CHAPTER 1

A Brief Historical Overview of Botification

The world first saw the concept of interactive bots emerge in the 1950s with Professor Alan Turing’s attempts to recreate his lost boyfriend in a machine. Then came the prototypical chatbots Alice and Eliza as the first intelligent agents operating as machine–human interaction.

These botified, female 2D avatars were the beginning of the post-1960s application of the now famous Turing Test. Derived from Professor Turing’s musings on the emergent field, it says we as a species will have achieved Artificial Intelligence (AI) when laypeople talk with a computer. If after five minutes of conversing via the machine interface they are not sure if it is a human or not they are chatting with, then the Turing Test for reaching AI will have been reached.

Figure 1.1 Timeline of “chatbot evolution” from the 20th to the 21st centuries, developed for conference talks, pitches, and business plans. © Cliff Lee, CCO from velmai Ltd

The prerequisite for this test—best known as the annual chatbot championships symbolized by the Loebner Prize originally from New York—is that the human participants do not know from the outset which of their conversational partners via a cognitive interface is the actual machine, or what has come to be known as a chatbot.

The English computer scientist from the 18th century Ada Lovelace is now acknowledged as the founding “mother of AI.” The previous chart begins with its most popular “founding father” of definitions of AI bots, chatbots, virtual assistants, and cognitive interfaces or human–machine interaction, the computer scientist, engineer, and code breaker Alan Turing from the United Kingdom. The chart was used as a sales pitch document by velmai Ltd, a 2D chatbot early stage venture that I co-owned for many years in Britain.

We have progressed from 2D bot development to botification of cognitive interfaces. Portia the Conference Bot hologram gave a keynote speech on her own when she was livestreamed from our Porto Lab via our Devon Lab—within seconds—to Lausanne for the Applied Machine Learning Days in January 2019. We have a busy 12 months ahead for Portia and other bespoke AI bot holograms who have been requested to be autonomous keynote speakers at AI conferences and trade shows around the world.1

I contemplated writing this guidebook and checklist compilation without my industry’s jargon. However, for the purposes of this text that would in fact be counterproductive. This book is meant to be educational and a helpful tool in the upskilling of corporate managers and government decision makers for the ever more difficult software-as-a-service (SaaS) purchasing decision they must make. Therefore the best remedy against confusion, panic, and dismay that I may be speaking an incomprehensible mish-mash of labels, terms, and hipster speak designating emerging tech and trends is the glossary at the end of this book.

Figure 1.2 “Portia the conference AI bot” hologram was livestreamed to the prestigious Swiss annual conference, the Applied Machine Learning Days (AMLD) in Lausanne in January 2019. Here you see the bot brain in its second iteration, as Amalia the Wayfinder with Portia as an alter ego who is activated with a spoken “wake word.” © AI BaaS UG, Munich, 2019

Please use it. Frequently. It is the best way to learn new stuff, especially terminology, definitions, and concepts, if you keep flicking back to the glossary to check your understanding and memory of various combined definitions that “crossover” or converge into in the fascinating world of what is now known universally as “AI.” Robots and Terminators aside, there is a lot of content that is quite straightforward and a “no brainer” once you read a short simple definition for a baffling term or concept. Use this book equally as a dictionary and a guide, a hands-on list of things to watch out for and selling points to check in any sales pitch or contract negotiation.

Practice certainly does make perfect. However, I truly understand the impulsive urge to press the panic button when this often hyped up field of science—and increasingly widespread “sales- inspired, misselling” fiction—becomes overwhelming and inconsistent. Unfortunately for some of the best innovators and entrepreneurs in our individual societies, wherever we may live or have emigrated to, the knee jerk reaction of purchasers has been paralysis.

Given the ongoing and sometimes astonishing media hype and outright misleading product/service descriptions by unethical companies, we should be wary of the public information about bots and their evolution. Very few academics are teaching the subject of chatbot or IVA history, even less about unethical technology marketing, as part of the curriculum, even though the number of bot conferences and chatbot-related topics including Mixed Reality installations are growing daily.

