Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
In our Executive’s Guide to Agentic AI we explored how AI agents can reshape enterprise workflows and unlock high-value work across industries. Recent data validates the transformative potential of enterprise AI: an IDC and Microsoft study reveals organizations are achieving an 18% improvement in customer satisfaction, employee productivity, and market share. The financial impact is equally compelling — companies are generating $3.50 for every dollar invested, with ROI achieved in just 14 months.
As AI agents emerge with capabilities that surpass generative AI, the strategic imperative for implementation has never been clearer. The opportunity cost of delayed adoption is beginning to exceed the investment risk. However, successful deployment requires a clear vision for the future and a framework for quantifying potential returns.
We've developed a practical approach for estimating the business value (or opportunity value, as we’ll discuss in this article) of AI agent deployments so you can catch the third wave of artificial intelligence early. This guide illustrates the approach with three use cases from highly regulated and complex sectors.
When workshopping AI use cases, prioritization is an essential step, and having a way to quantify value helps quickly identify some of the best opportunities for your strategic goals. In this article, we’ll use this “opportunity value” to reference improvements in:
From ideation with collaborators to pitching to stakeholders and decision-makers, providing an approximate understanding of a use case’s opportunity value instills confidence and creates momentum. While precise figures can be elusive, providing well-reasoned estimates is a powerful catalyst for action. Will implementing agentic AI save your organization $1-3M annually or $10-30M? This playbook will give you the tools to answer that question.
We’ll examine estimates for agentic AI use cases in:
While compliance requirements and access to sensitive customer information raise the stakes in these sectors, the opportunities for short-term success and long-term potential are inspiring.
But first, let’s establish our formula for this exercise.
As with most technologies, the value of agentic AI will vary depending on the applied use case and an organization’s size. We represent these variables in this versatile formula for estimating value generated:
Opportunity ($) = baseline key metric x improvement factor range (%) x scale factor
We define the formula’s components as:
Our previous article in this series explored how AI agents can free employees to pursue high-value activities. This concept directly informs our improvement factor, as we can quantify the value of redirected human effort. For instance, if AI agents handle 1-2 hours of equivalent human work daily (a 12.5-25% offload) in an organization of 1,000 employees, then we’re in the $10-20M range of potential value — and this baseline doesn't account for the compounding impact of redirecting human talent toward strategic planning, innovation, and relationship building.
$10-20M value = 8 hours/day (x 250 working days x $40/hour effective wage rate) x 12.5-25% offload x 1000 employees
The same approach can be applied to estimate agentic AI’s value in examples like:
The critical part of any use case is articulating how the new technology will improve the key performance metric — and by how much. This entails listing specific activities agentic AI will take over to calculate the improvement factor.
Let’s do exactly that for our sector use cases and calculate the opportunity value, which can help set help benchmarks for future ROI analysis.
A recent Conning survey reveals that 65% of healthcare insurance payers are already piloting large language models for claims processing — a clear signal of the industry’s technological trajectory. Payers staff thousands of medical claims specialists to manage a multi-step process of review, approval, and payment processing. In order to evaluate a submitted claim, these specialists must toggle between databases and applications while carrying out sequential administrative and analytical tasks. While an LLM can act as a copilot for some tasks, an AI agent can act as a lightning-speed coworker, making decisions on lower-complexity claims and deferring to a claims specialist as needed as a “human in the loop.”
From claims receiving to payment, that workflow could look like this:
A claims specialist’s role would evolve into that of a manager, relying on the AI agent for most of the low-value, time-consuming aspects of their job (e.g., researching missing information and flagging miscodings). Unlike the rigid auto-rejection patterns of some of the sector’s previous AI implementations, agentic AI powers intelligent automation that streamlines processes while preserving the nuanced judgment essential to claims processing.
Let’s create a baseline for the potential opportunity value of this use case by assuming these specs for a claims specialist:
So, the net average cost per claim (annual salary / (claims per day x working days per year)) is $6.25 at 15 minutes per processed claim.
