AI driving cost reduction : from automation to agentic transformation

In a demanding economic environment, artificial intelligence is redefining cost reduction strategies and process optimization across organizations.

AI enhances traditional cost-reduction approaches at multiple levels. First, through its analytical capabilities : it provides companies with powerful ways to assess existing operations, including time spent, productivity, repetitive tasks, and value-added activities, far beyond the limits of traditional methodologies.

It also transforms operational levers themselves : the automation of repetitive tasks through AI significantly boosts operational efficiency, generating remarkable productivity gains that can reach 40 to 50%, and in some cases even 90%.

With the emergence of agentic AI, artificial intelligence is now entering a new phase.

A competitive and geopolitical challenge

AI-driven task automation offers companies an unprecedented opportunity to improve efficiency and optimize operations.

By automating complex or repetitive processes, including activities traditionally considered “creative,” engineering-related, or highly expertise-driven, organizations become more agile, strengthen their competitive advantage, redefine their production models, and enable employees to focus on higher value-added activities while simultaneously improving quality and operational speed.

The key outcomes include greater resilience to pricing pressure, improved margins, and stronger positioning in M&A dynamics and broader market shifts.

Employment is often not the primary lever

Our projects and observed use cases show that many benefits can be generated without negatively impacting employment :

• Reducing waiting times for expert-level tasks
• Multiplying service quality
• Capturing and processing new internal and external information
• Creating new customer services
• Scaling commercial opportunity processing
• Eliminating human error in operational tasks
• Enabling employees to focus on higher value-added work

The central issue is competitiveness within an ongoing global race.

The United States : leader in integration and deployment

Among global powers, the United States currently holds a significant lead in the development and adoption of agentic technologies.

This advantage is built on a unique technological ecosystem combining massive private investment, cutting-edge R&D, strong government support, and a culture of innovation highly favorable to experimentation.

Some notable examples include :

• Walmart : using the AI assistant “My Assistant” to answer HR-related questions and summarize meetings, reducing time spent on administrative tasks.
• FedEx : leveraging AI to optimize delivery forecasting and logistics routing, reducing fuel consumption and operational costs.
• General Electric (GE) : deploying AI-powered predictive maintenance systems, reducing machine downtime by 20 to 30% in certain factories.
• Pfizer : using AI to accelerate drug candidate identification, shortening some research phases from several months to just a few weeks.
• Delta Air Lines : integrating AI into customer service and ground operations, with 70% of customer requests handled through chatbots and a 10% reduction in flight delays.

‍

China : a pioneer in innovation and experimentation

China is emerging as a major player in the development of agentic artificial intelligence, following a very different approach from that of the United States.

Strongly supported by the state, Chinese innovation in AI relies on a centralized strategy and massive public investment.

While the United States favors a private and decentralized dynamic, China is pursuing a long-term vision, integrating AI into large-scale national initiatives such as the “New Generation Artificial Intelligence Development Plan.”

The country aims to become the global leader in AI by 2030.

Some examples of initiatives supporting this strategy include :

• Alibaba : using AI agents to automate logistics management, including warehousing and delivery operations.
• Tencent : leveraging AI to reduce chatbot response times on WeChat by 30% and accelerate game development by 25% through automated testing.
• Baidu : deploying autonomous taxis in major cities using agentic navigation AI.
• Ping An : integrating AI into telemedicine to provide automated pre-diagnosis capabilities.
• SenseTime : developing surveillance agents capable of detecting suspicious behavior in real time within urban environments.

In conclusion, the differences between these two approaches reflect their respective priorities : the United States focuses on rapid adoption to maintain its competitive advantage, while China intensively explores technological capabilities to establish itself as a long-term global leader across strategic societal and geopolitical dimensions.

These strategies influence not only their domestic economies, but also the global competitive dynamics surrounding artificial intelligence.

While the worldwide adoption of AI is already strengthening corporate competitiveness, a new phase is now emerging : agentic transformation.

The challenge now goes beyond task automation. The objective is to deploy autonomous agents capable of acting, collaborating, and making decisions to optimize performance.

Agentic Business Redesign (ABR) : a new approach to process transformation

Agentic Business Redesign is based on integrating AI agents capable of autonomously or semi-autonomously performing certain functions within business processes.

This approach enables organizations to rethink operational performance improvement : rather than marginally optimizing existing processes, the objective becomes redesigning workflows based on what agents are now capable of accomplishing.

This requires robust technological infrastructure, but above all, a structured transformation approach.

1. Targeted process assessment

The first step consists of identifying areas with the highest transformation potential.

Priority is generally given to processes that are :

• Costly or redundant
• Sources of friction or value loss
• Built around standardized, repetitive, or highly codified tasks

This assessment can rely on internal data analysis or exploratory AI models capable of objectively identifying priorities.

