According to a recent study conducted by CB Insights among 50 strategy team leaders from large enterprises, Generative AI is considered the cornerstone of technological innovation in 2024. However, despite its obvious potential, its adoption is still in its early stages, hindered by obstacles such as security, regulation, and resource management. This article provides a summary of the study and strategies adopted by pioneering companies to overcome these challenges.

 

A Major Technological Priority

According to CBInsights’ survey, generative AI occupies a central place in companies’ technological priorities. Nearly 94% of respondents consider this technology a significant or moderate priority for the next 12 months, surpassing essential areas such as data management, cybersecurity, and even non-generative AI.

It offers unique advantages, such as automated content creation, optimization of customer interactions through virtual assistants, among others. These functionalities enable companies to improve their productivity while reducing operational costs.

However, this strong prioritization does not mean that generative AI is already ubiquitous; although 32% of companies have deployed generative AI solutions, the majority (54%) are still in the pilot or evaluation stage. By comparison, mature technologies like cloud or data management have a deployment rate of over two-thirds. This slowness is explained by the complexity of integrating GenAI into existing environments, particularly in highly regulated industries.

 

Deployment: A Progressive but Strategic Adoption

The contrast between companies’ enthusiasm for generative AI and its limited deployment is explained by several factors. The departments actively adopting this technology are primarily customer service (76% adoption or testing), marketing, and IT services. Customer service, in particular, benefits from tools like advanced chatbots and automated responses that reduce processing times and improve customer satisfaction.

Nevertheless, the relatively low deployment rate in other departments, such as sales or product management, indicates that some companies are still hesitant to generalize the use of generative AI. This is often linked to structural and organizational obstacles.

 

Challenges to Overcome for Successful Adoption

The main barriers to generative AI adoption are varied but significant. Among them, risks related to security (46% of respondents) and confidentiality stand out. It is clear that this hinders deployments. This concern is particularly marked in the financial (78%) and healthcare (73%) sectors, where sensitive data is omnipresent. Companies fear that security breaches could compromise information confidentiality while creating threats of regulatory non-compliance. “Any potential vulnerability is a major source of concern. We therefore proceed cautiously, even if it slows down our adoption rate,” from an executive in finance.

Other challenges include competing priorities within organizations (42%) and legal uncertainties (40%). Indeed, the legal framework surrounding AI remains unclear in many countries, which slows its large-scale adoption.

This challenge is particularly pronounced in organizations where technological resources (talent, budgets, computing power) are already committed to other projects. Companies must therefore prioritize initiatives offering the best return on investment.

One executive explains that shareholders favor projects with immediate ROI, thus creating tension between long-term innovation and short-term financial requirements.

In regulated sectors, complex legislative frameworks often hinder innovation. Integrating generative AI into these environments requires considerable efforts to ensure compliance with strict standards, such as GDPR in Europe or regulations specific to healthcare and finance.

Data quality issues (34%) and high development costs (30%) add an additional layer of complexity. These challenges highlight the need for companies to implement robust implementation strategies, relying on proven technological solutions and partnerships.

 

Initial Focus on Customer Departments and Internal Process Automation

Customer service (38%) and marketing (30%) are the first departments to benefit from generative AI: they offer concrete and measurable use cases. For example:

  • Customer service: Optimization of response times and satisfaction through solutions based on predictive analysis. A financial company uses GenAI to identify customers requiring specific credit solutions.
  • Marketing: Generation of personalized content, optimization of advertising campaigns, and improvement of customer loyalty.

Internal functions such as operations (42% pilot activity) and finance (36%) are becoming major targets for generative AI.

  • Compliance: Automation of regulatory controls to reduce time and costs related to report generation.
  • Finance: Automated invoice processing, risk assessment, and document management.

 

Build or Outsource: A Crucial Strategic Choice

One of the most important questions for companies concerns how to implement generative AI. The data shows that approximately 85% of organizations prefer to outsource this competency by opting for ready-to-use solutions, rather than developing their own tools.

This outsourcing trend is explained by several factors. First, solutions offered by specialized providers deliver technological maturity and expertise that are difficult to replicate internally. Second, companies can thus reduce initial costs and accelerate implementation timelines. However, this choice also raises questions about dependence on these providers and implications regarding data control.

 

Future Priorities: Between Maturity and Innovation

While generative AI attracts particular attention, other technological priorities continue to engage companies. Data management and cybersecurity, for example, record higher deployment rates (74% and 68% respectively). These areas support the infrastructure necessary for generative AI to be fully leveraged.

In this context, generative AI should not be perceived as an isolated technology. It is part of a broader technological ecosystem, where data quality, security, and integration with other solutions play an essential role.