How does an LLM-based chatbot improve customer autonomy while enhancing service quality?

Faced with changing usage patterns and the need to respond ever more quickly and efficiently to customer requests, G7 (France’s leading taxi company) has made an ambitious commitment: to leverage generative artificial intelligence to improve its customer relationship. To meet this challenge, the company collaborated with Dydu, an expert in conversational AI, to deploy a chatbot based on a language model (LLM).

This approach is resolutely focused on innovation, but also on control, mastery, and quality. Let’s take a look back at this transformative project.

A clear objective: streamline customer relations without dehumanizing them

G7 is a company built on a dual promise: to improve mobility for passengers and simplify the daily lives of drivers. With 32 million passengers transported each year, 15,000 affiliated drivers, and more than 10,000 corporate customers, customer relations are a central and ongoing challenge.

The company already had an AI-enhanced help center, but wanted to go further with a tool capable of understanding natural language, adapting to context, and responding reliably, while managing risks with a “restricted” and configurable LLM.

It was in this context that G7 saw an opportunity in LLMs: not to replace its teams, but to better guide customers, offer them greater autonomy, and free up time for high-value requests.

Three objectives were identified:

1. Enable customers to quickly find answers to simple questions without having to contact an advisor.

2. Ensure consistent response quality when human contact is still necessary.

3. Leverage data from interactions to identify weak signals and adjust business and product priorities.

An LLM chatbot co-built with Dydu: a hybrid and pragmatic approach

G7 and Dydu opted for a hybrid approach, combining the power of LLM and business structure. The chatbot, now deployed on the G7 website, is structured around three response levels:

• Level 1: for simple, frequently asked questions, the LLM generates a natural response based on the data available on the website.
• Level 2: for more complex topics (e.g., refunds, legal framework), responses are constructed using more structured formats (decision trees), with intelligent AI support.
• Level 3: for sensitive cases (e.g., a person with a disability who needs a specific service), the request is forwarded to an advisor via a ticket automatically enriched by the LLM.

This architecture combines responsiveness, quality, and security in request handling, while ensuring a smooth experience for the user. Above all, the Dydu solution was designed to be used directly by business teams, without requiring technical skills.

Measuring performance: a new challenge for the LLM era

The chatbot testing phase began with the sending of hundreds of simulated questions, which were analyzed manually to assess the relevance of the responses. However, although valuable, this method quickly proved too time-consuming to be sustainable. G7 and Dydu therefore worked together to implement an automatic performance evaluation system tailored to the specific characteristics of LLM models.

The first challenge was to define the right indicators. Traditional chatbot KPIs (such as the non-comprehension rate) become obsolete because an LLM almost always provides an answer, even if it is not sure of its relevance.

It was therefore necessary to design new evaluation criteria, focused on the confidence of the model in its own answers. The LLM now self-evaluates each response using several techniques:

• A factual analysis, based on the number and diversity of sources used.
• A semantic analysis, which measures the gap between the generated response and the reference documents to identify any discrepancies.
• A probability score that reflects the relevance estimated by the LLM itself.

At the same time, the chatbot also uses AI to analyze the intentions and classify the topics addressed by users. This makes it possible to identify the most frequent or sensitive topics and prioritize improvements, particularly in cases of low confidence scores. This intelligent and scalable system is an essential foundation for continuously managing and improving the quality of the responses provided by the chatbot.

Improving the chatbot: ongoing, structured, and iterative work

The approach adopted is based on a continuous improvement cycle structured around three levers:

1. Improving the algorithm with a dedicated prompt: to frame responses and avoid digressions or hallucinations.

2. Improve the knowledge base to make it suitable for use by AI: add missing content, enrich existing articles, and clean up obsolete data.

3. Create decision trees: for specific, complex, or sensitive cases in order to control the responses provided.

This approach has not only improved the quality of responses, but also enhanced the reliability of the service for users.

All of this is part of an iterative process, carried out regularly with the G7 and Dydu teams, well beyond the initial production launch. The project does not stop when the chatbot goes live: it continues to evolve to remain effective and relevant.

Managing escalations: uncompromising efficiency

Another important aspect of the project was managing escalations to human teams. G7 has adopted a clear policy: it is better to escalate a request than risk giving an inaccurate answer.

Thanks to AI, sensitive tickets are now:

• Automatically detected, based on context or certain keywords.
• Pre-filled by the LLM, which saves time for advisors.
• Prioritized according to their complexity or criticality.

The result: more responsive, more efficient, and better informed customer service.

Very encouraging results from the very first weeks

G7’s structured and gradual approach quickly paid off. From the very first iteration, the results were satisfactory:

Before optimization:

• 50% of responses were very satisfactory
• 45% of responses needed improvement (often due to a lack of clear content)
• 5% of responses were unsatisfactory

After improvement:

• 95% of responses were very satisfactory (including cases that were escalated appropriately)
• 5% of responses needed improvement (sensitive topics that were not suitable for 100% automated processing, but possible with a hybrid approach)
• Less than 1% were unsatisfactory

What’s next?

The chatbot is just a starting point. G7 already plans to:

• Extend it to other channels (social media, apps, etc.)
• Connect it to its back office to broaden its scope
• Adapt it to other audiences, particularly drivers
• Explore new formats, such as the callbot

A human partnership, driving innovation

Beyond the technology, this project is also the story of a strong partnership between the teams at G7 and Dydu. Clear project organization, regular communication, and a shared commitment to continuous iteration and improvement have enabled us to overcome the challenges of generative AI.

“AI doesn’t replace humans. It enhances them. It helps us serve our customers better, faster, and with greater relevance.” –Samir Dilmi, CRO at Dydu

The French public procurement agency UGAP (Union des Groupements d’Achats Publics) called on Dydu to deploy a conversational bot for its clients. It has racked up more than 1.8 million visitors since its launch and boasts a 95% interaction qualification rate.

Achylle EBELE, assistant director of the multi-channel contact center, discusses this success in an interview for Dydu’s Client Focus.

UGAP is the only “generalist” public procurement agency in France. It provides access to more than 1.4 million products in over 3,500 markets. Its clients include public bodies (ministries, local authorities, hospitals, etc.) and public service institutions.

Customer relations at the heart of UGAP’s strategy

UGAP recorded 5.6 million orders in 2022 and has made customer relations a priority by implementing:

• a regional sales network (25 regional offices) to be as close as possible to its clients
• an online sales site: UGAP.fr (more than 5 million visits/year)
• a multi-channel contact center (200,000 calls/year)

These three entry points provide clients with a wide choice of access to the company’s products and services.