In some cases, this diagnostic phase can be completed within a few weeks, whereas more traditional approaches, such as Zero-Based Budgeting or business benchmarking, are often slower and less sensitive to the real operational dynamics.

2. Defining an agent-oriented transformation roadmap

Once the target processes have been identified, organizations move into the design phase.

Task by task, teams evaluate automation opportunities based on repetition, complexity, and strategic value.

This is where the specific value of agents becomes clear : they can manage decision chains, interact with business systems, and even perform functions traditionally considered “human,” including creation, analysis, and coordination.

Each potential solution is then assessed through strategic trade-offs : internal development, the use of existing platforms, or hybrid approaches combining both.

The roadmap may also include targeted experiments designed to quickly evaluate the value generated.

‍

3. Reallocating resources and enhancing human expertise

The implementation of AI agents is often accompanied by a redeployment of human capital.

Teams can be redirected toward more creative, relational, or strategic activities. The objective is not to replace employees, but to reposition them, ensuring that productivity gains also translate into greater engagement and more meaningful work.

4. Fast and scalable implementation

One of the key advantages of the agentic approach is its ability to deliver tangible results quickly.

Whereas traditional transformation programs often extend over 12 to 18 months, initial automations can typically be launched within 3 to 4 months, with short iteration and improvement cycles.

This agility enables organizations to test, adjust, and progressively industrialize solutions without waiting for large-scale system overhauls.

The challenges and key success factors of Agentic Business Redesign

Agentic Business Redesign opens up a new way of thinking about operational efficiency, moving beyond traditional cost-reduction strategies.

Where conventional approaches such as Zero-Based Budgeting focused on optimizing or eliminating low-value tasks, this new model encourages companies to rethink processes through the lens of what an AI agent could perform independently or collaboratively alongside humans.

For this transformation to deliver on its promises, several conditions appear essential.

1. Designing a modular agentic infrastructure

The first challenge is technical, but highly strategic.

To avoid the uncontrolled proliferation of isolated or opportunistically developed agents, organizations need to establish a coherent infrastructure.

This involves identifying core functions such as document understanding, interaction with business systems, or constrained decision-making, and encapsulating them into standardized and reusable building blocks deployable across the organization.

The objective is not to centralize everything, but rather to provide a shared foundation : a library of maintained, secured, accessible, and documented agentic capabilities.

This approach enables organizations to mutualize efforts, ensure agent robustness, and establish the foundations of clear governance.

2. Enabling business teams to build their own agents

A second critical success factor lies in user adoption.

Unlike traditional automation tools, which are generally managed by IT departments, AI agents are intended to be used, adapted, and even configured directly by business teams, provided they are given the appropriate tools and autonomy.

This requires rethinking role distribution : IT departments become providers of agentic capabilities, while business teams become architects of intelligent workflows through no-code or low-code interfaces.

Naturally, this distributed model requires support through training, awareness-building, and governance rules, but it is essential for achieving large-scale adoption without overwhelming IT teams.

3. Governing agentic resources as a new form of capital

Finally, once agents become integrated into business processes, organizations must address the question of governance.

We are no longer dealing with simple scripts or automation tools, but with true “software actors” whose capabilities evolve over time.

Monitoring them, allocating them appropriately, and maintaining operational performance become strategic topics in their own right.

It may therefore become useful to think in terms of “agentic resources” and imagine new organizational roles dedicated to managing them effectively.

As an exploratory example, some organizations are already considering the creation of an “Agentic Resource Officer” role, mirroring traditional HR functions : someone responsible for mapping active agents, monitoring their performance, ensuring their proper use, and supervising their evolution.

Although still emerging, this type of role could become a critical bridge between IT, business teams, compliance, and innovation functions.

Conclusion

The implementation of “Agentic” diagnostics represents both a value-creating and necessary transformation given the rapid evolution of competitors and market dynamics.

Organizations that undertake this work early will be best positioned to capture value, improve competitiveness, and strengthen organizational agility, ultimately becoming market leaders rather than followers.

By adopting Agentic Business Redesign, companies can not only reduce costs, but also improve agility and competitiveness.

This innovative approach transforms challenges into opportunities, positioning organizations at the forefront of the digital revolution.

At this stage, employment itself does not appear to be the primary risk : with an AI agent-centered approach, companies can transform their operations while simultaneously offering employees more enriching and higher value-added opportunities.

To learn more about these topics and identify the impacts and opportunities relevant to your organization, contact our partners and experts :

Simon Georges-Kot (Partner et Expert AI)

Jean-Charles Ferreri (Senior Partner et Expert Corporate et Efficiency)