Customer Relations in the era of Artificial Intelligence

intelligence artificielle relation client

In June 2017, the cover of Relation Client magazine featured “The year of artificial intelligence”. As reported by the magazine article in question, “artificial intelligence, or augmented intelligence, made its entry into the domain of the customer experience in 2017. Combined with pertinent data, AI opens up perspectives for customizing customer relationships and anticipating consumer behavior”. To dig deeper into the subject, we asked 3 questions to Benjamin Alquier, our Director of Product Management.

Benjamin, in your opinion, what role does, or will, artificial intelligence play in our sector of activity ?  Will it fundamentally transform the customer relationship and the customer experience ?

1) AI in direct contact with the end-customer : this may concern new media such as chatbots (aka conversational agents), or new natural language interfaces, such as “Amazon Echo” (a connected loudspeaker capable of answering questions by means of the Alexa intelligent personal assistant), or the “Apple HomePod” (Apple’s intelligent loudspeaker based on its Siri voice AI technology, which will go on sale in the USA during December 2017).

Artificial intelligence can also be used in more integrated applications, such as personalized customer services – as in the case of the SNCF application that predicts your next destination – or more prosaically for improving the relevance of FAQ responses, or in the field of selfcare1.

AI is also heavily used to optimize customer interactions by means of predictive scoring mechanisms2 and can thus improve the performance of promotional or loyalty campaigns. This 1st level application of Artificial Intelligence brings more reactivity – to the point of immediacy – in processing customer requests.

2) AI for advisers : this provides assistance – one speaks of an “augmented agent” – or helps with prioritization and pre-qualification. For example, in the banking sector, where a full credit study used to be necessary to obtain a loan, banks are now able to give a very quick response to customers. It can also prompt call-centre agents with appropriate replies to customer questions, as is the case with Natixis. This 2nd level is more of an evolution than a revolution, and optimizes the relevance and effectiveness of the Customer Relationship.

3) AI for all levels of management : this makes it possible to understand, analyze and improve processes, products and experiences. In this domain, we speak mainly about support for decision-making. This 3rd level acts more deeply within organizations.

Although the 1st level – the direct encounter between AI and the customer – arouses buzz and fantasies, certain signs lead me to say that the other 2 application levels are deployed sooner because they are less risky (not head-on with the customer), and preserve a degree of humanity in the relationship with the brands. Experiences with chatbots have shown that this humanity is key: customers appreciate it (as long as the service works efficiently), and brands too because customers feel more engaged and involved. This is the way that Google, a pioneer, this year launched the PAIR program (People + AI Research), which aims to streamline the relationship between AI and humans.

How can Artificial Intelligence help to understand the wishes, expectations and needs of customers even before they are expressed ?

It makes it possible, for example, to correlate similarities in the profile of “customer A” with that of “customer B”, as Amazon and Netflix are already doing. For example, if customer A has bought a book on Amazon along with an additional product, the site will suggest this second product to customer B (who has substantially the same profile) if he buys or simply clicks on the same book as customer A. This use case has been successfully employed by Telecom operators for the past 10 years to prevent the risk of attrition and stimulate usage.

Historically, AI was first applied to quantitative data in order to predict risks (of attrition, of payment default…) or predilections (for purchase, for usage …). Today, AI is becoming more and more powerful for processing qualitative, unstructured data. This opens up other perspectives: for determining the root causes of these risks and predilections, and allowing brands to address them.

How is Artificial Intelligence incorporated into a solution such as Instant eXperience® ?

Benjamin Alquier

Benjamin Alquier, Product Management Director at MediaTech Solutions

Instant eXperience® is already an artificial intelligence solution, in particular for the enrichment of collected data:

– Semantic analysis (recognition of concepts and intonations in a verbatim),

– Automatic translation of verbatims classified in categories,

– Transcription (speech to text) of audio verbatims.

These three modules are built on artificial intelligence mechanisms, and continue to improve thanks to them.


AI is also present in the analysis of results :

Automatic root cause analysis : this report – in beta-version – analyses an indicator in order to identify the causes that have the greatest impact in an upward or downward direction.  For example, in the case of a major transport operator, our solution examined a hundred variables and demonstrated that the one which plays the most decisive role in customer satisfaction is “lateness”, with a threshold at 10 minutes; in second place is “environmental noise”. This has allowed our customer to put the importance of the “price” variable into perspective, a key insight for the company !

Detection of weak signals : Instant eXperience® allows trends to be detected, for quantitative criteria of course, but also in forwarded verbatims. Our solution offers the possibility of detecting the fact that certain verbatims about a particular concept, whilst still in a minority, are steadily growing and thus merit attention.

Scoring and prediction of satisfaction : the solution automatically predicts customer satisfaction after a contact, and immediately triggers corrective actions (eg an outgoing call) or intensifying actions (by boosting promotion and suggesting that the customer publish his/her opinion).

Optimization of response rates : it is also possible to predictively optimize the response rate and the level of customer solicitation by contacting the client at the optimum moment so as to maximize his likelihood of responding, and dynamically customize the level of customer pressure according to his profile…

Feedback Management3 solutions have a prominent place in the field of artificial intelligence. Although the customer satisfaction rating is the #1 prerequisite as a measured result, we have access to a considerable amount of information about customers and their interactions. The « data lake4 », a fashionable concept, is at hand !


Selfcare : a practice whereby the customer is given the possibility of carrying out actions, in complete autonomy, to manage his account or to make use of support information. In the domain of selfcare support, the customer can access knowledge databases, diagnostic tools, FAQs, forums and other items of community support (ex: the management of online bank accounts is a form of selfcare). Selfcare allows the initiating company to reduce its customer support costs and can be considered as a useful service by customers looking for autonomy, for services available 24H/24H and the avoidance of any waiting time on the telephone (www.definitions-marketing.com).

Scoring : a technique that allows a score to be assigned to a customer or a prospect. The score obtained generally reflects the likelihood of an individual responding to a marketing solicitation or belonging to a target audience. It therefore measures the appetite for the potential offer (www.definitions-marketing.com).

Feedback management: Feedback Management consists of capturing, enriching and exploiting the feedback of individuals (customers, employees, and partners) and/or information systems, in order to maximize the performance of an organization, by optimizing the customer experience and developing the commitment of all stakeholders.

  • Capture feedback and data for key customer experiences and at key moments, across all channels of interaction (points of sale, phone, digital)
  • Enrich this raw information with voice/text analysis and artificial intelligence technologies to maximize its exploitability
  • Leverage the enriched feedback by disseminating it in real-time, in a personalized manner, to all stakeholders in an organization, along with the actions to be taken – both strategically and operationally in each department and at all levels of the organization

By enabling continuous improvement of customer, product and service experiences, but also by enriching managerial tools for a greater “customer centricity”, as well as by integrating an efficient means of engagement for all the stakeholders, Feedback Management offers the creation of sustainable value for the company. From the strengthening of brands to the development of turnover and profitability, all these levers can be impacted in a tangible way.

Data lake: a space for storing large quantities of heterogeneous raw data, allowing cross-exploitation by means of reports, visualization, machine learning… (https://en.wikipedia.org/wiki/Data_lake)


By | 2017-09-04T16:06:13+00:00 09, 2017|Feedback Management|
Please wait...
Envoyer un message

Nous ne sommes pas disponibles. Envoyez-nous un mail !

* Nom
* Email
Login now

Need more help? Save time by starting your support request online.

* Votre nom
* Email
* Message
Nous sommes en ligne

Help us help you better! Feel free to leave us any additional feedback.

How do you rate our support?