The multi-channel contact center, created in 2011, provides simple answers to a range of queries within a short timeframe. UGAP set two objectives to achieve high satisfaction levels:

• 90% pick-up rate
• 90% next-day resolution rate

Digitalizing the customer relationship

UGAP wanted to provide its clients with a more innovative and digital experience. The aim was to offer a simplified user journey, with ever shorter response times to simple, recurring questions. So, the company turned to conversational AI solutions.

The Dydu chatbot was officially launched in February 2022 on the UGAP.fr website. We installed an API to enable the bot to inform clients in real time of their order delivery status. The tool reduces the number of requests and increases satisfaction with a shorter response time.

The bot recorded more than 16,000 interactions in the first year. Clients use it primarily to request information about products, despite the fact that the website includes a FAQ section.

Both channels have recorded more than 45,000 interactions to date. The sharp increase in the number of users has led to new challenges for the company.

Given the use of the bot and the large number of user questions, UGAP decided to implement MailJet.
This tool retrieves complex client requests and e-mails them to the multi-channel contact center. The support team then contacts the client once the request has been processed.

This solution is the first step before implementing a live chat system, which will increase customer support.

TotalEnergies is the 3rd largest electricity and gas provider in France. It provides electricity and natural gas to more than 5 million consumers and businesses, as well as innovative services to optimize their consumption. Innovation and customer satisfaction are at the heart of the company’s strategy.

After launching a live chat service in 2014, TotalEnergies wanted to further innovate its customer experience. The aim was to be available 24/7 while relieving its support teams who were overwhelmed with requests.

In order to meet these new challenges, we helped TotalEnergies implement a conversational chatbot for consumers that redirects users.

This conversational agent now has 300 knowledge articles and generates more than 2,500 conversations per day, thus automating tier 1 customer service requests.

In an interview for Dydu’s Client Focus, François Rémuhs, Tier 1 Customer Service Activities Manager at TotalEnergies, looks back at this success. He discusses the company’s initial challenges, the project implementation, results and next steps.

Download our client case study to find out more about TotalEnergies’ customer support chatbot and its key figures:

TotalEnergies is the 3rd largest electricity and gas provider in France. It provides electricity and natural gas to more than 5 million consumers and businesses, as well as innovative services to optimize their consumption. Innovation and customer satisfaction are at the heart of the company’s strategy.

After launching a live chat service in 2014, TotalEnergies wanted to further innovate its customer experience. The aim was to be available 24/7 while relieving its support teams who were overwhelmed with requests.

In order to meet these new challenges, we helped TotalEnergies implement a conversational and redirectional chatbot for individual customers.

This conversational agent now has 300 knowledge articles and generates more than 2,500 conversations per day, thus automating tier 1 customer service requests.

In an interview for Dydu’s Client Focus, François Rémuhs, Tier 1 Customer Service Activities Manager at TotalEnergies, looks back at this success. He discusses the company’s initial challenges, the project implementation, results and next steps.

Download our client case study to find out more about TotalEnergies’ customer support chatbot and its key figures:

The car manufacturer Renault called on Dydu to deploy an internal conversational bot for its salespeople and sales managers. Stéphanie Lecomte, Head of Digital Transformation and Sales Tools, looks back at the implementation of the chatbot and shares the next steps for the future.

Can you describe your company and your role?

My name is Stéphanie Lecomte. I’m the Head of Digital Transformation and Sales Tools. I’m in charge of all the sales and CRM tools and the digital transformation of our network (e.g., highlighting the network stock on our website or booking test drives online).

What are your current IT support challenges, and what issues did you want to address by implementing a chatbot?

We wanted to address cost and availability issues.

We wanted to end phone support while maintaining a high level of support for our salespeople. Chatbots are efficient solutions that cost less than phone assistance.

Chatbots are also available 24/7, whereas our call center was only available during working hours.

Before we set up our chatbot Maeva, we couldn’t help our salespeople on open days because they mainly take place on weekends.

We implemented the chatbot for Renault and its network in France, which includes around 370 businesses and 5,000 users (4,000 salespeople and 1,000 managers).

Why did you choose Dydu?

We’ve been working with Dydu on our internal HR chatbot for several years. We’re very pleased with it, and it meets all our expectations. So, when the IT team wanted to set up a support chatbot, we naturally turned to Dydu.

What scope does your chatbot cover?

Our chatbot, Move Maeva, assists our salespeople and sales managers. It handles tier 1 issues relating to their sales positions or CRM tools.

How many people manage the chatbot internally, and what do they do? (bot administration, updating the knowledge base, monitoring dialogs, analytics, etc.)

We have a team of 6 people managing the chatbot, each specializing in a different area.

We organize reading workshops once a month to analyze the chatbot’s dialogs and improve our knowledge base.

How did you build your knowledge base?

To get started, we used an Excel file that was created during our sales and CRM tools audit.

We then added any questions raised by our pilot group of salespeople to build up the new chatbot’s knowledge base.

Since rolling out the chatbot, we’ve continued to manage and update the knowledge base.

Which features do you use the most?

We connected our chatbot, Move Maeva, to our ticketing tool.

Maeva can answer users’ tier 1 questions directly. If users have a tier 2 or 3 problems, the bot automatically offers to create a ticket so that they can get the assistance they need.

Which KPIs do you monitor to manage your bot’s activity?

We analyze the following KPIs to manage our chatbot’s activity:

• number of dialogs
• salespeople’s satisfaction rate
• number of tickets created via the chatbot
• qualification of dialogs (which allows us to check that the chatbot is answering the questions correctly) 

What benefits have you observed since implementing the solution?

We’re able to monitor incidents more closely with the creation of tickets via the bot.

Before implementing the chatbot, we used several different channels (emails, tickets, and phone calls), so we weren’t able to follow up on our sellers’ requests properly.

Do you think the project is a success? Why?

We think that the project is well on its way to being a success. Our support teams already appreciate it, as do the salespeople who have used it. We’re now trying to encourage our sales network to use it more, with a video to promote the chatbot and introduce the team that manages it. We’re confident that the bot’s success will continue to grow. 

What developments and updates are you planning to improve the bot and encourage its use?

To improve our user experience, we’re thinking about adding an option to enable users to attach a file when they create a ticket via the chatbot. This feature is part of the Dydu roadmap and will give us all the information we need to solve our users’ problems. We’re also thinking about adding the dialog history to our tickets.

