The future of AI: ridiculous, dangerous or obvious?

Artificial intelligence disturbs, scares and questions. My Digital Week asked Mohamed Senhadji, Manager Sales Strategy & Customer Experience at SQLI, to share his thoughts and experience on the subject.

MyDigitalWeek:  Can AI be seen as just another technology that helps to improve and perfect existing processes or is it more than that?

Why?Mohamed Senhadji: Science fiction and Hollywood have really stirred up people's imaginations about AI, raising fears in the most sceptical among us (we're still a long way from the singularity: i.e. self-improving higher intelligence). AI is currently a catch-all term. There are several such technologies and today these are mainly technologies used to process and analyse very large quantities of data in order to identify system patterns or models.In my opinion, in the same way that the development of the oil industry and the discovery of radio waves revolutionised and shaped our world today, AI will shape the world of tomorrow and have a major effect on society. It will definitely do more than just perfect existing processes and will create new ones.

In the early 2000s, the main aim of AI research was to "intelligently" delegate repetitive and low-added-value tasks: e.g. autonomous driving and automatic message processing. Certain recent exploits have seen it display a form of intelligence: this was particularly the case when DeepMind's AlphaGo programme beat the best Go player - this was not expected for another twenty years.

The Chinese AI that enables much more accurate identification of brain tumours successfully established 87% correct diagnoses out of 225 cases in just 15 minutes. And the doctors? They achieved 66% accuracy in 30 minutes. For expanding cerebral haematomas, the AI established the correct diagnosis in 83% of cases compared to only 63% for the group of doctors.Similarly, in the legal profession, Casecrusher, the AI lawyer, beat by practically the same rates a group of lawyers examining cases of mis-selling of insurance policies.

Clearly, in situations involving the processing or compilation of a large quantity of data for analysis, humans have already been left behind, but what's more striking is when AI focuses on artistic fields: i.e. AI used to create classical music (Aiva, which released its debut album last year) or imitate painting styles (CAN, which was used to present works to experts, who couldn't work out whether they had been produced by a master painter or a machine).As I mentioned earlier, we're still a long way from the singularity, we're only in the early stages of what AI can do as all the examples I mentioned are based on supervised learning.

But as Amara's law tells us, we tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.So yes, AI will do much more than improve existing processes: they'll be significantly effected. And the effect that, in my mind, hasn't yet been examined enough is the one that all these technologies will have on society. Do I need to see a doctor or should I go AI? Hire a lawyer or go AI? So should we keep training doctors and lawyers in the same way? Will we still need certain specialities? Will museums exhibit works produced by a machine? Can they win awards? Or in a different vein, how will driverless cars be insured? 

MyDigitalWeek: How will artificial intelligence revolutionise user experience?

Mohamed Senhadji: We switched from a product-based approach to a service-based approach in the 2000s (with printers and Nespresso machines that developed a business model based on consumables), then to an experienced-based approach. The focus is no longer on the service or product being proposed to the customer, but rather on how they will perceive them, and even the emotions that will be generated or sustained to develop satisfaction or engagement. For me, AI will have a very important role to play in this transition.It is already partially used to anticipate needs, process requests more quickly or delegate low-value tasks, e.g. making appointments.

Everybody was blown away by Google's video showing Google Duplex, a robot imitating natural language - even including impressions, which makes appointments at the hairdresser's or books restaurant. Increasingly-advanced semantic processing in chatbots and voicebots will provide a new after-sales (but also before-sales) service design for providers of products or services. For customers, these after-sales services will be perceived differently, which will raise a number of questions. I'll get back to that later.So CX, Customer Experience, will slowly but surely make way for its successor - HX: Hybrid Experience.

Suppliers will no longer have to contact an individual customer but rather a Customer – Assistant pairing. A sort of tripartite relationship between the provider of the product or service, the customer and their AI companion. Part of the customer's knowledge and experience will be delegated to this AI, and the supplier will have to interact with the AI to make sure its service meets the customer's expectations as well as possible. Just imagine a "Jarvis", like in Ironman, who helps a customer with a range of purchasing experiences: in practical terms, AI that records the customer's trends, style and purchasing history to match an item with what they've already got in their closet.

This tripartite relationship can also be designed the other way round, this time with the AI on the supplier side. The customer doesn't contact a salesperson representing the supplier, but rather a Salesperson-Assistant pairing (and also have no way of knowing whether they're contacting a human salesperson or a bot). The complementarity between the simplicity of processing a large quantity of data provided by the AI and the human touch of the salesperson should improve this customer experience.

But are we ready for these revolutions? Will we need to say that it's a machine that's calling to make the appointment or that you're talking to a machine when you contact after-sales service? How will this be perceived?Another key issue that comes to mind is the disparity in enthusiasm for these technologies between the 3 regions: US, Europe, Asia (more specifically China). There are Asian countries where nearly 70% of people accept these new tools and the positive effects they're set to have on society, but these rates drop to 30% in the US and are even lower in Europe. We at SQLI are here to help to change things. 

MyDigitalWeek: What is the publishers' strategy? Can you describe the Lab's missions?

Mohamed Senhadji: At SQLI, and more specifically at SQLI Consulting, we offer a Customer Experience range that, in addition to other aspects, includes a significant AI dimension. So we work on several levels in this field:

  • A Strategic dimension: we help customers to identify the impact of the implementation of this type of solution on their processes, in terms of their sales strategies or the customer pathway, including highlighting of their potential return on investment.
  • An Operational dimension: we help customers with implementation or assistance with the implementation of off-the-peg solutions. Our works includes defining the customer's need, and managing change with employees, as well as coordinating the project, by identifying the right indicators to manage the project and the advantages of the solution adopted.
  • A Prospective dimension: with the set-up of technological and societal monitoring tools for customers in their area of business, by proposing in-house monitoring centres or labs. These labs' activities include launching POCs (Proofs of concept) to test features with end users. Otherwise, we also rely on the work of our SQLI Lab, which adopts more of an all-out prospective approach rather than focusing on a single area of business. They attend various trade shows to learn about the latest trends for us, interact with the industry's start-up ecosystem to identify newcomers then develop interesting partnerships based on the requirements specified by customers. Of course, SQLI lab also helps to launch labs for customers, as I mentioned earlier. In a different sphere, we also have an AI and Customer Experience Chair project with the academic world in France: a group of engineering schools, business schools and university laboratories. It's still early days for this project: I'll look forward to telling you about it at the end of year when it's a little further along.

MyDigitalWeek: Got a message to get across?

Mohamed Senhadji: Despite all these questions, I'm certain that we're seeing the first signs of a revolution. And like any revolution, I'm sure it will go through these 3 phases. Ridiculous: it'll never work, Dangerous: the survival of humanity depends on it, and Obvious: there's nothing weird about having a personal digital assistant, or speaking to a robot to access certain services.At SQLI we want to be active participants rather than simple spectators faced with all the opportunities that Artificial Intelligence could generate and get our customers involved, too. So, get ready to get on board for what promises to be quite the experience.

Published in MyDigitalWeek