Organizations are embracing advanced technologies powered by artificial intelligence (AI) with Generative AI (GenAI) tools becoming increasingly accessible to the modern workforce. According to Deloitte’s latest quarterly executive survey, State of Generative AI in the Enterprise, just under 40% of the workforce gained access to GenAI tools over the past year. As the workforce continues leveraging GenAI capabilities, and the underlying large language models (LLMs) to generate content at a prompt, AI agents capable of reasoning are being utilized to support end-to-end transactional work, analysis, and decision-making.
Rise of AI Agents
Foundation models help enable users to “chat” with data for generating content, code, and some level of insight. The evolution of foundational LLMs to advanced reasoning models have opened new gateways of opportunities for Agentic AI, empowering AI agents to play a more significant and impactful role.
AI agents are autonomous software systems endowed with reasoning capabilities to process large volumes of data, comprehend context, reason logically, design complete workflows by interfacing with other systems and undertake actions including executing tasks.
When multiple AI agents are put together to form a multiagent system, these AI agents can help revolutionize how processes are automated and analyzed, often leading to faster processes. These AI agents can interact and collaborate when tailored to the specific needs and complexities of processes—handling sophisticated tasks and improving overall efficiency. Many aspects of enterprise architecture can be elevated, going beyond mere automation of current processes and tasks, to fundamentally reimagining and improving them.
Powering functions with AI agents
Modern enterprise workflows follow a relatively linear model where junior staff enable transactional processes including data collection. Mid-level teams analyze details to help identify trends and extract actionable insights and top management leverage these insights to formulate strategy. While that is an ideal scenario, top management often face an overwhelming volume of data with technology and human-related challenges often limiting their ability to interpret and monetize the insights. This gap presents a significant opportunity for innovation.
Digital workers or AI agents powered by advanced reasoning models help enable enterprises to streamline all four phases—transacting, analyzing, decision-making, and actioning —delivering increased efficiencies and, more importantly, quality insights in minimal time. By leveraging these agents across various functions, organizations can gain an advantageous position where they not only enhance operational efficiency but also empower human workforces for a smarter future, allowing employees to concentrate on more critical, creative, and strategic tasks. This shift can allow employees to focus their skills towards problem-solving and innovation, supporting the drive for greater business value. Below are some examples of functions where AI agents can have significant impact:
- Finance: At an operational level, AI agents can perform transactional tasks such as managing financial data, invoices, cash applications, and closing accounting cycles. They can analyze and advise on sales and profitability, working capital, and expenses to deliver qualitative insights to chief financial officers—helping to enable better assessment of each function’s financial performance while providing recommendations on potential actions to take.
- Sales and marketing: AI agents, powered by advanced reasoning, can perform contract renewals, billing, invoicing, and sales order management. They can analyze and advise on trends for chief commercial officers looking to identify high performing and critical areas to address.
- Procurement: AI agents can efficiently perform routine tasks such as supplier data management and purchase requisition, while also managing and advising on more complex work related to category management and supplier intelligence with minimal human intervention.
There are many other functions such as customer experience, recruiting, and human resources where agents can play a significant role in enhancing processes and developing innovative solutions that help create measurable outcomes and enhance value.
Enabling Agentic AI-powered transformation
By advancing from text-to-text tasks to text-to-action capabilities, enterprises can leverage AI agents to help automate complex workflows. As these agents evolve in their reasoning and analytical capabilities, enterprises can be well placed to connect additional agents and assign extended processes, driving enhanced operational efficiency and strategic agility. The following actions can help organizations accelerate their agentic AI journey:
- Starting small: Enterprises can get comfortable with agentic capabilities and outputs by starting with one component or layer of a larger end-to-end process. Starting small can assist with achieving tangible return on investment more quickly while being able to get feedback faster for delivering informed decisions on scaling broadly into more complex workflows.
- Optimizing data for Agentic AI architecture: Data serves as the cornerstone for the efficiency of multiagent AI systems in producing cognitive output. It’s critical to compile and integrate authoritative data sources for each use case, and pre-training and fine-tuning the models on this information ensures that AI agents can interpret the data within the applicable context. By employing knowledge engineering, enterprises can structure available data into a taxonomy to enable agents to efficiently access and utilize the necessary information.
- Building trust and upskilling in the workforce: The development and implementation of Agentic AI can necessitate significant contributions from the workforce, including capabilities in data engineering, business process engineering, machine learning, and ongoing application architecture. Upskilling the talent and earning their trust in these new advancements is important for success.
- Appointing apt technology for every stage: While an array of technologies of varying sizes and capabilities are available in the market, specific enterprise requirements often require a tailored approach. To select the most suitable technology stack, consider developing an evaluation framework to assign scores to potential options at each stage of your agentic architecture development.
- Preparing the workforce for transition: The transition to an agentic workforce—where humans and AI agents work in concert to help create exponential value—requires more than new skills. It requires a re-architecture of how work gets done, how decisions are made, and how performance is scaled. Organizations will likely need to confront the paradoxes shaping this shift: from managing people to managing performance across human-agent systems; from task-based roles to role-based design; from capacity growth through headcount to scaled performance through role redesign; and from use-case led AI adoption to AI-native business and operating models. This moment requires courage, speed and shared accountability—anchored in systemic co-dependency across domains. The path forward starts with where you stand, operates at dual speed, and moves from vision to scaled impact through bold, strategically coordinated work redesign.
Agentic AI platforms, such as Zora AITM by Deloitte offer a specialized digital workforce that can boost workforce productivity by streamlining processes and enhancing efficiency, particularly in addressing complex, functionally-nuanced work.
The role of trust
AI agents and multi-agent systems could likely prove to be a game changer for various functions across an enterprise, driving both speed and efficiency, while unlocking new sources of value. But while this can give rise to an evolved workforce that is able to deliver better outputs faster, earning human workforce trust will be essential. To achieve successful human-AI collaborations in agentic AI applications, organizations should explore how to build this trust (such as employing Deloitte’s Trustworthy AI™ framework) by incorporating fairness, transparency, and accountability into their enterprise AI applications.
Karen Pastakia, Human Capital leader for Deloitte Global, also contributed to this piece.
To learn more about specialized agents visit the Zora AITM by Deloitte page on Deloitte.com.
