Agentic AI is helping businesses across industries move beyond the hype cycle into real productivity and revenue gains.
Author: Macaulan Servàn-Chiaramonte
The AI that waits for instructions is giving way to AI that pursues goals. For most businesses, Agentic AI is the difference between AI as a hype-driven project and one that will provide meaningful returns, as it will be able to replace many of the monotonous tasks where valuable time is spent by people who could be better utilized in driving strategic initiatives or putting their creative thinking to work on bigger, thornier problems.
Agentic AI refers to systems capable of autonomous decision-making, planning, and multi-step task execution. These aren’t chatbots with better prompts. They’re digital workers that analyze situations, develop strategies, take actions, and adapt based on results, all with minimal human oversight. The potential of AI agents represents the most significant shift in enterprise automation since cloud computing.
The business case is already compelling: 79% of US organizations have deployed AI agents according to PwC’s May 2025 survey, with US enterprises projecting 192% expected ROI, the highest globally according to Landbase’s analysis of industry data.
Of course, all opportunity comes with challenges: Gartner predicts over 40% of agentic AI projects will be canceled by 2027 due to unclear value, escalating costs, or inadequate governance. This means that aligning your Agentic AI projects to your specific business goals is more important than ever. In fact, organizations scaling industry-tailored agentic solutions for core processes are three times more likely to exceed ROI expectations compared to those still experimenting.
Gartner named agentic AI the #1 Strategic Technology Trend for 2025, defining it as “AI systems with agency, within defined guardrails, to go beyond merely augmenting workflows to fully automating them.” Where Generative AI responds to prompts and traditional automation follows rigid rules, Agentic AI sets goals, creates plans, adapts to changing conditions, and takes action autonomously.
Four characteristics define an AI Agent:
The technical architecture enabling this autonomy combines large language models with reasoning loops, memory systems, and tool integration capabilities. Agents use patterns like ReAct (Reasoning + Acting) to interleave logic with action, calling external APIs, querying databases, and executing workflows while continuously adjusting based on results.
McKinsey’s research reveals the pace of capability growth: the length of tasks AI agents can complete with 50% success has been doubling every seven months since 2019, reaching approximately two hours of autonomous work as of late 2024. At this trajectory, AI systems could potentially complete four days of continuous work without human supervision by 2027.
Of course, these projections assume optimal task selection. Success with agentic AI requires identifying workflows with clear decision points, measurable outcomes, and tolerance for autonomous operation. Organizations achieving strong ROI focus agents on repetitive processes like claims processing and document review while avoiding tasks requiring nuanced human judgment. Automating the wrong work, no matter how long it runs, delivers little value.
While 2024 was a time for building the business case for Agentic AI, 2025 was all about moving beyond projections to documented results. PwC’s May 2025 survey of 308 US executives found that among adopters:
A few of the largest companies in North America have already seen serious returns.
Wells Fargo demonstrated agentic AI’s ability to handle complex workflows. Using LangGraph-based agents, the bank re-underwrote 15 years of loan documents with agents autonomously retrieving files, extracting data, cross-checking systems, and performing calculations with minimal human intervention. The Fargo virtual assistant handled 245.4 million interactions in 2024, up from 21.3 million the previous year.
JPMorgan Chase exemplified comprehensive transformation, with an $18 billion annual technology budget and roughly $2 billion dedicated specifically to AI. Nearly 250,000 employees have access to the LLM Suite platform, with CEO Jamie Dimon reporting the bank’s $2 billion AI investment now generates equivalent annual savings.
Walmart transitioned from model-centric to system-centric AI architecture with Wallaby, a proprietary retail-specific LLM trained on decades of company data. Results included a 30% logistics cost savings and a 25% improvement in customer satisfaction scores. Now, the company is deploying AI tools to 1.5 million associates.
It isn’t just the world’s largest corporations seeing meaningful ROI. Across nearly every industry, companies pursuing strategic use of AI have found real value.