If we assume this updated process enables a 25% time savings, allowing 40 claims to be processed per day by a claims specialist, then the new net average cost per claim is $5.00 at 12 minutes per processed claim.
That’s a 20% cost reduction per claim.
Now, let’s scale the impact. If a national health insurer manages 30M planholders who make five claims per year, agentic AI has the potential to provide a $188M annual savings opportunity.
$188M value = $6.25/claim x 20% savings x 30M planholders x 5 claims/year
(Opportunity ($) = baseline key metric x improvement factor (%) x scale factor)
Future enhancements could substantially increase the value of this agentic AI use case. For example, integrating retrieval augmented generation (RAG) systems into electronic health records (EHRs) could power auto-submission of claims, allow decision support for medium-complexity cases, and train AI agents on the adjustments made by human claims specialists — all of which could rapidly improve agentic AI accuracy rates.
Let’s project how the savings seen above might scale with increased efficiency:
This opportunity value estimate assumes that time savings equate to more new claims processed, but the potential exists to deliver better healthcare outcomes for customers as well. Payers could use some or all of the time saved to process audits for pain points, improve compliance methods, or enhance the planholder experience.
Overall, this use case imagines a paradigm shift in claims processing. By combining unprecedented efficiency with enhanced accuracy and insights, thoughtful implementation of agentic AI can benefit all stakeholders in the Healthcare ecosystem — from payers and providers to patients themselves.
JPMorgan Chase recently rolled out LLM Suite, a proprietary generative AI assistant, to help 60,000 employees streamline tasks like writing emails and reports. This decision aligns well with the sector's administrative and analytical workload — it also underscores the potential for agentic AI to improve operational efficiency within Financial Services.
Consider the role of a Small Business Association (SBA) Business Development Officer, whose primary focus is building a robust client base through prospecting, community involvement, and relationship development with centers of influence. While these high-value activities require a human touch, agentic AI can enhance productivity by handling decision-making for complex supporting tasks:
Let’s approximate that outsourcing these tasks to an AI agent will free up six hours (15%) of a Business Development Officer’s work week. This improvement increases their weekly time for client meetings from 24 hours (60%) to 30 hours (75%).
A six-hour increase from 24 hours to 30 hours for new client meetings is a net 25% increase, and this additional time for face-to-face meetings increases the pool of prospective clients reached — let’s estimate that results in closing 25% more new clients annually.
Assuming an average SBA business development officer produces $10M in loans annually and their bank sells the guaranteed portion of SBA loans in secondary markets for a 13.5% net premium (i.e., $1.35M in revenues), that nets an extra $0.34M in revenues per employee.
Scaling to the size of a large regional bank with 100 business development officers, that’s an opportunity of $34M.
$34M value = $10M loans x 13.5% net premium earned by selling guaranteed portion in secondary market x 25% more business development x 100 SBA business development officers
(Opportunity ($) = baseline key metric x improvement factor (%) x scale factor)
Compelling opportunity value like this provides a foundation for Financial Services organizations to implement agentic AI across their business development operations. While staffing models and financial structures vary by institution, the fundamental advantage remains consistent: organizations can scale their revenue-generating capacity without proportional staffing increases — and existing teams can focus on the high-touch, client-facing activities that drive conversion and foster long-term relationships.
Customer Experience (CX) has emerged as the prime target for AI innovation in Telecommunications, according to McKinsey research, which found 73% of senior executives ranking it as a key priority. This focus is well-justified — despite significant advances in network speed, data capacity, and reliability across major carriers, customer satisfaction continues to lag. Still struggling to shed their traditional utility-provider image, Telecommunications providers consistently rank among the lowest in net promoter scores (NPS) across all industries. This persistent gap between technical capability and customer satisfaction creates a compelling case for agentic AI deployment.
The traditional Telecommunications customer service model presents a tension between cost and quality. Live agent calls, while commanding the highest "cost to serve" (CTS) due to expenses that include labor, benefits, and overhead, generally provide the best customer experience. Alternative channels, such as self-service portals and automated recordings, offer lower CTS but often at the expense of customer satisfaction and NPS. This inverse relationship between cost efficiency and service quality has long challenged Telecommunications providers seeking to improve their customer experience strategy.
