Registration is wrong! Please check form fields.
Publication detail:

Dialogue with the Invisible Jeeves

Dialogue between humans and machines is a reality of our life at present. But where is it taking us? Will technological developments make robo-advisors and personal assistants assume quasi-human characteristics eventually – from artificial intelligence at its core, to the pleasant voice modulations of a good actress or actor inviting emotional attitudes. The answers are not clear yet, but we can foresee that the ability to place basic human needs and interests at the centre of the entire ecosystem, and serving them most effectively in terms of cost, speed, and quality will shape the future course of both technological advancement and market developments. Most importantly, it will unfold as the learning process in a dialogue with us through a Voice User Interface (VUI).

What is this dialogue about – an invisible and rather unobtrusive Jeeves responding to daily needs of a typical modern-day Wooster? Most of the projections focus on automating lifestyle services such as basic financial planning – from moving savings to an account with higher interest rates to balancing the portfolio in line with risk profiling and analysis, ordering food and household goods, or using a VUI for conversational programming of IoT[i] objects, booking hotels and arranging things for dining out with friends, travelling, going for sports, using health services, planning and paying for education, insurance, etc. Even advising on the style of a dress or gift is a possibility. Constant background analysis of personal choices and preferences as well as smart use of statistics and big data will stand behind the higher cost/quality ratio than the best human PA could deliver. The psychology of communication, sentiment analysis, and conversational excellence will probably be among other methods of achieving excellence.

Machine dialogue is with us alreadyIt starts from a simple voice-driven IVR[ii], where a customer of a telecom or a bank does not have to wait on the line listening for all the complex menu options of an old-style DTMF[iii] IVR forgetting the first few options by the time of getting to the last ones, and pressing buttons randomly to get to a human operator (unless the connection is interrupted suddenly by the call centre, which can be truly maddening). Instead, you simply say what you want and machine deals with it, as it has been trained specifically to understand all the possible ways and forms in which you can formulate, articulate, and voice the reasons why you are calling. It can also get a sense of whether you are happy with the service or not. Even if you are mad, it will transfer you to a specialist who can help (it’s not a joke).

What makes this dialogue possible: speech recognition and analytics, semantic interpretation techniques enabling contextual understanding, voice biometrics to verify human identity for processing highly personal and confidential data, a VUI layer for universal chatbots and AI platforms based on machine learning, 4-5G and computing power, big data, and importantly – psychological preparedness to communicate with the machine based on an increasingly smooth customer experience.

What are the main obstacles:accuracy of speech recognition, technical imperfections (noise filtering and, generally, managing acoustic environments)? As it stands today, it is increasingly about innovation in algorithms and the right amount of data for machine learning. Technologies like those of Spitch AG will deliver the greatest accuracy if trained on a massive amount of client audio data from the same very end-users that the products based on these technologies will be serving.

Any concerns:(1) can the machine/AI harm humans, and (2) can we harm ourselves by treating the machine as a slave. This invites a more philosophical question of whether artificial intelligence, when/if it fully materializes, is a form of semi-autonomous artificial life and personality, essentially involving a new kind of ethics, or human relationships ethics would apply. In the foreseeable future AI will remain fully dependent on human programming. Chatbots will too, by default. The danger, therefore, can only be rationally conceptualized as being associated with human mistakes. The powerful image of “Skynet” terminating life on Earth or mismanaging your pension funds and deleting all the records is disquieting enough to make us think about real but less conspicuous and more distant threats. The old notion of prevention being better than cure is valid and important in this context.

Personal Assistant (Travel Case)

Where are banks and money in this picture?KPMG is envisioning an “invisible bank” of 2030:

“… there is no ‘banking app’ – access to money is interwoven with health, time management, leisure and friendship. Visiting a bank will be as alien a concept as picking up the landline (“What’s that?” my five-year-old son said recently). Banks will be just as invisible, but just as vital, as the manufacturers of 4G base stations are today”[iv].

The KPMG vision for retail banking in 2030 is one of a disaggregated industry – with three distinct components. The first layer, represented by universal personal assistants like EVA in the KPMG example, is the Platform Layer. Together with Product and Process layers, the banking industry is set for a period of significant structural reform. At present, one can see only incremental improvements in customer experience such as contactless payments, faster onboarding processes etc. Genuine, transformational innovation is rare.

“But customers are increasingly using other channels to fulfill functions previously dominated by banks. The arrival of services such as Apple Pay – to which most banks have signed up – hints at a future where financial brands are hidden behind devices…

Bank brands remain highly trusted. Some would argue they could develop lifestyle layers to compete in this Platform space. This is one possible scenario…”[v].

Other scenarios are dependent on the evolution of smartphones and other personal mobile devices and their ‘always-on’ voice user interfaces enabling access to essential banking services, robots, and PAs surpassing the scope of human capability as far as the speed and volume of data that they can process.

At Spitch AG we are excited to be taking part in the effort to transform this space for you and your customers alike. We wonder what are you feeling about it? Please share!

  • [i] IoT – internet of things
  • [ii] IVR – interactive voice response
  • [iii] DTMF – dual tone multi frequency – is the signal to the phone company that you generate when you press an ordinary telephone's touch keys
  • [iv] KPMG: «Meet Eva» (downloads PDF)
  • [v] As above