Industry Insights

Robots come out of the AI winter to give brands a hug


Image of Robby the Robot from the film ‘Forbedden Planet’ from Wikipedia

When I write about robots, what image does that conjure for you? For many I am sure it will be one of the many science-fiction “mechanical men” such as Robby the Robot from the 196 film “Forbidden Planet” or for those of my generation his (her? it?) later incarnation in the family television series lost in space. The cultural impact of these robots has been significant (“danger Will Robinson!”), but the reality is taken a long time to catch up as the level of artificial intelligence (AI) needed to make these robots a reality has been beyond us.  In fact, AI has suffered from a cycle of hype and disillusionment for 50+ years (the so-called “winter” of the title).


Robots are more than mechanical men

If you look at the definition for ‘robot’ in Wikipedia you will find something quite interesting. Wikipedia recognises that robots can also be disembodied and virtual creatures: think about chat-bots on e-commerce sites. Perhaps the “HAL9000 in 2001” is closer to the kind of robots (or bots) that we are more likely to encounter in the near future.

Think about this from a brand point of view. You want to achieve reach but also engagement. By 2050 there will be over 9 billion people to reach and engage with. This is clearly the domain of automation, but it needs to be a form of automation that can communicate intent of brand managers to consumers.

Take native advertising: wouldn’t it be great if you had a robot that could correctly promote a tweet on twitter or suggest a post on Facebook?

The challenge is that brands have personality, and they have an emotional side and all of this is core to their equity. When I use the term robot I am thinking of software that is capable of emotionally intelligent, personal and interactive support regardless of whether it is embodied in a cute physical persona like this one at the reception of Hilton hotels, or working next to people on a factory floor like these from Rethink Robotics.  The new generation of “collaborative robots” are designed to work safely with people, and some (such as Sawyer) have features designed to build rapport with teammates and to help with visual cues about movement (it has a face, for example!)


The Artificial Intelligence (AI) Spring

My view of robotics for fast-moving consumer goods (FMCG) really starts with what has been learnt from good old desktop automation. You know, those macros in spreadsheets such as Excel that continue to carry some of the load of repetitive tasks. A more sophisticated example would be what is referred to as Robotic Process Automation (RPA) as used in finance and HR functions, often to help with self-service.Let’s hear from Mark Fisher, IBM Partner in Finance, Risk and Fraud (Mark is ex P&G and so knows the FMCG world well):

“The advent of RPA has created a real buzz across Finance and the wider back office over the past two years, with many leading organisations seeing this as an opportunity to not only reduce costs (robots are faster, cheaper, can work 24 hours a day and don’t take holidays…) but also to build in better controls (robots make far fewer manual errors).

This has allowed CFOs more time and focus to support the business in commercial decision making and develop more insightful, analytical robotic tools that can help to optimise sales and predict growth rates”

Despite the demonstrated benefits of these approaches they have been limited by needing to work with structured data and simple rules.  These approaches work well for well-defined tasks, but they don’t address the need to work with messy, unstructured data and, most importantly, it is obvious you’re dealing with the machine.

Even the smartest macro is pretty dumb and if you are a Marketer working with brands and consumers these techniques allow you to stay just the right side of spam in a direct marketing campaign, but they are not going to move the 1:1 consumer relationship needle very far on the dial.

However, the huge breakthroughs in deep learning technology over the last few years (IBM Watson for example) have enabled companies such as IBM to start building a completely new generation of robots and service innovation is everywhere. These robots bringing an awareness of context (for example they would understand that somebody asking about a Mars bar is not talking about a recreational facility on the red planet).

Their language ability is vastly superior to what is being available before, to the point where it can be very difficult to detect whether you are interacting with the human being not. Perhaps most importantly these new “cognitive” robots are capable of a degree of autonomous decision-making that does not require hard embedded rules and all of the maintenance that goes with that-they are learning systems after all.

In my view we are still some way away from true artificial intelligence (think of ‘Her‘ from 2013 to continue the movie references), but we can already achieve a lot. In a recent eMarketer study (“Artificial Intelligence 2016: What’s Now, What’s New and What’s Next“) looking at the use of AI techniques in large companies, voice recognition and response (dialogue) techniques were being used by over 30%. 15% had virtual personal assistants.

You’ll note on the chart below that robots appear as a separate line, and I believe this still reflects the older view some hold that robots are electro-mechanical machines slaving away on the shop floor:


eMarketer also report:

 “Large tech companies, including IBM, Google and Amazon, dominate the market and are jockeying for leadership positions. There are also hundreds of smaller startups focused on developing niche products that meet specific business needs.”

