Evènements T.I.C. - Informatique

[Replay] Artificial Intelligence: opportunities and limitations. What are the critical success factors for the implementation of the technology?


  • Hesham Fayed, VP - Head of Sales Middle East, Atos
  • Prof. Theodoros Evgeniou - Pr. of Decision Sciences & Technology Mgmt at INSEAD
  • Moussa Zaghdoud, Executive Vice-President Cloud, Alcatel-Lucent Enterprise
  • Moderator: Valérie Hawley, Executive Director, Sorbonne Centre for Artificial Intelligence - SCAI Abu Dhabi

View the video

Key takeaways

All companies wanting to implement AI strategies go through similar set of stages with different levels of maturity:

  • Stage 1 is where a central unit supports the business units to experiment in AI, leading to slow cultural changes but nothing in terms of business value creation - it's more about soft business value creation and usually just costs money.
  • Stage 2 for more mature companies is when they move to a product oriented approach to AI and data technologies and that’s when they start creating value. On top of data scientists they need to hire business unit managers who understand business cases and product managers who create the products.

What it means to be a data-driven organization:

  • It is a mindset change: by moving from a gap-filled decision-making to an evidence-based decision-making organization.
  • Then it's about having tools available at all parts of the organization for users middle managers or all managers and all employees to use to access data
  • You then have to utilize those tools
  • And eventually you have new products with AI embedded on them

AI strategy is only successful if it has been adopted so adoption and the right level of engagement from organizations is essential. To do so the most important things are:

  • To ask the basic question ‘what is the problem we want to solve’
  • Make sure that we take into account all the ecosystem and the source of information
  • Ensure we rise the right level of expectation
  • Make sure that the user interaction is as simple as possible
  • The new trends that are going to affect AI and the domain of AI are 5g smart destination, autonomous cars, AEG technologies, real-time data analytics and fidget analytics, video analyzing and crowd management. It can also be an enabler and accelerator of decarbonization.

AI requires very specific technological capability and human capability so the success of projects rests on communication between data scientists, product managers and business experts, the storytelling with data in order to communicate and convince people. This is where data and information are used as part of the political process to enable organisations to make decisions: data is influence.

You have to have data scientists, product managers and business people talk to each other and understand each other. So you need everybody to have a mix of technical/ analytical and business skills. Organizations should have a combination of internal and external AI skills.

Key factors for a successful implementation:

  • Have the whole organization embrace that strategy: by engaging everyone in the conversation and making sure everyone adopts it - so the benefits of the AI solution need to be easy to demonstrate within a process.
  • Upskill the whole workforce the managers more than anybody else
  • Connect the data capabilities with the business side all the time
  • Consider risk adjusted business value
  • AI projects can not be implemented the traditional way: they need to be managed in an agile manner with continuous improvement in mind.   

Evénements sur le même thème

Partager cette page Partager sur FacebookPartager sur TwitterPartager sur Linkedin