Imagine a new CX channel that combines the cost efficiency of automated systems with some of the nuanced, personalized services of human agents. As agentic AI becomes more autonomous, conversational, skilled in translation, and attuned to the intricacies of voice and natural language, it presents the potential to efficiently handle customer voice calls at a fraction of the traditional cost.
To quantify agentic AI's potential impact on CX costs, let's analyze an example of current channel costs and then compare them to a scenario incorporating AI agents:
Now, let's introduce voice-equipped AI agents to lessen costs while potentially improving NPS. These AI agents would replace the IVR channel and could handle calls previously routed to recorded systems and make decisions on simple, recurring issues like device setup, troubleshooting, and account management. Here's how this could transform the cost structure:
The impact is significant: a monthly savings of $15,700, potentially scaling to $188,400 annually. This represents a 40.8% reduction in overall CX costs while maintaining the same query volume.
$184,000 value = $3/query (original live agent CTS) x 87.5% savings (midpoint of 75-95% range) x 70,000 annual queries shifted from live agents to agentic AI
(Opportunity ($) = baseline key metric x improvement factor (%) x scale factor)
According to this PWC report, 55% of surveyed customers say they would drop a company they liked after several bad experiences, and over 71% of customers expect companies (including Telecommunications providers) to deliver personalized experiences. With the help of artificial intelligence providers can begin to transform their service operations into revenue-generating centers focused on personalization. As AI agents efficiently manage routine customer inquiries — such as modem resets, billing adjustments, and pricing questions — service staff evolve into consultative sales roles, focusing on identifying customer needs, recommending premium services, and driving higher average revenue per unit (ARPU) through personalized cross-sell opportunities.
Even marginal increases in a customer's monthly value or slight reductions in customer churn can significantly impact a Telecommunication provider’s bottom line. In the following example, we take a Telecommunication company’s average margin per customer, across a variety of services, and then apply it to the enterprise’s millions of existing customers. We also assume monthly churn rates around 2%, which is standard for the sector.
Now, let's project the impact of AI. By enabling human CX specialists to focus on high-value interactions, we can expect:
While small increases in individual customer margin and a slight reduction in churn (0.1%) may not seem significant at a glance, our example shows that such gains could add $100M in aggregate customer lifetime value.
$100M aggregate CLV improvement = $1.1B baseline aggregate CLV x ~10-12% margin/churn improvement x 3.5M customers
(Opportunity ($) = baseline key metric x improvement factor (%) x scale factor)
As Telecommunications providers face increasing pressure to differentiate their services and improve customer satisfaction, agentic AI can catalyze sustainable competitive advantages — driving immediate cost savings and long-term revenue growth through enhanced customer relationships.
The opportunity value framework we've explored provides a foundation for evaluating agentic AI investments, but remember that organizations can always encounter pitfalls that skew results and lead to unrealistic expectations. Key areas of overlooked costs for AI implementation include training, integration, ongoing maintenance, and regulatory compliance. Regulatory costs, particularly crucial in highly regulated industries, involve managing AI-related risks such as data privacy concerns and algorithmic bias. These costs include risk assessment, mitigation strategies, ongoing monitoring, and compliance with current and future AI regulations.
By understanding and proactively addressing these challenges, you can ensure a more reliable assessment that will lead to more accurate ROI predictions.
Consult with your data science, legal, and operations teams to identify potential hidden costs for your infrastructure. A thorough and honest assessment provides a more accurate picture of potential returns and helps build a stronger case for adoption of this new artificial intelligence.
Looking for personalized guidance on your agentic AI journey? WillowTree's Data & AI team is here to help you assess your organization's readiness, identify prime opportunities for implementation, and develop a roadmap for success. Contact us today to explore how we can help transform your AI investment into measurable business value.
One email, once a month.
AI-fueled experience transformation.
Smarter products. Faster builds.
Automated marketing and operations.
Proudly a telus digital company