Finally, we’d like to set up an internal organization that involves training our teams and implementing a live chat via the chatbot so that we can provide employees with tier 2 support.

Bpifrance, the French public investment bank for entrepreneurs, wanted to relieve its IT support team while continuing to provide employees with a high level of service.

As part of its modernization and digitalization process, the company implemented an IT chatbot to handle the top 20 recurring requests from employees.

Every month, the chatbot handles more than 1600 questions related to applications and hardware: in addition to getting personalized answers, users can use the chatbot to create incident tickets or switch to a livechat operator if necessary. 

Find out how Dydu’s conversational agents help Bpifrance optimize employee relations management. 

TotalEnergies, the 3rd largest electricity and gas provider in France, has implemented a comprehensive and efficient self-care tool for customers combining conversational AI and livechat technology.

With innovation and customer satisfaction at the center of its strategy, TotalEnergies wanted to provide 24/7 assistance for its customers while optimizing the way incoming contacts were handled.

With more than 77,000 conversations per month, the TotalEnergies chatbot automates 40% of incoming enquiries : it provides customers with personalized answers and lets them switch to a livechat for cases requiring an agent’s intervention.

Find out how Dydu’s conversational agents became a key aspect of customer relations management at TotalEnergies. 

Bpifrance, a French bank, called on Dydu to deploy a conversational IT bot for employees. Hugo Parlier, ITSM and Chatbot Product Owner, looks back at the chatbot’s implementation and shares plans for future developments.

Can you describe Bpifrance and your role?

Bpifrance is a bank for entrepreneurs. We support companies at every stage of their development. We help them with financing, guarantees, innovation projects, exporting products, international expansion, etc. We also offer training, coaching, and support services.

I’m Hugo Parlier, ITSM and Chatbot Product Owner. I work in the IT department, made up of almost 1,000 people, to meet the needs of Bpifrance’s 5,600 users.

What are the current IT support challenges at BPI?

For the past 2 years, we’ve been in the process of modernising and digitalising the
IT department to provide Bpifrance employees with new tools.

What issues does your chatbot address?

Our tier 1 internal support team, dealing with incidents and requests, was under a lot of pressure, so we had to find new ways to communicate. We wanted to keep the same level of support while providing greater flexibility. A lot of our users were encountering the same problems and we thought that a chatbot could solve this issue. We analysed the top 20 recurring incidents. Our support team already had the knowledge to deal with these topics, which we then adapted to the chatbot.

Take VPN, for example. We regularly get requests for help with this complex technology, especially since the first lockdown. The chatbot has made it easy to provide users with a solution and to help them solve problems by themselves.

The same applies to authorisation requests and access rights management. We integrated a dynamic list that enables users to identify their IT correspondents and contact them directly, without going via the support team. OID (OpenID Connect, an open authentication protocol) also allows us to identify users and personalise the chatbot’s answers.

Why did you choose Dydu?

We benchmarked all the tools on the market and reviewed the existing editors. Dydu stood out for its innovation: its chatbox, integration with Teams, and, more generally, its roadmap. Dydu also hosts data in France and has a competitive offer. Since we use Teams as a collaboration platform at Bpifrance, it makes sense to integrate the chatbot into the tool, so that users always use the same interface. We’re currently working on implementing authentication in Teams.

What scope does your chatbot cover?

Alfred, our IT chatbot, covers user questions and requests related to apps and IT equipment within Bpifrance. We’ve also connected our chatbot to EasyVista, our IT ticketing tool. The chatbot can therefore provide users with tier 1 support for any computer or app problems and create an incident ticket.

How many people manage the chatbot internally and what do they do?

Three people currently manage the bot, including a tier 1 support staff member from my team. The tier 1 support team helped to integrate the chatbot and qualify its knowledge articles. Today, the support team adds and manages knowledge independently. We assist with any technical problems or to implement new developments (e.g. APIs, etc.).

We’re currently deploying a new chatbot for the short-term financing team, to assist all our salespeople in France. We’re helping the team deploy and train their chatbot until they can manage the knowledge by themselves.

After we’ve deployed 3 or 4 chatbots, we plan on reorganising our chatbot team to provide the best possible support and guidance.

How did you build your knowledge base?

The knowledge bases for our IT chatbot and sales chatbot, which is currently in production, already existed. We helped the bot managers with the syntax of their answers and showed them how to use the different features so that they can integrate the knowledge base as soon as the chatbot is deployed.

Which features do you use the most?

The features we use most are the decision trees, side panel, top knowledge, and carousel.

We use the features proactively wherever possible. For example, we’ve added an app status overview to the side panel, using an API, which also informs users of any incidents. This means that users don’t need to create incident support tickets. We also use the chatbot to highlight important information.

Which KPIs do you monitor to manage the bot’s activity?

We support the business teams by monitoring the following KPIs on a weekly and monthly basis:

• Number of dialogues (between 250 and 300 per month)
• Number of interactions
• Dialogue qualification rate
• Accuracy rate
• User rating
• Number of incidents opened via the chatbot (approx. twenty per month)

Bpifrance has an IT portal that users can use to report the same type of incidents as on the chatbot. We’ve found that users still tend to use this portal out of habit. The integration of the chatbot into Teams should resolve this issue and significantly increase the statistics. We also always check that users are satisfied.

What benefits have you observed since implementing the solution?

Our IT chatbot was officially deployed in July 2021. The project is currently in a stabilisation phase. We want to tailor it to users’ needs. It’s difficult to put a number on what we’ve gained regarding the support team’s workload. But we’ve noticed within the IT department that employees use the chatbot more for major incidents, like when a VPN is down. We instantly see a spike in the chatbot’s use, which takes some of the pressure off our internal support teams.

Do you consider the project to be a success? Why?

Overall, employees at Bpifrance are very positive about the chatbot. Users are increasingly aware of the value of getting answers immediately, rather than waiting for support to become available.

I think that the more conversations employees have with the chatbot, the more it will become a part of their day-to-day life and impact their way of working. This transformation has only just begun. When we deployed the chatbot, we carried out an internal communication campaign to raise awareness, but we believe this should happen naturally, through word-of-mouth. We also have change management teams working on the bots. Many users are used to working a certain way, so it will take time for the chatbot to become part of all the employees’ daily lives.

Some of the conversations take place outside office hours, which provides more flexible support.

How do you plan on improving the bot and encouraging its use?

