We hear about generative AI and its impact all the time, yet companies are still struggling to get to grips with the subject in a concrete way.
Contrary to what one might think, the spread of generative AI in the enterprise is still sporadic. While some employees use it in the workplace, they often operate it with their own ChatGPT license. Despite the presence of AI in the TOP 3 of corporate priorities, just over 10% of companies have launched a significant business impact study.
What are the current obstacles, and how can you take the first steps towards enterprise AI?
What are the main obstacles to adopting AI?
Over the past 12 months, I've led nearly 40 conferences for SBF120 COMEXs. Three main trends are holding back the deployment and use of AI within our companies: cybersecurity, lack of training and insufficient ability to showcase use cases.
First hurdle: cybersecurity fears
For most of the organizations we interviewed, cybersecurity is the main obstacle. In their view, generative AI is easier to hijack, and competitors or malicious individuals could gain access to highly sensitive data.
But are companies really taking the best route? Because the real threat, from a cybersecurity point of view, is that if they are slow to embrace generative AI, employees will use their own private licenses for business purposes. And this represents an even greater risk of data leakage!
Second obstacle: lack of AI training
ChatGPT’s huge success with the general public may have led some to believe that these new tools based on generative AI were easy to learn and operate. But this is not the case, and the lack of AI knowledge among managers and employees remains significant. The multiplicity of artificial intelligences and their novelty in the daily lives of employees explain the need for support. On this subject, several levels need to be involved: the company itself, but also solution providers, must be able to train and acculturate teams to these new solutions.
Third obstacle: the legibility of AI’s positive effects
Within organizations, there is a preliminary and quite natural need to identify uses based on joint reflection between managers, Business Unit directors – HR and IS/Digital. For many leaders, the concrete applications of generative AI are still unclear, and the benefits to be derived from AI-based tools require collective intelligence.
Finally, even if COMEXs are aware that they are facing the next technological revolution, few of them are taking the plunge.
Of the 35 companies scrutinized for AI in 2023, only 4 (or 11%) have conducted a robust impact analysis to identify the jobs most at risk.
The 20 statistics on AI adoption are available on our article“the real adoption of AI in business“.
TheAI ACT in Europe: between ethical regulation and workload for companies
The AI ACT, passed in Europe on March 13, 2024, will come into force in June 2024. It marks a significant turning point in AI regulation in Europe, while the USA has simply passed an executive order stipulating that every company must provide the results of safety tests to the administration: two rather different legislative visions coexist.
The AI Act aims to provide a framework for its use, emphasizing transparency, safety and the protection of fundamental rights. For companies, this translates into a double opportunity: strengthening confidence in the ethical use of AI, and obtaining CE marking, a guarantee of compliance with European standards.
However, compliance also represents a substantial challenge. The requirements impose a definition of the level of dangerousness of the system developed, from minimal risk to unacceptable risk. Following this diagnosis, companies marketing these devices have between 6 and 36 months to comply. In particular, 24 months will be given to systems assessed as having a limited and high level of hazard.
What are the 5 steps to compliance?
- Map their AI systems: identify existing and future AI projects to ensure they comply with new regulations.
- Set up a dedicated project team: include compliance, technology and legal experts to steer compliance.
- Assess associated risks: analyze each AI system to identify potential risks and implement corrective measures.
- Testing systems for bias : carrying out rigorous tests to ensure that algorithms do not unfairly discriminate against certain groups.
- Allocate adequate budgets: set aside financial resources for initial compliance and ongoing monitoring.
This approach enables companies to structure their compliance process efficiently, identifying priorities and allocating resources optimally.
Examples of positive-impact corporate programs
A number of organizations, both public and private, have already set up working groups or internal projects to explore the influence and benefits of artificial intelligence. These include L’Oréal and even France Travail. Because generative AI is not just a lever for productivity, we’ve illustrated 2 use cases to inspire you:
Business transformation at one of France’s leading gaming operators:
This company has also initiated a major reflection on the impact of AI on all its activities. We are working with them to identify the ways in which generative AI is transforming their activities and jobs: automation, the emergence of new tasks within their current positions. In this way, the company equips itself with the tools it needs to guarantee the employability of its staff, and to plan for the future of skills critical to its performance.
Diversifying the profiles recruited by a consulting firm
Another example: we recently worked for one of the world’s leading consulting, auditing, legal and tax firms. They used AI to match skills candidates in their recruitment process. Whereas before, the profiles selected were reproduced, the AI injected enabled them to diversify the talent retained.
The company has gone from having 3% “atypical” profiles to around 15%!
The proof is in the pudding that AI is shedding new, positive light on HR practices.
Key success factors for AI adoption
1. Strategic Alignment
Ensure strategic alignment between key executives (CEO, CFO, COO, HR) to define the transformations to be carried out. This involves understanding the challenges and opportunities offered by AI, and integrating them into the company's overall vision.
2. Strong governance
Establishing solid governance around the AI project is crucial. This includes the creation of a dedicated project team, made up of technology, compliance and legal experts, as well as business experts, to ensure effective coordination and ongoing monitoring of systems and initiatives in the design stage or more advanced.
3. IT & Data investments
Investments in IT and Data technologies are essential to support AI adoption. A strong HR/IT pairing needs to coordinate these investments, assessing existing skills and planning the training needed to bridge the gaps.
4. Data & AI-based tools
Use data and AI-based tools to meet the specific needs of each business. This means developing or integrating tailored AI solutions, automating repetitive tasks and optimizing decision-making processes.
5. Social strategy and ROI management
A social strategy focused on reskilling is essential to support employees in the transition to AI. ROI management must combine HR costs and operational gains, using clear metrics to measure the effectiveness and impact of AI initiatives.
Four-step methodology for adopting AI
Step 1: Awaken awareness and acculturate to AI
Start by raising awareness of AI among all stakeholders. Educate executives about current trends, emerging technologies like deep learning, and the potential impacts of AI. Then extend this education to all employees to help them understand how AI can transform their everyday working lives. This acculturation phase is crucial to prepare the ground and get AI accepted as an ally rather than a threat.
Step 2: Mapping and analyzing needs
Accurately map your company's current systems and processes. Identify pain points and bottlenecks in your operations, where inefficiencies and errors are common. Prioritize these areas to focus your efforts where AI's impact will be most significant. Then explore opportunities for improvement by considering how AI can automate repetitive tasks and optimize decision-making processes.
Stage 3: Developing and integrating AI solutions
With a clear vision of needs and priorities, develop and/or integrate appropriate AI solutions. This involves not only automating time-consuming tasks, but also optimizing high value-added processes. Use specific SaaS software for each business line, such as marketing or HR, and make sure these tools are easy to use and perfectly aligned with end-users’ needs.
Step 4: Continuous training, evaluation and adjustment
Set up ongoing training programs to ensure that all employees are comfortable with new AI technologies. Establish feedback mechanisms to gather user feedback and adjust strategies accordingly. Rigorously track performance indicators to measure the impact of AI initiatives, combine HR costs with operational gains and effectively drive ROI.
Conclusion
By following this structured four-step methodology, companies can turn the challenges of AI adoption into opportunities for growth and innovation. AI, strategically integrated and rigorously governed, becomes a catalyst for positive change, aligning internal skills with market needs and strengthening the company's overall competitiveness. It's an adventure that takes time and resources, but when done right, delivers tangible and lasting results.
