Automation AI and Robotics command daily headlines given the pace of innovation. It therefore makes for good conferences as this one indeed proved to be. I was chair and this is a flavour of what was explored on the day.
The inimitable Brian Manusama from Gartner started proceedings with a disturbing quote. “The country that leads artificial intelligence will rule the world.”. This is Putin’s view of its strategic value. I think he is right which incidentally is why our own national success matters so much.
Brian had wise words for enterprise decision makers.
- Be patient. AI needs the right expertise and workflow discipline.
- Focus on augmenting employee value such as improving the quality of decision making.
- Figure out the business problem before choosing the technology. In other words, have purpose before shopping.
He finished with one final pearl – The business value of AI will be proportional to how thoroughly you reinvent your business.
A great scene setting contribution.
In terms of putting those principles into practice there were some very clear examples to learn from. Mike Migliore, Head of Customer Value at News UK was one.
It’s no secret that selling news is tough and newspapers need deep reinvention to offer more value than online news aggregators. Mike showed how subscriber loyalty is a function of engagement and habit. This is driven by highly personalised subscriptions. The question becomes how can this be affordably scaled?
Enter James – an anthropomorphised machine learning driven engine. James orchestrates the right content, right time, right format and right frequency for each individual subscriber and of course being an ML algorithm, gets smarter over time. The net result is increased engagement which translates in increased lifetime value.
The results showed James was especially effective in the first few habit-forming weeks of a new subscriber. Dramatically reducing their rate of churn.
Fast forward to the next example. Verco is a leading energy management and sustainability business advising organisations how to transition to a zero-carbon economy. Tim Kay. Commercial director presented how they were starting to use AI for in the context of that mission.
As you might imagine energy management generates tons of data. However, turning that into realised benefit is tough. Especially when searching for those opportunities is human driven and therefore tends to be inconsistent in terms of skills and focus across different organisations. As Brian’s Gartner research showed, improved decision making is the top use case for AI. It’s ability to unearth patterns within complex datasets is precisely what Verco aims to achieve for its clients.
The project is still in early days. Tim and his team are using 2019 to pilot the capability of spotting patterns and opportunities. Anyone interested in being part of that beta testing programme should contact Tim.
After lunch we kicked off with another market overview. Leslie Willcock’s views the world of automation through the lens of LSE as a curator and author of how organisations are evolving their use of automation. In line with growing market maturity, we are moving from ‘easier’ – structured data (robotic process automation) to ‘harder’ unstructured data and basic decision making (cognitive automation).
My own observation is that RPA has been popular. A quick show of hands from the audience confirmed that its use is widespread. The benefits are readily quantifiable in terms of classic cost reduction metrics: less time and faster responsiveness. There is a lot of low hanging fruit in the form of simple administrative tasks that can be automated. This makes RPA faster to rollout out than assembling a team of data scientists and engineers which more complex AI initiatives classically demand.
However, once we enter the more complex world of cognitive automation, the ROI becomes tougher as the quality of data becomes a key issue.
We were then offered a wonderful example of why AI is going to be a force for creative disruption in every sector. As one of the audience remarked at the end of Stuart Stock’s presentation, ‘I had no idea rubbish collection was so interesting!”
Stuart Stock is CIO of Veolia. He loves data and uses it to great effect. He has assembled a very bright team of 40 data scientists to make rubbish collection smart. Within customer service he has implemented automation, chatbots, predictive analytics. He seems to have handled journey mapping and omni-channel which others still struggle with. The fact that customer churn has halved in the last two years is evidence of the team’s impact.
However, it was adding sensors to bins that really sparked the audience’s interest. This allows a whole new approach to the scheduling of bin collection. Instead of pre-determined collections, new schedules are dynamically generated based on bin capacity SLAs. This fundamentally changes the business model of waste collection echoing the point Brian made at the start of the day.
We saw a glimpse of what smart cities are all about. I can’t wait to see what Stuart’s team get up to once they get their hands on 5G technology.
Amy Mitchell from Leathwaite took the stage to talk to us about recruiting AI talent. She has tons of experience and well worth tapping into if that’s where you are in your own journey. She offered what I would describe as strategic common sense. Don’t invest in roles until you understand the business problem (another Brian Manusama point). Set people up to succeed in terms of workflow and support or they will under perform and eventually leave.
I’ve just read a great article on the important differences in skill sets between data scientists and data engineers. Both are needed in typical AI deployments and they are poor at duplicating each other’s skill set. The point is to make sure that real expertise is brought to the discussion and decision making around the required team and the support needed to make it a successful investment. As such, Amy concluded with the following observation – Adopting AI is primarily a people problem.
ThoughtRiver provide contract pre-screening. It’s a classic use case for augmenting humans. Trawling through legal contracts for the few clauses that require attention is slow, expensive and prone to error when undertaken by people. ThoughtRiver takes the pain away and expedites contract approval at lower cost. No brainer really.
‘Digital first’, ‘cost reduction’ and ‘winning through data’. These are any retailer’s mantras in today’s world. Phil Jones used them to scene set what M&S is doing around intelligent automation. Phil approach was detailed, educational, well planned and by the looks of it well executed. An agile approach providing the momentum. Like most large organisations, M&S employees waste time doing tasks that automation would do better.
Here are Phil’s hard won insights.
- Get the right ownership.
- Find the right problem to fix – it took a few iterations for M&S to hit gold.
- Improve the process before automating it.
- Focus on the data – quality, governance etc
- Think long term – scalability
That’s it. The day was full and insightful. Delegates walked away with a pocketful of practical insights to guide their own efforts. Things are definitely happening. Let’s hope UK plc keeps forging ahead in their AI deployments.