This year, we will deploy 2 chatbots, one for our short-term financing department, which will cover a legal and sales scope, and another for HR. Then, we’ll launch a larger communication campaign within Bpifrance to encourage employees to use them. We want to continue to deploy Teams and to set up a live chat to better meet employee needs. We firmly believe that robots and humans can work together.

Do you have any anecdotes about the deployment of your IT chatbot?

We called our IT bot “Alfred”, after Batman’s butler, so that users would view it as their butler. When we deployed Alfred, we noticed that the question “who are you?” came up a lot, so we decided to create a knowledge article to introduce the bot to users. Alfred has his own avatar, and today IT and other departments think of him as an employee in his own right.

In the long term, we’d like to widen the chatbot’s scope beyond IT.

The French Development Agency – Agence Française de Développement (AFD) – called on Dydu to deploy two HR conversational robots within their company: an internal chatbot for AFD employees, and a recruitment chatbot for applicants. Meng-Han Chiang, Head of HR Digital Transformation at AFD, looks back at the implementation of these two chatbot projects.

Can you describe AFD and your role?

The Agence Française de Développement (AFD) funds, supports and accelerates the transition towards a fairer and more sustainable world. Climate, biodiversity, peace, education, urbanism, health, governance… Our teams carry out over 4,000 projects in the French Overseas Territories and 115 countries. We contribute to the commitment of France and French people to the Sustainable Development Goals (SDG).

AFD has a network of 85 branches and 17 regional directorates in the world. We have more than 2,600 employees and 80 different nationalities in 5 continents.

I am Meng-Han Chiang, Head of HR Digital Transformation at AFD. We conduct digital HR projects.

What problems were you trying to solve by implementing a HR chatbot and a recruitment chatbot?

We noticed that employees were asking our HR managers a lot of recurring questions, to which they were providing the same answers. We wanted to provide a tool that could answer simple and recurring questions, to help my colleagues focusing on specific cases. Our 85 branches operate in different time zones, so it was important to opt for a tool that was available 24/7.

How did the project come about?

We began to look into chatbots in 2018, the year I joined AFD. At the time, AFD’s innovation hub was launching an entrepreneurial programme which enabled us to push new projects and test our ideas on employees. We put together a small HR team to take part in the programme’s first edition. We tested different tools on user groups, such as interactive FAQs and chatbots. Following these tests, we realised that chatbots were better suited to our needs. Our executive committee then approved the project for a pilot phase. After conducting several POCs with different editors, we decided to launch our chatbot projects with Dydu.

What were the different chatbot implementation stages?

We started with the recruitment chatbot. Recruitment was our primary focus because there was a limited number of recurring questions, and we could cover a fairly large population. We receive 3,000 applications per month, so we wanted to improve our employer brand through the chatbot, and to use this digital tool as a lever for social interaction. For example, our chatbot shares job videos on applicants to promote careers at AFD.

The HR department’s scope is very wide, so we defined themes for the internal chatbot, according to the different HR areas of expertise. We first worked on appraisal interviews, for which our HR policy is the same for all employees. We then addressed internal mobility and training.

We launched both chatbots in October 2020. The recruitment chatbot on AFD’s careers site and the HR chatbot on our 3 HRIS tools: career management, training and e-training. We communicated on the launch of our internal bot with infographics and chatbot presentation videos via our HR newsletter.

Why did you choose Dydu?

We met several chatbot builders and chose Dydu because the solution offered the most advanced features. In particular analytics, anonymous dialogues and the possibility of creating decision trees allow us to provide an answer according to the employee’s situation.

With Dydu, we can manage two separate chatbots in one environment. In the Dydu back office, I can manage our recruitment chatbot’s knowledge articles, dialogues and analytics internally. The other editors did not offer this type of all-in-one functionality.

What do you want to achieve?

With this chatbot project, we want to make information easily accessible to everyone, lighten our HR management teams’ workload and strengthen AFD’s employer brand.

We set ourselves goals and indicators to reach within 1 year:

• Identify the most popular 20% of knowledge articles in each HR field to answer 80% of recurring questions.
• Provide information at the right time.
• Free up time for our HR teams, so that they can focus on more added-value tasks.
• Achieve an 80% user satisfaction rate with the answers provided by the bot.

Overall, we have achieved these goals. We are continuing to feed the bot and hope that it will free up time for HR employees in the long run, when it has become an expert on all HR topics and deployed on all our HRIS platforms.

What benefits have you seen since implementing the solution?

Since implementing our recruitment chatbot, we can answer most applicants’ questions, whereas we struggled to do so without the tool. We can only gain from this new service. It also helps develop our employer brand. Our recruitment chatbot has 1,200 dialogues per month with a 96% qualification rate of interactions. A lot of our dialogues, therefore, have an excellent accuracy rate, and we are delighted.

Our internal chatbot is also a new internal support service. There are on average 200 dialogues per month, with an 86% accuracy rate. The questions are logically tied to major HR events, such as appraisals, internal mobility, etc. The rest of the year, the questions are often about training.

How do you manage the project?

Today, in run mode, I’m the only chatbot administrator. I spend an hour per week reading the dialogues to identify any trends (it’s an excellent tool to know what is preoccupying employees), improve the bots’ understanding and identify any new knowledge articles to create. Once every two months, we have a meeting with experts from our different HR teams to review the most frequently asked questions. We then decide together if we need to create a new knowledge article and which ones to highlight in the Top 3.

Which KPIs do you monitor to manage the bot’s activity?

We monitor several indicators in the Dydu back office:

• Number of visitors per month
• Percentage of visitors with a dialogue
• Number of dialogues per month
• Qualification rate per interaction
• Number of knowledge articles that cover 80% of user questions
• Customer satisfaction for the knowledge articles highlighted in the Top 3. If satisfaction is not sufficient (below 80%), we try to improve the knowledge article’s phrasing and add additional information to our answers.
• Contact channel: this allows us to know which tools applicants use to communicate with the recruitment bot (Windows, Android, iPhone, or Mac) and to improve our user experience.
• Platforms of use: this allows us to know which HRIS platforms employees use to interact with the internal HR bot.

I check these indicators regularly and share them with our HR experts during our steering committee.

What prospects or developments do you have planned for the project?

We plan on deploying the internal chatbot on other HR platforms and to integrate all HR knowledge. We also want to translate the knowledge base into English and possibly Spanish. With regards to translation, we will identify the most used knowledge articles and focus on them first.

