Marshall McLuhan said “the medium is the message”, meaning the medium changes how the message is perceived. I tell a story about how we came to prototype a conversational UI, and how this new medium challenged the team's thinking. This is less about the 'how' of constructing a Conversational UI and more about the 'why'. What thinking we needed to challenge and why this approach helped achieve that. In a broader sense it reflects the evolution of the industry in the past 5 or so years.
3. 3|Top: A mangled quite from Marshall McLuhan
But, sometimes we need to stop replicating old
methods with new tools
Have you ever had the feeling a project was
just finding more efficient ways to ‘fight the last
war’?
This is the story of an experiment to change
the medium and hopefully, affect our thinking
6. 6|
It’s fast
Icons from Creative Stall @ The Noun Project
For a more direct fiber path between
Digging through mountains and rock
$300 Million
New York Chicago
A company spent
3 millisecondsTo save…
9. 9|
It has all the
screens…
Trading floors can be
overwhelming.
Though not all work is
done in these
environments
Image source: Getty (Royalty free),
11. 11|
Sparking a
platform war
Source: Simon Wardley
• Something upsets the status
quo
• Parts of the value chain are
commoditised
• A period of rapid change
• A fight for survival
• Often played by the old rules
• New, higher order activities
emerge
• There is lots of experimentation
and failure. The future is
uncertain
• New winners begin to emerge
at new points of value
• Mature markets
• A time of relative stability
• The rules and winners are
established
• Success breeds inertia
Peace War Wonder
We were here
Industries and technologies go through a continual cycle.
13. 13|
Our project brief What we were tasked to do in 2012
Create a platform that was flexible and
adaptable to meet client needs and let them
shape how they use services
01
The platform should allow the bank to be more
responsive to changing demands02
Create common tools, components and
more to drive efficiency and allow iterative
evolution
03
Provide a rich UI using the latest technology
04
Allow 3rd parties to contribute to the ‘open’
platform.05
14. 14|
We had settled
into a structure
and rhythm
Source: Idea from Robert X. Cringely and Simon Wardley
Town
Planners
Settlers
Product teams
• Establish Apps
• Analyse requirements
• Optimise experience for
a given app and
features
1 central platform team
4 product teams
+ other distributed teams throughout the bank
Platform team
• Industrialise platform
• Establish common tools
and patterns
• Consistency, efficiency,
utility, access
Roles
15. 15|
Not everything
was awesome
Some issues to overcome
The Apps were still defined by the internal
organisational structure01
Needed consideration of how the platform
would evolve02
Needed to consider how the industry and
clients might evolve03
Were we just making it easier to build and
distribute the same hard to adapt
monoliths?
Were we just gluing Legos together?
04
17. 17|
The role of
pioneer
Pioneers operate in unchartered waters and
are comfortable with ambiguity01
Challenge assumptions and reveal local
optimisations02
Prototype novel ideas for competitive
advantage03
Seek to make future success possible
04
Settlers
Town
Planners
Pioneers
Draws fromDraws from
What do they do?
20. 20|
It started with a
conversation in
Spanish
Learning a language gives you a new
appreciation for structure, nouns, verbs…
Could I map the structural grammar of our
entire platform? Down to each interaction,
button and event?
What if we tried to do everything in a
conversational UI?
We would need to break everything down to
it’s lightest, most temporal, common parts…
21. 21|
I mapped all the
Nouns and Verbs
Defining the grammar for
the entire ecosystem
resulted in a clearer view
of what we were dealing
with.
Not so many special cases
after all.
All tasks could be recalled in
linguistic structures
23. 23|
The prototype was
pretty rudimentary
Syntax hintsCall an app Inline buttons, charts and content
Triggering actions
I can’t show you the real thing because of strict NDA.
However it had many similarities to Slack (not yet launched).
Though, being a prototype, it wasn’t very polished.
The features (helpfully demonstrated by Slack)
24. 24|
Initial lessons Changing the medium brought some welcome challenges
You can’t have a button for everything
01
Not everything will be in view. You need to
think about time and space more02
Your existing silos and concept of an app will
be challenged03
You need to have a holistic understanding of
your system04
Not everything is suitable. You need to
decide when to offer a jumping off point05
26. 26|
Discoverability
and the uncanny
valley
Triggers, signals, flow and handoffs
Optimum number of capabilities are one or
everything (Ben Evans).01
Need to set expectations about what it can
do. You realise how much GUIs and Nav do
this job.
02
More suited to a domain that constrains what
you want to do. Or where you tell it what to do
and move on.
03
Discoverability can flow from context…
04
Traders are accustomed to commands
27. 27|
Anticipatory
design
It is a powerful constraint that focuses you on
context and jobs-to-be-done05
Emphasises that paths can/should adapt
with the user, context, and expanding
capabilities
06
We needed to learn new skills to design for
conditional states and flow07
It challenged the assumptions of how we were
bundling and integrating apps…09
Forcing a new framework
28. 28|
Design systems
not destinations
Moving from linear connections and singular
structures to flowing configurations10
Apps aren’t screens. They don’t need to be
a single thing we open, with users initiating
every action.
11
The primary way to interact with your ’App’,
may not be through an ‘App’12
It made us think about our job differently
29. 29|
No more
cathedrals,
just bazaars The constraint of Conversational UI demands
modularity because modularity aids
adaptability
13
Doing something new shouldn’t require an
extensive redesign14
The flow of conversation should just get more
intelligent, be more anticipatory, have less
steps…
15
Expect others to scale out in unanticipated
ways without repeating effort16
Designing to scale iteratively
30. 30|
“It is the framework which changes with
each new technology and not just the
picture within the frame”
Marshall McLuhan
Closing thought