AI for everyone
Bouygues Telecom achieves rapid innovation by scaling AI on AWS
Put AI to work for customer service in telecom Subscribe to AI insights
Cropped shot of a person using mobile app device on smartphone

Bouygues Telecom’s innovations set it apart among French communications service providers (CSPs). From 1996, when it offered the country’s first Short Message Service (SMS), to today, when it is helping lead the rollout of a nationwide fixed and mobile 5G infrastructure, it has been first to the marketplace with many technological advancements.

Bouygues Telecom also focuses on enabling innovation within its internal organization, most recently by accelerating use of AI across the enterprise. Toward this end, it teamed with IBM® Consulting on a multiphase initiative to empower all business and IT functions to create, develop and deploy their own cloud-native AI apps.

“AI was starting to appear everywhere, and we didn’t want to miss the train. Also, as a CSP, we have access to loads of interesting, valuable data,” explains Romain Dutot, IT Innovation Leader with Bouygues Telecom. “We knew that if we didn’t use data and AI to improve our business, in a few years our competitors would do it and seize the advantage.”

By helping teams infuse their processes and operations with AI capabilities, the company could more effectively generate actionable insights needed to improve decision-making, accelerate workflows, increase efficiencies and launch new services.

Solution components

IBM® Consulting

IBM Consulting for AI at Scale

AWS Consulting Services

IBM Enterprise Design Thinking®

IBM Garage™

Faster Go-Live

 

Scaled 4 cloud-native AI application from concept to production in 4 months.

Faster Deployments

 

Deployed a new AI app to triage B2B leads, significantly saving costs within 2 weeks.

AI is no longer an experimental subject for us. Everyone is now AI ready, and that’s a real game changer for us. Romain Dutot IT Innovation Leader Bouygues Telecom
From AI ideation to production

Bouygues Telecom had already experimented with AI/machine learning (ML) models by developing a few simple, one-off apps, such as for chatbots and text analysis. Yet the proof-of-concept (POC) approach used to develop these models was unaligned with corporate IT standards and couldn’t support full production. In addition, many line-of-business (LOB) and IT managers had trouble envisioning how to best use AI and hesitated to take the time to help develop solutions.

The company needed to effectively involve managers in brainstorming viable AI use cases. It also required a uniform, integrated approach to scaling AI solutions beyond experimentation. In other words, if Bouygues Telecom wanted to drive enterprise innovation, it had to fully democratize use of AI.

Yet the company lacked in-depth AI knowledge and skill sets. “We knew we had to go fast and strong to build a global AI offering for our company. However, AI is a very specific domain that requires not only technical but also management, ideation and design skills,” says Dutot.

To fast-track AI adoption, Bouygues Telecom sought a consultancy company with deep, multivendor AI expertise and experience. The company had a complex IT landscape, with numerous data systems and governance requirements. The consultants would need to support Bouygues Telecom’s evolving multicloud strategy, which gives teams flexibility to select cloud and AI providers based on their departmental and application needs.

Sustainable, affordable AI on AWS

Bouygues Telecom chose IBM. A leader in ethical AI, IBM embraces an open-technology, collaborative approach to digital transformation and has highly experienced data scientists. In addition, IBM provides the IBM Consulting for AI at Scale services offering, a comprehensive consult-to-operate approach for effectively scaling AI enterprise-wide. The offering guides companies on their AI journeys by delivering the framework, methods, assets and technology for creating and scaling AI solutions, from minimal viable product (MVP) to production, with rapid time to value.

Integrating AI design, ethics, data science, architecture and engineering, the AI at Scale offering helps businesses consistently integrate and scale AI/ML PoC and other tested models into production and then continually revise and optimize those models over time. It also includes components to help businesses evolve their AI culture and governance. To structure this project, the team followed the IBM Garage™ Methodology — proven practices that guide a company through designing, building and scaling solutions for end-to-end transformation. IBM co-designed custom data and AI reference architecture covering multiple cloud scenarios that can be extended to all AI and data projects across the Amazon Web Services (AWS) cloud as well as other cloud/on-premises platforms.

The initiative to democratize use of AI across Bouygues Telecom began with an AI Boost program led by the corporate Innovation team, which was tasked with proving the value of AI by creating nine MVPs of cloud-native AI apps. It also provided training to fuel an enterprise-wide cultural transformation around AI and data.

To help generate ideas for the MVPs, Bouygues Telecom participated in several IBM Enterprise Design Thinking® Workshops where consultants held brainstorming sessions with the Innovation team and several LOB managers.

“Imagining what you can do with AI can be quite difficult and requires specific skills and methods. The IBM Garage approach really helped accelerate the ideation process,” says Dutot. IBM consultants also assisted in rapidly building the MVPs following agile principles.

Once the Innovation team successfully completed the AI Boost project, the company was ready to move beyond experimentation. At this point, the IT Innovation Department stepped up to operationalize and scale AI capabilities across the organization, including setting AI-at-scale standards for the company. A small team that tests and scales new technologies, the department could have chosen to work with a different consulting company, but it wanted to keep the momentum going with IBM.

“IBM did really great work during the AI Boost program. We wanted to bring some of the MVPs — and also some new ones — to scale in just a few months, but we didn't want to start from the beginning. We wanted to accelerate the process, and IBM was the perfect partner,” says Dutot.