Our IT department has also provided support from the start of this adventure and is very interested in the potential of a conversational robot.

In 2019, PwC France, one of the Big 4 specialising in consulting, audit and legal services, called on Dydu to create Alex, a “concierge” chatbot for its 6,000 employees.

KEY FIGURES

• 2,800 knowledge articles around 10 subjects
• 15,000 interactions per month with a 97% comprehension rate
• 4,000 dialogues per month, of which ​​85% are successful

[Only in french]

PwC’s Human Resources chatbot currently includes 2,800 knowledge articles, based on ten recurring subject areas (Human Resources, HelpDesk, work premises, events, corporate strategy, etc.). With 15,000 questions per month and a comprehension level of 97%, the chatbot has become a strategic tool for PwC’s internal support management. It simplifies employees’ working life and helps to onboard new recruits.

During our client meeting in March 2022, Nelly Fercoq, the PwC chatbot project manager, revealed that the chatbot had been particularly useful when implementing remote working for everyone. PwC recorded a 4% increase in the number of chatbot requests.

The tool’s managing team regularly promotes the bot. They also work continuously on improving and updating its knowledge base. “We read the conversations every day. This enables us to see the sentences that haven’t been understood correctly or any little bugs, so that we can correct them and continuously improve the bot. We add phrases and we create knowledge articles in accordance with current events. Any that are no longer relevant are deleted,” Nelly Fercoq explained.

PwC is planning on interconnecting the bot with their HRIS to perform more transactional tasks, such as checking an employees’ leave balance or requesting leave via the bot. They also want to deploy the chatbot on more internal channels, such as Teams.

SNCF Connect and Dydu have been working together for more than 10 years. The first chatbot, created at the time of Voyages SNCF, took the form of a Helpbox available on the website and passenger app.

KEY FIGURES:

• 10 million questions handled by the bot in 2021
• 6 languages available
• 5 times more requests made to the SNCF Connect bot since its launch and the merging of all SNCF bots
• 95% of questions asked to the bot have been answered since the migration

In addition to the Helpbox, SNCF Connect deployed their own bot in 2017, Oui Bot, which answered travellers’ questions about buying and booking tickets.

In anticipation of the OUI.sncf website migration to SNCF Connect, they decided to use one chatbot instead of two. “We wanted an easily identifiable assistant everywhere, all the time. We also wanted to end the OUI Bot (SNCF bot) / Helpbox (Dydu bot) split, so that customers would only see one bot,” Franck Gbenou, Head of Passenger Information and Customer Support at SNCF Connect, commented. The aim was therefore to make the customer experience smoother and more efficient, without losing the richness of what had already been put in place with Dydu.

Today, with the merging of the two bots, the SNCF bot routes requests with the Dydu API, depending on the user’s question:

• The SNCF bot handles transactional or booking queries.
• The Dydu bot handles customer support and follow-up requests.
• If the question is not covered by the SNCF or Dydu knowledge base, the Dydu bot answers.

SNCF Connect has therefore succeeded in interfacing these technologies in a smooth and intelligent manner with a single dialogue box, and the possibility of switching to an operator via live chat.

Thanks to a more integrated and visible journey, the number of bot requests has been multiplied by five since the merging of the two.

Read all our client testimonials.

AFD, the French Development Agency, who funds social and environmental impact projects, wanted to leverage conversational AI to modernize its employer brand and become more efficient.

For AFD’s HR digital transformation unit, the purpose of the project was twofold: first, to provide better support for employees on HR topics, and second, to offer assistance to future talent in the application process.

The two Dydu chatbots implemented now handle over 40,000 questions every month, thus relieving the HR department of recurring requests – a resounding success for the team !

Find out how Dydu’s conversational agents became a key aspect of employee relations management at AFD. 

PwC France and Maghreb called on dydu in 2018 to create Alex, a conversational robot for employees. Nelly Fercoq, a chatbot project manager at PwC France and Maghreb, looks back at the implementation of the project.

• Can you describe your company and department ?

PwC provides consulting, auditing, and legal and tax expertise, with the strategic ambition of reconciling businesses, the economy and society. 6,000 employees currently work for PwC in France and North Africa.

I’ve been a member of the IT department for 13 years and have been a chatbot project manager for the last two. I’m responsible for managing and developing the chatbot with another colleague, who is in charge of the knowledge base, reviewing the dialogues and monitoring the analytics and KPIs.

Problems and solutions

• What problems does the implementation of a concierge chatbot seek to solve ?

Our project was born after interviewing a sample of employees. Our goal was to identify any obstacles in their day-to-day life. This survey revealed in particular that the intranet was seen as dense and complex, and that it was difficult to carry out a search.

We also wanted to help newcomers become autonomous more quickly. Indeed, nearly 2,000 new employees join us each year. The chatbot enables them to understand our whole ecosystem in a simple and quick manner.

To make everyday life easier for employees and to facilitate the onboarding of new recruits, we wanted to implement a tool that could instantly answer all their questions. We decided to deploy an internal chatbot and the project began as a pilot with a handful of employees. We started to feed the bot with knowledge from our intranet. We then launched it throughout the organisation in January 2019.

Project implementation

• What scope and use cases does your chatbot cover ?

The chatbot is internal. It acts as a concierge and helps employees with various day-to-day issues: HR, IT, office space, events, corporate strategy, etc. For example, it can provide information about how to access payslips, fill out expense reports, request leave, etc. The topics vary but the questions mainly relate to HR and IT.

• What channels is the chatbot available on ?

The chatbot can be accessed on the internet via an icon on our home page. It is also installed directly on desktops and mobile phones with a web app.

• What were the different project implementation stages ?

Design – We ran workshops with representatives from each department, with a mix of junior, experienced and senior profiles. We asked them to ask any questions they wanted a quick and clear answer for. This helped us determine which themes to include and the content to create. We also conducted an employee survey to choose the bot’s name; Alex was the most popular of 4 choices.

Building the Knowledge Base – Based on the questions we identified, we collected the official answers from the department representatives. There were 15 of them, divided up by theme. We built a first list of knowledge articles with their answers.

Launch – We communicated a lot when we launched the bot, using all of our internal communication platforms: internal newsletter, cafeteria screens on every level, annual convention, etc. We also presented the solution via a stand during our innovation days, and we always present the chatbot to newcomers during our onboarding programme.