As we have seen in other tech’ areas, start-ups tend to be highly effective looking for an unmet need or a specific business problem and targeting their innovations there. Here are a few examples from sales, marketing and supply chain:

Caspera                                   marketing chatbot for night owls

Jibo                                          emotionally intelligent robots

Teabot                                     custom tea blend at point of sale

Momentum Machines         customized gourmet food production

Marble                                     automating last-mile logistics

Impact of robotics on FMCG e-commerce

One area that I believe robotics will have a very important impact on in FMCG is e-commerce. To quote again from eMarketer:

 “e-commerce is not just a sales channel it’s a marketing channel, where content will be the defining factor in the battle for screens and attention.”

 The opportunity for big brands is not to be the best at the shopping transaction (important though that is) but to make that transaction relevant and meaningful and, yes, emotionally fulfilling. For me this means robotics rather than a buy now button.


Robots and the future of manufacturing

I haven’t abandoned the idea of robots as physical presences whether they are in a retail store or in a factory. Recently I was a guest of the Financial Times in London at their excellent “Future of Manufacturing” Summit. The themes of the day were “big data analytics, advanced robotics, the internet of things and additive manufacturing.”  Yes, robots!

It strikes me that many developed nations are re-igniting their passion for making things (see the MADEHERENOW Project as an example). The economic choices made when so much mass-manufacturing was off-shored over the last 30 years no longer seem valid in a world where robots walk the earth (or sit next to you on a production line) and where mass-customization starts to take over (for an up to the minute discussion, take a look at Peter Marsh’s book “THE NEW INDUSTRIAL REVOLUTION : CONSUMERS, GLOBALISATION AND THE END OF MASS PRODUCTION“).

If the latest 3D printers can make over 100,000 per day small items per day, and robots become ubiquitous, then labour arbitrage is no longer the dominant competitive factor, especially as more and more personalization of label, pack and product is demanded by consumers.

When I first went out to work there were robots on the shop floor in the car plant already but they were difficult to program and they had to be kept in cages away from people. They were completely unaware of my need to be treated like a human being-they were just machines.  Likewise, their design and software did not lend itself to emotional engagement either (see here for a creative approach to making robots likeable).

Fast forward to today and there are robots using cognitive computing techniques out there working side-by-side with human workers to make and pack a wide variety of everyday items. This new generation of robots do not require guards, and they are not programmed in the conventional sense-they are taught.  For example, when I recently visited DHL’s innovation centre in Bonn, Germany, I was shown many examples of how robots were augmenting human beings in their distribution centres.


Should we fear our machine overlords?

Should we be concerned about the rise of this new generation of robots? The word ‘robot’ itself has its origin in a dystopian play (“R.U.R”) by the Czech writer Karel Čapek.

For academics such as Nick Bostrom this question is linked to broader concerns about the rise of artificial intelligence and the likely impact on employment. But hasn’t the human race always had to handle new technology? If this was 1794 and we were looking at the introduction of the cotton gin wouldn’t we have similar concerns? It does seem to me that we are creative enough to find new ways of employing ourselves or even new models for operating our societies as the writer Jaron Lanier described in his book “Who Owns the Future?


Last words on robots

robotsThis is all getting very philosophical! So let me come back to reality. We’re still a long way from being able to hold a 100% convincing conversation one-on-one with anyone on the planet. See how well it works for you here.  But the rate of improvement is fast.

The practical extent of robotics in most FMCG companies today is bounded by use of enterprise RPC tools such as Blue Prism at the one extreme and exciting proof of high-concept and boutique projects at the other.

In my view cognitive computing will rapidly create a new generation of software robots or bots…perhaps we should we call them “cogs”…which will allow us to engage with  individual consumers at the level of the global population.

The direction of travel for robotics gets clearer every day whether you are in marketing or in the supply chain. Whichever way you look at it the field is vibrant and investors are investing (see this research from CB insights here and here).  Are you?




IBM, Distinguished Engineer

More Industry Insights stories

Build your own wonder

Wonder-ing out of the uncanny valley The “uncanny valley” is a phrase used to describe machines or systems that at the same time appear close to human, but elect a sense of revulsion rather than wonder.  Creepy is another word that gets used.  It is an old idea (first used by a Japanese roboticist working […]

Continue reading

More wonders

In my last post (“A time of wonders“) I finished on the idea that “machines like HAL or Her are referred to as super-intelligent or as strong artificial intelligence…and that most experts agree that we are probably 60 years away from being able to create systems like that, but there are many wonders available to […]

Continue reading

A time of wonders

Wonders indeed.  Last year a man’s life was saved in Missouri when his Tesla car drove him to hospital on autopilot. It seems that every day we read headlines like that one: you could call it a time of wonders, and all of them seem to be the result of artificial intelligence or AI. It […]

Continue reading