Turning data into insights

The IT Innovation team established several objectives to enable successful AI transformation:

  • Establish key capabilities required for enterprise-grade AI
  • Quickly experiment and pilot AI solutions to drive business value creation at scale
  • Accelerate time-to-value with AI while minimizing operational risk
  • Build an open, scalable, cost-efficient, security-rich infrastructure
  • Upskill AI talent across both IT and business teams

With the AI at Scale framework and the IBM Garage approach, the team successfully codesigned and built an AI platform that enables rapid training, testing and deployment of AI and data services. Featuring a customized reference architecture integrated with best-in-class external services, the platform supports standardized IT processes, a variety of data systems and multiple cloud deployment scenarios.

Next, the team developed enterprise-grade cloud-native AI solutions for two of the nine MVPs and two new projects on the platform. They also identified five previously developed AI solutions that could potentially be migrated to the platform.

The first AI application Bouygues Telecom deployed at scale helps the Financial Department operate with greater efficiency by automatically detecting inconsistencies in system inputs. Previously, department employees regularly used manual processes to review accounting processes. During these reviews, they often detected numerous double billings sent to business associates or double payments received. Now, employees rely on the invoice validation app to automatically find daily accounting discrepancies.

“Bouygues Telecom is a large company with multiple subcontractors, and our billing is quite complicated. Detecting inconsistencies is important but time consuming,” says Dutot. “Now, the AI app tells employees which invoices look strange. That’s a direct time gain for our employees.”

Another AI application rapidly triages and qualifies information submitted by potential business contacts using an online form on the Bouygues Telecom website. The form is meant to attract new B2B sales leads. However, nonbusiness prospects often use the form to submit requests irrelevant to the B2B customer center. Using the new AI app, the qualification call center can automatically sort out irrelevant information, speeding lead generation for sales teams and dramatically reducing costs for Bouygues Telecom.

Faster time to value

Working with IBM, Bouygues Telecom scaled its first four cloud-native AI apps from concept to production in just four months. The apps rapidly delivered value. For instance, within two weeks of being deployed, the AI app for triaging incoming B2B leads significantly reduced the associated costs and time spent on the effort.

“Four months is an extremely short time period for us, and it is for a lot of companies. Time to value is critical to generating business value with AI projects,” says Dutot.

Dutot also found that the AI at Scale framework provided his small team with much-needed flexibility. “The framework helped us understand the amount of work that it takes to scale AI and the steps we needed to take,” he says. “We chose to not take all the work on at once but rather go step by step, depending on Bouygues Telecom priorities.”

With the new AI platform on AWS, Bouygues Telecom can not only develop POC models and scale them into production faster but also minimize costs and risks. In addition, it can enable data scientists to work with greater efficiency, purpose and satisfaction by allowing them to spend more time on complex, high-value AI projects rather than launching standalone, low-value solutions.

AI-fueled innovation

The quest to democratize AI has just started at Bouygues Telecom, Dutot emphasizes. Now that the AI platform is operational, the focus shifts to continuing to facilitate cultural transformation and upskilling AI talent across departments, including through training, codesign and codelivery. New users to the platform are given access to AI tools relevant to their roles and assisted in learning how to effectively use them. The team also plans to identify more use cases, improve the AI/ML models and build a robust MLOps framework to standardize at-scale processes.

That said, the work accomplished to date represents a huge leap forward toward helping Bouygues Telecom gain competitive advantage through enterprise AI adoption. Now, nearly any team or department can optimize virtually any business application and work process using AI. Bouygues Telecom is also better positioned to capitalize on its vast data stores and continue rolling out a multicloud strategy.

“AI is no longer an experimental subject for us,” says Dutot. “Everyone is now AI ready, and that’s a real game changer for us.”

Bouygues Telecom logo
About Bouygues Telecom

Founded in 1996, Bouygues Telecom (link resides outside ibm.com) is France’s third-largest B2B provider of services for mobile telephony, internet and IPTV services. It strives to bring people closer together by providing a high-quality fixed and mobile network, innovating in a constantly changing market and fostering collaboration. Part of the Bouygues Group, the company serves 25.3 million customers, operates 500 stores and employs 9,550 people.

Next up:
Read the PDF Technology and business newsletters

Get insights into emerging trends and technology innovations.

Subscribe
Reckitt

Building the factory of the future today.

Read the case study
Autostrade per l’Italia

A singular focus on the road ahead.

Read the case study
Legal

© Copyright IBM Corporation 2022. IBM Corporation, IBM Consulting, New Orchard Road, Armonk, NY 10504

Produced in the United States of America, February 2022.

IBM, the IBM logo, ibm.com, Cloudant, Enterprise Design Thinking, and IBM Garage are trademarks of International Business Machines Corp., registered in many jurisdictions worldwide. Other product and service names might be trademarks of IBM or other companies. A current list of IBM trademarks is available on the web at https://www.ibm.com/legal/copyright-trademark.

This document is current as of the initial date of publication and may be changed by IBM at any time. Not all offerings are available in every country in which IBM operates.

The performance data and client examples cited are presented for illustrative purposes only. Actual performance results may vary depending on specific configurations and operating conditions. THE INFORMATION IN THIS DOCUMENT IS PROVIDED “AS IS” WITHOUT ANY WARRANTY, EXPRESS OR IMPLIED, INCLUDING WITHOUT ANY WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND ANY WARRANTY OR CONDITION OF NON-INFRINGEMENT. IBM products are warranted according to the terms and conditions of the agreements under which they are provided.