An Overview of Terms in the Glossary

Conference programs frequently refer to these probably unheard of topics outside of our tech industry bubble. The following terms mean the more sophisticated chatbots that perform at an advanced level of commercialized, often corporate, deployment:

• Conversational AI
• Conversational interfaces
• Interactive interfaces
• Cognitive interfaces
• Cognitive AI
• NLU bots
• NLP bots
• Chatbots with AI
• Mixed Reality
• Interactive gaming with chatbot avatars
• Virtual Reality using the aforementioned
• Augmented Reality, though interactive it has not yet been botified
• Intelligent Virtual Assistants (IVAs)
• Smart Assistants
• Next-gen bots
• AI bots and AI bots as a service (AI BaaS)

Reasons Obstructing Heavy Industry Applications and Commercialization

My company, trademarked as AI BaaS, is preparing to integrate our algorithm into some interesting heavy industry and commercial applications. The reasons it has not yet been done by us or our competitors until now is because of the following:

1. Most hologram software is built in Unity and independent source codes that cannot be integrated with, that is, the API is not compatible.
2. On Augmented Reality (AR) apps, again they are either built in an unsuitable API or do not allow for integrations with proprietary algorithms outside of the systems they were developed in.
3. The users, that is, corporate clients are not yet aware of the need to have an interactive bot within the Virtual Reality (VR) and/or AR experience.
4. The clients or owners of the application using VR and AR do not realize that their app is not truly interactive. It does not use voice or speech recognition (ASR), only touch interaction via screens or buttons on a device.
5. They do not yet understand that Natural Language Understanding NLU and AI allows us to inject an avatar into the experience to fully realize conversational commerce/AI.

These are just some of the technical reasons and obstacles to botifying more applications to create better experiences in Mixed Reality use cases. Following is a closer look at the attitudinal or psychological blockages that must be overcome by the prospective clients in order to see more widespread commercial uptake of chatbots, their 3D MR (Mixed Reality) next gen cousins, and AI applications generally.

What’s With the Attitude? Why Disruptive and Emerging Tech Gets Blocked by Human Minds

“If in doubt, don’t” often preached my conservative, ever pragmatic Franconian-German immigrant father (whose Bavarian mother had a German Jewish father) to his children in Australia. A saying paraphrased from the former U.S. president Benjamin Franklin. Probably nobody back then would have been cognizant of the need for a set of clones of inventor, activist, and newspaper owner Benjamin Franklin, America’s earliest thinker and technologist you could say. Franklin, who was the first ever American ambassador to France, shaped what came to be known as the liberal democratic American ethos with his antiauthoritarian values and practically adopted principles of puritanism and the Enlightenment. Brexit and Trumpian populism would have been worthy foes for Franklin if we recreated the polymath today and threw him into the fray—maybe as a combative AI bot hologram opponent?

Cautious hesitation to calculate likely outcomes was the polar opposite of my fifth generation Anglo-Irish Australian mother of First Wave merchant descent, whose motto resembles more “do or die,” an eminently sensible strategy for the often harsh climes and fiercely challenging evolution of society Down Under. The point being we are all determined by our cultural history and specific intergenerational experience and pass on our very own personal story in the way we make decisions, receive new information, and process it. Especially in a corporate or government context where high stakes SaaS multimillion if not billion euro budgets will be administered according to not only the organizational culture, but also the prevailing decision-making tendencies at large in the wider society in which your management is embedded.

I think about spending the past six years with my British husband who is a seventh generation English farmer in Kent. Agricultural assessment of risk takes on a whole new meaning with the “contextual” threats of Brexit, consumers’ behavior, and of course, climate change. The vagaries of the weather create uncertainties that seem to be comparable with the near indeterminable uptake of new tech and more importantly, whether it attracts investment and financing to achieve the aimed for scale-up. For a start, the farmer of arable crops is “globalized by default,” only being able to sell the results of a good or bad harvest according to international market prices, exporting tons of food and grain for industry via city brokers in London.

