, they become less productive and less happy.

Giving people greater autonomy and allowing them to group together at different levels of scale to make things happen is good for everyone. This applies to companies, institutions, healthcare and even how we assist refugees and displaced persons.

Create a context in which people can get things done to help themselves.

This is why much new thinking about organisations starts with idea of the small autonomous team or squad and scales up from there.

 

We no longer think of the brain as a kind of CPU that controls our actions and thoughts. We are starting to understand that many functions of intelligence are distributed, and even if we might not be quite as decentralised as an octopus, we at least understand that distributed intelligence is a product of evolution and emergence, not top-down control.

Right now the dominant model of hierarchical management assumes intelligence and leadership exist only at the top. This is incompatible with what we know about how humans work, both individually and in groups.

 

The possibilities for distributed cognition with the right combination of machine and human intelligence are truly exciting, but we need to apply AI and cognitive computing in the right way.

If we can use them to automate the boring stuff and augment rather than replace human intelligence, then organisations in the age of algorithms could be both more machine-like and more human at the same time.

I believe we will see technology platforms replace much of the management meat that dominates organisations today, and on top of these platforms, talented individuals (including a new generation of leaders) and groups will find new and creative uses for the underlying services in pursuit of value creation. Of all the low-value repetitive jobs that automation can replace, perhaps middle management (as in the coordination of work) will be the biggest area of cost saving, whilst freeing up organisations to work more autonomously. Management should be a business function, not a reward for loyalty or a political cartel, and at the point where this expensive luxury starts to hinder organisational performance, Boards need to step up and think about the long-term health of the organisation.

Connected organisations will allow people to apply their cognitive powers over a much wider area, and with greater impact.

 

The future is not a battle between humans and AI any more than it was between humans and bronze, or books or weaving machines. I like Kevin Kelly’s idea of the Technium – that technological evolution is an extension of human consciousness and evolution, not something separate. We shape our tools and thereafter our tools shape us, as McLuhan famously quoted John Culkin as saying.

Meat and machines are changing together, but at different speeds. Our culture is fast moving, our cognition less so, and our bodies even more slowly, whilst software is evolving in real-time. But the more we learn about our own consciousness, such as this apparent recent discovery that our nervous system is even higher resolution than we previously thought, the less we should feel threatened about it being replaced by a form of engineering ‘intelligent design’.

 

In conclusion, I think that organisations in the age of algorithms will need to make the most of our distributed cognition in concert with AI. In that sense, they need to become more brain-like, more ‘hive mind’ and less directive.

Tech platforms will replace much of the management meat that dominates organisations today, and on top of these platforms, talented individuals and groups will find new and creative ways to create value.

By automating the boring stuff, by removing the politics and posturing from how we coordinate work, and by letting people work autonomously in networks and groups, we can create the conditions for organisations to make the most of the immense power of the human mind, without gradually killing us through cognitive dissonance, alienation, frustration and boredom as some of them do today.

The first Boards with the courage to reform byzantine management structures, and free up the organisation to work in a more agile and autonomous way, will reap huge rewards in terms of cost savings alone. But more importantly, they will leave a legacy for the future that is more important than just enjoying their time at the top of the pyramid.