Telecommunications offers a masterclass in agentic AI deployment. AT&T’s autonomous assistants achieved an 80% reduction in fraud attacks and cut GenAI application development time from seven days to one. Verizon’s Project 624 delivered a 40% sales increase for service teams, while T-Mobile’s IntentCX platform, built through a multi-year OpenAI partnership, is part of the carrier’s strategy to reduce person-to-person customer service interactions by 75% through proactive AI that executes tasks autonomously with customer permission.
Financial Services leads adoption of AI. For instance, Bank of America runs 270 AI and machine learning models across operations, with 18,000 developers using coding agents. And, their virtual assistant Erica handled 676 million interactions in 2024.
In another example, JPMorgan Chase is deploying agentic AI to handle complex multi-step tasks for over 140,000 employees, with systems that create investment banking decks in 30 seconds. This is completing work that previously required teams of junior analysts hours to complete. The bank’s LLM Suite platform, updated every eight weeks, doesn’t just assist, it autonomously executes cascading actions, pulling data from multiple sources and assembling comprehensive analyses without waiting for human approval at each step.
Healthcare shows agentic AI’s potential for autonomous clinical workflow optimization, though deployments remain carefully supervised due to regulatory requirements. Kaiser Permanente’s KPIN (Kaiser Permanente Intelligent Navigator) system, rolled out to 4.9 million Southern California patients in October 2024, allows patients to describe care needs in their own words. The system autonomously analyzes requests, determines appropriate care pathways, and produces a set of recommended offerings tailored to each case and completing digital encounters that would traditionally require call center intervention.
Retail (and Supply Chain) operations showcase agentic AI’s ability to orchestrate complex, dynamic systems. Walmart’s Trend-to-Product system is a multi-agent engine that autonomously tracks social media trends, generates mood boards, creates product concepts, and feeds them directly into prototyping and sourcing processes which shorten traditional fashion production timelines by 18 weeks.
Utilities are starting to use AI Agents to continuously monitor energy consumption patterns and autonomously balance load in real-time between different regions. When evidence of high demand is detected, the system independently redistributes power, triggers backup resources, and adjusts grid configurations without waiting for human operators to analyze the situation and issue commands. This autonomous orchestration maintains grid stability while optimizing efficiency across distributed energy resources.
While all AI projects benefit from strategic planning, with Agentic AI, it’s even more important to have the i’s dotted and t’s crossed before launching so you can extract the most value from your new workforce. Start by:
The enterprises that treat Agentic AI as comprehensive organizational transformation, rather than technology deployment, will be best positioned to capture the productivity and cost advantages as capabilities mature.
At RevGen we view AI as a tool for augmented insights and efficiencies, whether it’s Agentic AI, Generative AI, or Machine Learning. Our approach connects inputs to value, helping clients operationalize AI-based solutions within their existing data and analytics ecosystems.
The RevGen Generative AI Readiness Framework assesses organizational preparedness across eight dimensions: vision clarity, executive sponsorship, use case prioritization, data quality, existing AI capabilities, scalability readiness, integration capability, and change management capacity.
Our five-phase AI methodology (Define Vision, Prove Out, Operationalize, Realize Value, Drive Continuous Value) provides a structured path from use case identification through scaled deployment. We’ve used this method with most leading AI platforms, including Microsoft Copilot Studio, Anthropic, OpenAI, and Google Vertex AI. And RevGen’s Data & Analytics capabilities provide the foundational infrastructure agentic systems require: modern data architectures, governance frameworks, and AI-ready data pipelines.
Contact us today to discuss how our AI experts can mature your current capabilities or to schedule an AI Workshop to help you identify high-impact agentic use cases and build a roadmap for responsible deployment.
Macaulan is a Managing Consultant specializing in artificial intelligence, data governance, and enterprise automation. Through his expertise, he helps organizations navigate the rapidly evolving landscape of AI technologies.
How is a data & AI company handling their own AI Governance? This article gives a peek behind the curtain into our own processes and guidelines.
Read More
In the world of the new Smart Factory, companies are leveraging AI and big data for predictive maintenance, anticipating issues before they can impact performance.
Read More
AI-powered quality control is transforming manufacturing through smarter defect detection, real-time data analysis, and enhanced productivity
Read MoreGet the latest updates and Insights from RevGen delivered straight to your inbox.