Updates – The number of questions asked is gradually increasing. We add new knowledge articles to a Google Sheet, along with our department representatives. We hold regular calls to clean the knowledge base by deleting or updating any obsolete articles. Every month, we look at the most asked questions with the representatives. My team centralises everything and carries out any updates in the bot management system (BMS) for all the department reps.

• How did the dydu teams support you in setting up and monitoring the project ?

We assessed various existing solutions. In the end, we chose dydu because of the quality of the BMS, that can be used with many functions by a non-developer, the transparency of pricing, client references and the possibility for on premise of SaaS hosting.

During the implementation phase, dydu’s teams trained us to use the BMS. The aim was to become as autonomous as possible to build knowledge articles and use important functionalities, such as decision trees and variables.

Today, we are fully autonomous. We do however still benefit from dydu support and we have a dedicated customer success manager (CSM) for any questions, update requests or bugs. I’ve had three contacts at dydu since I started working on this project, and I’ve always found them quick to respond. They’re always ready to solve problems and find workarounds when on very short deadlines.

Benefits and good practices

• What benefits have you seen since implementing the solution ?

When we all began working remotely full time, bot requests increased by 4%. Employees broached new topics as a result of the crisis, which brought about several changes: adjusted working hours for some departments, curfew, guidelines for coming on site, etc.

We updated or adapted a lot of knowledge articles to the situation. We added answers to questions such as, “Who is the Covid rep?”, “How can I get psychological support during the health crisis?”, “Where can I find a permission form if I have to travel for business reasons during curfew?”, or “What are useful Coronavirus contacts?”.

We try to guide our employees as best as possible, so that they can easily find answers during these difficult times. Since the beginning of the year, there are on average 4,000 dialogues per month, with more than 15,500 interactions. More than 300 knowledge articles are used, 136 of which represent 80% of the bot use.

• Which KPIs do you monitor to steer the bot’s activity ?

We look at the number of dialogues and the number of visitors per day to ensure that employees are still using the bot. When these numbers go down, we communicate about the bot again. We also check the satisfaction rate and review the dialogues between the bot and employees on a daily basis, to make any necessary corrections. We carefully monitor all the KPIs and check the themes to understand employee concerns.

Today, we consider that we have reached our goal. The bot is used regularly, and the satisfaction rate is high. As for any innovative tool, we regularly interview a sample of around 100 employees with different profiles. This enables us to check that our tools still meet their expectations and to identify any new needs.

• What advice or good practices do you have about the relevance , use and management of the bot ?

If the chatbot cannot answer an employee’s questions several times, they are unlikely to return. I therefore recommend having all the necessary resources internally, to offer the best possible service. This involves reading the dialogues, adding matches, and creating new knowledge articles. We pay particular attention to the quality of the answers we provide, to ensure that they are precise and current.

Alex also has its own email address. We make a note of the name of any employees with unanswered questions and reply a little later by email. This is particularly appreciated internally.

• Why would you recommend the dydu solution ?

Using the BMS to add knowledge articles is very clear. The tool is easy to use because it is very intuitive. Everything is simple, clear, and quick for end-users too. All our contacts are accessible and friendly. Our CSM is very responsive. We know that we can contact her if we have a problem. Dydu immediately takes our issues on board and finds a solution. I also really appreciated the exchange workshops with other dydu clients about developments we would like to see for the software; very few publishers do this.

Next steps …

• What prospects or developments do you have planned for the projetc ?

The next step will be to interface the chatbot with some of our HR tools so that it can provide fully personalised answers. The chatbot is already connected to our authentication system, which means Alex can address employees by their first name. We’re also working on personalising the answers according to the employee’s site location, ranking and service line.

Harmonie Mutuelle deployed a chatbot in 2017 to help its employees navigate the company’s digital environment. Marc Pasquet, Head of User Support Manager, and Bruno Walter, User Support Project Manager, look back at the implementation of this conversational robot and its benefits.

To download a PDF of the client case study, click here.

• Can you describe your company and department in a few words?

Harmonie Mutuelle is a health and life insurance company. We are part of the user support department, which handles and resolves employee incidents with the digital environment and escalates them to a specialist team, if necessary.

Problem(s) and Solution(s)

• What tools are available for managing employee IT requests?

ITSM EasyVista is an end-to-end tool that handles incidents and requests, entered directly by employees or support teams following a telephone or email exchange.

What problem(s) are you looking to solve by implementing a Chatbot?

Our main reasons for deploying a chatbot were to answer questions at any time, encourage employee autonomy and reduce the support teams’ workload. Simple requests entered into the chatbot are sent directly to the competent team, along with all any information required. This allows to bypass the first level of support to qualify the request, making the process faster and less resource-consuming.

The chatbot and its services are currently available for Harmonie Mutuelle’s internal employees and contractors, which represents about 6,000 people.

The Project

• What is the chatbot’s scope?

The chatbot primarily focuses on our digital environment, access to applications, printing services, videoconference procedures, etc. It frequently refers users to documents or pages on the Intranet. The chatbot can also open a user file, such as for password reset requests.

• How many people manage the chatbot internally and what do they do?

The knowledge base is currently administered by two people (IT knowledge scope) half a day per week, with certain periods requiring more or less responsiveness.

Can the chatbot escalate issues to specialist teams?

Yes, the chatbot creates and escalates tickets to the right support team, the Webservice we use is recognised by the ITSM tool and routed accordingly.

• What major changes have been made to the bot since its implementation?

The first version of the chatbot answered simple questions about our digital environment. The second one allowed requests and incidents to be sent directly via Webservices to the ITSM tool and specialised support teams. The chatbot currently has 480 knowledge articles, 55 of which represent 80% of the questions asked.

• What are its main functionalities?

We mainly use decision trees, data retrieval in the form of variables and webservice.

Benefits

• What benefits have you seen since implementing the solution?

Implementing the chatbot has enabled us to offer users an additional service, available 24/7, when the helpdesk is overwhelmed. The productivity gains (solutions provided by the chatbot without any helpdesk involvement) represent 8% less helpdesk requests, and up to 13.3% during the first lockdown.

Which KPIs do you monitor to steer the bot’s activity?

We use the following analytics: number of dialogues, number of interactions, quality of dialogues, most used knowledge articles, use of webservices (incidents and requests opened via the chatbot).  

• Do you consider this project to be a success? Why?

This project is a success because it fulfils the role it was set up for. The chatbot holds 500 dialogues per month on average (approximately 1,350 interactions), with an 80% comprehension rate. Now, we need to encourage even more people to use it. We need to take advantage of the opportunities for use with maximum benefit (simplicity and efficiency). We will also communicate more about the bot via video.