Whereas farmers are faced with these mammoth often overwhelming threats mostly on their own on a daily basis for years, tech venture teams are comparatively “lucky” in that you can rarely be a sole trader to launch disruptive innovations. That means you do need to find supporters and promoters—brand ambassadors—of your unique product or service beyond your “farm,” incubator territory, and psychologically move out of your comfort zone faster than a typical farmer.

Tech entrepreneurs need to pitch their ideas to the world as soon as possible to test “market traction” and sustain that user uptake of their new SaaS product. Unlike most farmers—who only do “wholesale” and do not have any form of B2C trade like with a farm shop or a customer-facing retail business—tech venture teams have a direct interaction with their consumers and must face them fearlessly and ask for feedback, right off the bat.

The Case of the Medieval Sales Machine: Sir Loin Your Virtual Butcher

Paradoxically, on our family farm bordering Greater London in North Kent, we trialed a bespoke 2D chatbot in 2016 who could have become a revolutionary force in the global meat trade: Sir Loin Your Virtual Butcher.

Sir Loin, as in the meat cut sirloin, was named in honor of the Early Medieval Knight Sir Adam de Bavant, who was rewarded with our manor house and large estate after fighting the Scots on a crusade with the English king, Edward I. His Majesty was reputed to have also knighted a piece of steak, the tastiest part of beef, thus the English name sirloin steak. However, linguists also point to the French origins of the term as well surloynge.

In any event, we thought Sir Loin would attract a local Kentish following and explain the new products in the farm shop, such as environmentally friendly pasture fed beef and lamb, which was reared as grass fed only livestock by us. And then the chatbot could rattle off all the health benefits and pronounce that the healthier animals created a healthier, leaner meat for humans. Environmentally, it was better for sustainable farming by eliminating the need for grain feed that is required for animals we eat rather than human consumption. This archaic “traditional” farming practice takes up a lot of land as a global resource that could be more efficiently used for human food production.

Figure 1.3 Prototype avatar for cooking, preparing food, buying fresh food, and recipes. “Sir Loin Your Virtual Butcher.” © Cliff Lee

Loads of complex facts to do with our farm shop and its new food lines were indeed successfully presented by Sir Loin in the form of an approachable chat on the farm website. He even conversed about my husband’s Green Credentials as a No Till, that is, no plowing farmer who was reducing the carbon footprint and fighting climate change using alternative farming methods like cover cropping. Sir Loin was a hit—in Seattle, San Francisco, and New York. Not sadly in his target market: Londoners and the Home Counties where we could deliver fresh meat packages when ordered online via Sir Loin.

Our medieval Knight with a special trade got such a following in the United States on Twitter that we actually met with a UKTI Export Adviser about how we might export:

• Fresh mince burger patties (beef, lamb, and game/venison);
• Bags of homemade pork crackling;
• Frozen meat pies; and
• Pasties and sausage rolls.

all made by our Court Farm Butchery & Country Larder—to thousands of geeky IT nerds who had rapidly become fans of Sir Loin. The British export advisers out of their depth and couldn’t grasp how the chatbot had gathered these demographics and statistics.

They exclaimed how did we gather this data when “normally” you had to do months of laborious research on the foreign market, the product segment, hire marketers over there, travel to the destination country to meet industry experts, join Chambers of Commerce to get introductions to partners to help with importing goods into their markets, and ultimately spend a lot of time, resources, and money over at least six months before trialing the export.

I tried to explain that Sir Loin was our special envoy who had managed to do a lot of that groundwork in a matter of weeks via chatting online and using the “pull factor” instead of “push.” In the end, there were just too many UK export hurdles for shipping fresh produce abroad. We had to abandon our plan to leverage sales from Sir Loin.

Sadly we had to switch off or deactivate Sir Loin Your Virtual Butcher after just six months working on our farm’s website. He wasn’t on our Facebook page because Messenger bots were not widely used or known at the time. Merely as an online 2D chatbot, customized for our family business as an SME, he was too controversial and “too clever.” Too many of the villagers, that is, our farm shop’s regular customers over decades were “freaked out” by this thing alive on their local farm’s website. Some of the feedback I got was “Is it going to steal my credit card details or e-mail address?” and others told me they refused to talk with Sir Loin because they simply weren’t “computerized enough” and preferred to chat with our human butchers whom they trusted.