• Do you think the chatbot was even more useful during lockdown?

Yes, the chatbot was very useful during lockdown. It enabled us to answer a lot of user questions about remote working tools, as well as the usual requests, at a time when the helpdesk was under a lot of pressure.

 Why would you recommend the dydu solution?

The dydu solution allows to implement a quality chatbot that fits perfectly into your intranet or website, for example.

Next Steps

• Are you planning any changes or updates to further improve the chatbot? And the management of IT requests in general?

Enabling users to fill out forms in the chatbot to create more complex service requests via webservice. We also plan on making it possible to book appointments with the helpdesk via the chatbot.

Feel free to download a PDF of the client case study.

Dydu has signed three new accounts in the e-commerce, automotive and distribution industries.

Logista France

Strator has implemented a chatbot for tobacconists and newsagents using the dydu solution. Strator (part of Logista France) is the leader in POS terminals, tools dedicated to the sales and management of shopkeepers’ businesses. Their brand new chatbot answers retailers’ questions about how to use Strator: stock management, inventories, dematerialised products, dashboards, etc. They have deployed the chatbot on their online Help website, that can be accessed via Strator terminals, PCs and mobile phones. The conversational robot, an expert in Strator’s software, helps save time by answering user questions immediately, 24/7. 

Renault

The Renault Group is deploying a new internal chatbot, using dydu’s technology. The car manufacturer already has a HR and IT bot and has just implemented a new selfcare tool to improve the experience of the 11,000 employees at their Technocentre (R&D hub), and to relieve the busiest departments. The conversational robot, that is still in a testing phase, answers questions about practical services (company rules and regulations, entry passes, lost and found, etc.), on-site services (media library, works council, concierge service, taxis, etc.) and events (activities, exhibitions).

Cdiscount

Cdiscount used dydu’s technology to create an FAQ tool on their website. It provides visitors with immediate answers to their most frequently asked questions, on topics such as tracking orders, returns and refunds, breakdowns and warranties, customer account management or the marketplace. This new selfcare tool handles more than 20,000 interactions per day, which saves customer services a significant amount of time and increases productivity.

S2E deployed and now manages conversational robots for 4 major banking and insurance companies, in order to answer their recurring customer questions about employee savings. Stéphane Palvadeau, the Chatbot Administrator and Call Centre Supervisor at S2E, shares his experience managing these 4 bots using dydu’s software.

Who is S2E (Service Epargne Entreprises)?

S2E was born in 2006, grouping together 4 major players in employee savings plans, to support their development. S2E manages the administrative processing of employee savings schemes for 70,000 companies and close to 3 million employees.

S2E works with dydu’s solution to manage and monitor 4 chatbots aimed at employees with an employee savings scheme.

Stéphane Palvadeau, who are you and what is your role?

I joined S2E in October 2018 to finalise the employee savings business aspects of the chatbot, drawing on the expertise I previously acquired working in account and business administration. With regards to the shared management of the 4 bots, I am currently responsible for publishing content, analysing the dialogues, and reporting to the custodian account keepers.

More generally, my role is to continually improve the customer experience by identifying how the online customer journey can be improved, as well as any navigation issues. I also liaise with the custodians’ call centres to ensure that the chatbots and operators are providing consistent information. I begin most days by analysing the previous day’s dialogues and I optimising the knowledge base accordingly. The rest of my day is then spent improving and enriching content, liaising with call centres, producing reports, requesting updates, etc.

What issues do company pension plan bots respond to?

The bots specialise in employee savings in the banking and insurance industry. They answer generic user questions and help navigate the custodian’s website. The aim is to relieve call centres of requests with low added-value. One of the custodian account keepers already had their own chatbot. So, when S2E was contacted to deploy the same service on other custodian websites, we merge the content and management for several clients with the same issues. A large part of the knowledge base is therefore based on a regulatory framework, which can be shared among our clients. Similarly, the insights gained from analysing the dialogues can often be shared too.

In order to share this knowledge, we therefore first had to define nearly 200 common knowledge articles in collaboration with the custodian account keepers. This phase took about a year. During this time, we also identified a dozen differentiating topics to be managed outside of the collective framework, by the custodian’s individual bot. Some content is customised for each of our 4 clients, based on their schemes’ specificities or their communication strategy.

This enables us to offer each of our custodian clients a personalised and scalable service, while ensuring a common knowledge base, particularly on the regulatory side. In order to target the best period for each client, the final deployment phase for the 4 chatbots took place between October 2019 and February 2020.

How does this project work?

The 4 bots benefit from the same overall services, including the knowledge base administration, bot management or the production of reports. Content is updated and added by S2E, who then informs the custodians. This mode of operating ensures maximum agility when using the product.

The number of potential users for each custodian varies greatly, so the volume of dialogues observed on each bot is naturally heterogenous. The most used chatbot carries out more than 5,000 conversations per month. But whatever the numbers, the volumes observed justify this chatbot self-care approach.

Drawing on user questions to the 4 bots to update and add to the shared knowledge articles, also ensures that all our clients’ users share the same experience. This is a particularly virtuous model that has a positive impact on the bots’ performance.

How do you measure the bots’ performance?

We send a monthly custom report to each of our clients, based on indicators made available by dydu and the 30 most used shared knowledge articles. The bots have now reached a level of maturity that will enable us to make these reports quarterly.

Once a year, we send the custodians a global report, to provide them with an overview of the activity, and identify areas for improvement and development. The initial objective was to maintain a high qualification rate for the interactions. We therefore monitor this indicator carefully. We continually monitor the knowledge articles, in a centralised manner, and improve them accordingly, which explains why the 4 bots have interaction qualification rates that are over 98%, compared to a 93%* industry average.

We also, of course, monitor and regularly seek ways to improve the level of other customer satisfaction indicators. User feedback and reasons for dissatisfaction are therefore always analysed to ensure that the content published is continually improved.

How have the bots been used during the health crisis?

The first lockdown didn’t change user behaviour, but we have observed a change in questions, particularly with regard to stock market trends and the impact of the fall in prices on employee savings plans. We were able to quickly create new knowledge articles and update others, in order to answer these questions. Dydu’s product makes it possible for us to react quickly, because we can adapt our answer model to the situation in a very short amount of time.

What are your chatbot administrator best practices?

When it comes to employee savings plans, certain key terms that appear in user questions, such as “payment” or “transfer”, can have different meanings depending on the context or even the time of year.