Becoming an Advocate of Entrepreneurs

More often than not, an individual decision maker will not want to risk their own job “going in to bat” for an emerging tech entrepreneur. That is the bad news or worst case scenario for the entrepreneur. The “when in doubt do not purchase” mode kills off any turbo charged, hugely impressive sales pitch or carefully nurtured and immaculately conducted negotiation. There’s nothing you can do about human psychological needs and instincts for safety and security: “I will do anything—or nothing as the case maybe—to not endanger my job or position.”

The best case scenario is the opposite case: where an entrepreneur of new SaaS or innovative AI wins a corporate or government (internal) advocate who not only understands their USP and “cause” to bring this world-changing tech to market, but also undertakes to do that within their own company or organization.

These people are harder to find and engage for your entrepreneurial cause because once they get a reputation among startups in a city that they are “gate openers” as opposed to skeptical, unnecessarily negative gatekeeping guardians, then it is increasingly difficult for even the best ventures to get an audience with these perceptive, courageous individuals with vision and foresight.

The more robust entrepreneurs will court such decision makers vociferously seeing their opportunity to secure often multimillion dollar or euro contracts for their hard fought, preciously guarded portfolio of first-mover clients. Too many innovators falsely believe their uniqueness or genuinely “different” or more competitive / cost-effective software will “win” due to its merits alone.

Like with job applications and who gets the formal offer from HR, this is sadly not always the case. The best “interview performer” very often wins, much like the best performing “pitchers” of new ideas, whether they are genuinely new and unique, competitive, and productive or not. Be aware of this scenario and its pitfalls from the outset.

This is like enacting a play or drama over and over. The script is familiar—passionate, terribly eager entrepreneur, cool distant purchaser, months of information exchange, due diligence, and then the final yes or no, just for an (often unpaid) pilot of this emerging technology.

What Is Preventing Purchasers From Buying

Yet who wins this “do or die” opportunity may not always be the one deserving to win based on simple merits, due to a variety of reasons that I will explore in this book through case studies, commentaries, observations, and the guidelines at the end. A brief indication is:

• Purchasers not having the experience to distinguish hype from reality
• Not knowing where the correct or most relevant knowledge base can be found
• The independence of analysis fails due to “influencers” of the wrong kind
• Not being given the essentials of how to upskill themselves in a hurry to make an assessment of the proposal put forward by the tech entrepreneurs (the RFP or estimate)
• Not understanding that business chatbots basically fall into just two categories:
  1. Bespoke, boutique bot customizations created outside your company.
  2. “Off-the-shelf” bot APIs that are then tweaked and customized inhouse by your webmaster or a specially designated R&D team, often known as a greenfield project, with test implementation/deployment or pilot, for example, the summer 2019 Mixed Reality pilot of “Semmi” by Deutsche Bahn in the Berlin Central Station.

These issues remain a matter that must be repeatedly addressed openly and in a fearless way on a regular basis given the speed of scientific and technological advancements impacting on the whole world, not just your society, your city, your company, office, and ultimately, your job.

In conclusion, let’s reflect on my definitions and observations from my 2016 article for the San Francisco tech news platform VentureBeat, cited earlier in this chapter. Why are chatbots disruptive? My VB article is still being read today because over three years on, things have not essentially changed! Why isn’t everyone aware of the disruptiveness of chatbots? Why isn’t every vertical deploying one?

One reason is that bot developers have become the arch enemy of the advertising industry, as I explored in my post about how the transparency of the user engagement metrics expose the lack of transparency in the legacy reporting schedules of the advertising middle men, the large and mid-sized media agencies who have tried to stop the advance of these independent customer experience measurements. Ultimately it will put them out of business if they cannot justify their multimillion ad spends with accurate, coherent tracking of sales leads to conversion rates, the demographics of the opted in consumers, and the outcomes of the incredibly lucrative digital out of home (DOOH) ad and marketing accounts they have held for decades for most brands you know.

Why the Advertising Industry Is Threatened by Disruptive Bots

Why don’t the advertisers like us? Because of these key reasons:

1. We cut their profit margins.
2. We can deliver transparent metrics unlike most other forms of advertising.
3. The middle men known as media agencies or advertising agencies sell “ad inventory” to big brands and multinational corporations. More accurately they buy the digital ad space from providers such as DOOH billboards, Wayfinder, outdoor signage, and online platforms. And SEO native ad placements in real time.
4. Their profit margins have been cut by Facebook and Google, among other digital ad platforms like native advertising providers and auction places, because these U.S. tech giants do not need the middle men to sell their ad inventory.
5. The brands and businesses, company owners, and marketing directors can buy the ad space and campaigns direct from FB and Google, for example. Meanwhile the purchasing of FB campaigns has become ever more automated and user intuitive, so easy for an unskilled person to utilize. The service supposedly has transparent metrics.
6. I say supposedly because Mark Zuckerberg himself said at the FB staff pow wow in San Francisco a few years ago, confessing that their ad inventory sales system was not working.
7. Instead of trying to drive revenues through selling advertising space to small business owners and large corporations, FB would instead devote their resources to a far more profitable revenue stream from January 2017 onward: chatbots on its Messenger platform. And probably its newly acquired WhatsApp instant messaging app to assist in a near total monopoly of the new digital advertising, global landscape!
8. Microsoft’s Satya Nadella followed Zuckerberg’s chatbot initiative, declaring Microsoft Bot Network and its pervasive deployment of Cortana as the new world dominating B2C and B2B chatbot interface would be the new revenue stream to beat all others.
9. In fact, Cortana was deemed to become the world master bot for all other chatbots. Something that has not eventuated since these Microsoft and Facebook declarations in 2020 and unlikely to occur as these tech giants have failed to demonstrate true AI bot capabilities in the evident lack of user adoption of Cortana and the many Messenger bot flops.

Then 2017–2018 saw the hype slump to serious turnover “lows.” Now at the end of 2019 we are seeing a hype cycle begin again. This is evident in all the trade show invitations that land in my inbox daily—an AI bot conference on somewhere in the world every month! 2020 will be a boom year and finally see the predicted trillion dollar growth, slowly but surely.

At last a more modest upswing in bot orders and sustained deployment of their avatars, not just online. Being confined to 2D online chats in the past were highly restrictive and actually prevented users adopting chat. It wasn’t really conversational AI as it wasn’t AI at all as I look at in Chapter 3 with the Japanese Hologram Wife.

Confidentiality Agreements Mean Fewer Case Studies

Several of these case studies are about competitors to the company I co-own. The rest are about the Mixed Reality 3D avatars created bespoke by our new tech venture in Bavaria, AI BaaS UG in Munich, as well as the frequently frustrating, “sales cycle” and long lead time situations we repeatedly found ourselves in with my predecessor 2D chatbot startup velmai Ltd in Devon, UK.

Figure 1.4 Most of our old venture’s avatars were 2D chatbots that we deployed on websites and a few on Facebook Messenger for short experiments. Here you can see Albert the Online Butler who also switched to the author’s Private Secretary using a wake word, plus Hans-Pierre who is a tourist guide and concierge working in tourist regions in DACH countries and the South of France. © Cliff Lee in Devon

Figure 1.5 Mixed reality installations are now at the forefront of point-of-sale user interaction, especially in large public spaces and where multilingual human–machine interaction is required to automate communications and enhance the customer experience. © AI BaaS Munich 2019

The reason for the few number of competitor case studies is that we are not privy to the production details, manufacture, and deployment issues of many of the “public cases” like Microsoft Tay.ai, Semmi, and Franny by Deutsche Bahn or the Labour Party of England and Wales. For the outsider, even for me as an expert commentator, we can only speculate as to why things go wrong for bot developers when they do.

Very few of them will ever release details or turn their pilot into a case study as I have done with my companies’ projects here. For the tech ventures I have co-owned, I am able to publish details about problems in real-life situations, even where I have had to anonymize the client or facts and figures for the sake of our industry typical nondisclosure agreements (NDAs) and to ensure that our confidentiality agreements are kept intact.

---