To overcome this issue, I avoid updating knowledge articles with ambiguous phrases and allow the chatbot to suggest alternative phrasings instead (suggestion of several knowledge articles), so as to not provide a potentially irrelevant direct answer. In my experience, users are able to select the knowledge article that reflects their situation. In some cases, I think that indirect matching should be favoured over direct matching. If I notice that a knowledge article’s phrase is misleading users, I delete it.

It is also very important, in my opinion, to take great care when adding phrases to knowledge articles. Indeed, during my daily dialogue analysis, I add to the knowledge articles if I think a user question is likely to be asked again. This means that the bot can now understand a question that it misunderstood the day before. Before adding these new phrases, I correct spelling mistakes, delete any irrelevant contextual elements, and chose the most suitable matching groups. This process has enabled to gradually improve the matching rate over time.

What developments are you planning for the future?

Some of our clients want to expand their service with LiveChat, so that users can continue a conversation they started with the chatbot, with an agent. Dydu’s solution allows to define the triggering of knowledge articles based on behavioural targeting rules (time of inactivity, determined number of unsatisfactory answers, etc.).

We would also like to add an onboarding step before users can enter a question, in order to contextualise the relationship and explain the chatbot good practices. This will avoid any confusion between the chatbot and LiveChat.

The more personalised the information, the more relevant and appreciated by users it is. With this in mind, we plan on providing the chatbot with custom information, so that it can publish individualised content. This will meet the specificities of some of our clients’ operations that are not currently covered by the shared knowledge.

And finally, why not deploy the chatbot on a mobile app, which has proved hugely successful over the past few years?

* dydu data for the entire year 2020, based on more than 15 customer relationship chatbots in the banking/insurance industry.

Pixelraise partnered up with dydu to implement a gamification layer to the Jules chatbot, a homework helper for secondary school students, deployed by the CNED with dydu’s solution. Moezz Sedkaoui, founder and managing director of Pixelraise, looks back at this collaboration and explains the advantages and mechanics of a chatbot-gamification association.

Can you describe your company and role in a few words?

Pixelraise is an IT consultancy firm and web and mobile developer, with an entity in France and Switzerland (WebGoeSon). Our 25 employees develop custom IT solutions, e-commerce sites, native and cross platform mobile apps, and CMS-based sites. Pixelraise has a large client portfolio: Merck, EMS Medical, Produits Dentaires, Carglass Suisse, the French Ministry of Education, Sword, Hermes, Plezi, etc.

How was the Pixelraise and dydu partnership born?

Dydu put out a call for tender looking for a partner that could quickly gain the skills required for their chatbot technology and, above all, enhance the existing Jules chatbot app through gamification. A shared contact brought the decision makers from both companies together, giving rise to a great collaboration and successful project over a 10 month period.

Can you describe the role Pixelraise played in the CNED Jules project?

Pixelraise was responsible for developing the gamification layer (excluding UX and design), with 3 main objectives:

• develop digital uses for the “Devoirs faits” programme (personalised support and help with homework for secondary school students)
• reward effort with prizes, to encourage students to use the app 
• motivate users to play regularly

What does the Jules chatbot currently offer in terms of gamification?

The gameplay we developed enables the user to win flames and gifts. By earning flames, the user can move to the next level, and is rewarded with a gift at every level. These gifts include outfits for the user’s avatar or backgrounds to personalise the environment.  

What were the different steps to implement gamification in the chatbot? How does this combination work technically?

The first step was to understand and perfectly master the dydu component (Assembly.js), the chatbot’s engine that contains the knowledge base. Then, we developed all the APIs to interact with user actions, to reward and encourage them to learn while playing. Technically, the interactions between the quizzes and dydu’s chatbot are done via a function call in Js, using JavaScript Window Option. Finally, the app authentication is based on an API managing the creation of user accounts (student and teacher profiles – web and mobile) and the connection.

What are the main game mechanics and keys to a successful game or challenge?

Gameplay is the first aspect, including playability and the rules of the game. The player should be immediately immersed in the action with an adapted design. The second important aspect is finding the right balance between constraints and possible actions; too many rules make the game difficult to play and too much freedom leads users to drop out, without any retention. The last aspect is the pleasure or fun of playing, based on innovation, whether technical, design or narrative. 

The keys to a successful game can be summed up in 5 points:

• choose a unique gameplay; use well-known game mechanics that the majority of gamers are familiar with, and add customisation to make the game unique,
• opt for an innovative design; UX is very important, and it is essential to carefully design the game so as to offer a unique and coherent experience from the very first session. The game should be fun. If the user’s first impression is negative, there is little chance they will play again (and they will probably uninstall the app),
• work on retention; for example, create addiction with random rewards,
• optimise the response time; make sure that your game takes no more than 2.5 seconds to load, in order to reduce your abandonment / unsubscribe rate, 
• market the game with a suitable budget to make it visible among the many games available, in a very competitive market.

What other web and mobile developments have you carried out for the Jules project? (application, authentication interfaces, etc.)

We developed an external back-office on the Java / Springboot / BD Postgress / J2ee stack, in order to manage certain items such as the flames, quizzes, levels, outfits, backgrounds, etc. It also contains user information.

Do you think there will be other opportunities for dydu and Pixelraise to work together, beyond the Jules project?

Absolutely, we are complementary and perfectly capable of integrating dydu’s chatbot solutions. We recommend dydu to our clients as an expert partner in chatbot solutions.

In your opinion, how is UX design evolving in web projects?

Over the past 10 years, the user experience (UX Design) has taken a new direction, and users are now at the heart of projects (User Centric). UX design is not just about the design aspect of the interface. It also involves a reflection on the standard user journey in order to meet user expectations. The science behind the user experience is central to all web projects because more then 7 out of 10 internet users will not return to a website if their first experience does not live up to their expectations.
P.S.: by the way, I’ll soon be announcing the release of our UX addict websites

How do you think Artificial Intelligence will impact UX?

I believe that artificial intelligence will revolutionise the user experience. Artificial intelligence and user experience are closely linked in today’s world, where we are now facing a turning point in a new era including big data, voice search, chatbots, etc.

Real AI allows a degree of customisation that is hard to find on other more traditional channels. Within the context of voice search in particular, AI makes it possible to introduce greater personalisation, fluidity, and intuitiveness, all of which greatly benefit the UX. It is important that UX designers offer three options at the most per interaction, because the user will struggle to retain more than three.