By Luca Collina
In working with a new client who wondered about Agentic AI solutions, I found that the simplest thing was to ask that providers explain their solution without technical complexity, so that we could read their underlying logic and process flows. Using this approach allowed a more informed decision based on the goals of the business instead of just the technical specifications.
Having experienced this firsthand, my belief that process-minded professionals (those with credentials in Business Analysis, Six Sigma Lean and process-focused methodologies) will be critical in determining which Agentic-AI can both be embedded in existing business processes. IT professionals have a unique opportunity to take on a pivotal role in AI-enabled transformation, as no other role marries automation with process efficiency, cost reduction, and ERP systems.
My position is based on:
Certified professionals have an in-depth understanding of process efficiency, waste minimization, and quality improvement. This ensures Agentic-AI tools complement rather than increase the friction to business. This is to enable seamless AI integration, as they can map, analyse and optimise workflows. Certifications also signal some specific skill sets, such as the statistical and analytical knowledge required when assessing how AI-powered automation can affect efficiency and performance.
Professional experts are skilled in requirement gathering, communication with stakeholders, and aligning technology solutions with business objectives. Generation AI’s capacity for automation and predictive analytics may lower costs and improve time-to-market, but without guides for how these set AI use cases to pursue for maximum return on investment, many companies will have difficulty knowing what areas of the business to apply AI. Digital transformation professionals and those with significant ERP implementation experience are the most able to plug Agentic-AI into enterprise systems and architectures.
Agentic AI: A Simple Overview
Agentic-AI may be defined as a system of automated artificial intelligence interacting in the environment over time, adjusting to dynamic conditions pursuing predefined goals. Instead of adhering to traditional AI models that operate on pre-written commands, Agentic-AI processes information through machine learning, reinforcement learning and decision-making frameworks to assess the available data and act accordingly.
How Agentic-AI Works:
- Information Gathering: The Agentic-AI system gathers information from different sources like databases, and real-time operational data.
- Analytical Decision-Making: This step involves analyzing the information and determining what actions to take using sophisticated algorithms.
- Task Execution: After making a decision, the system will perform tasks such as automation, generate reports and recommendations.
- Adaptive Learning: Because of feedback loops, the system can adjust its processes over time, making the processes more accurate and efficient.
Agentic-AI refers to artificial intelligence systems that operate autonomously, adapting to dynamic conditions while pursuing predefined goals. Unlike traditional AI models that rely on structured commands, Agentic-AI employs machine learning, reinforcement learning, and decision-making frameworks to evaluate data and take appropriate actions.
Case Studies on Agentic-AI Implementation Approaches
Agentic-AI implementation is a heterogeneous enterprise across organisations. While some companies focus on working collaboratively between process experts and AI solution providers, some of them go to the extent of working with the vendors only. The following case studies highlight each of these strategies:
Collaboration Between Process Experts and Solution Providers:
Accenture and NVIDIA in Manufacturing
In predictive maintenance and quality control, Accenture collaborated with NVIDIA to implement Agentic-AI solutions. Accenture’s process specialists worked with NVIDIA’s artificial intelligence experts to create predictive maintenance applications that reduced machine downtime by 20% and defect rates by 25%. By leveraging its process domain expertise, this partnership shows how AI technology can help within its operations, minimizing excess resources while maximizing product quality.
UPS’s ORION System
UPS built the On-Road Integrated Optimization and Navigation (ORION) system, an Agentic-AI that autonomously adapts to traffic, weather, and package volume, among other factors. Their backend engineers worked with the developers to make sure the AI was working toward things like delivery efficiency, reducing their fuel costs, optimizing delivery routes, things like that, so they could make sure they were operationalizing this.”
Sole reliance on AI solution providers.
Morgan Stanley’s In-House AI Development
Morgan Stanley launched AI @ Morgan Stanley Debrief, a generative AI app that summarises meetings and drafts emails. While there are other cases Morgan Stanley developed the solution internally, without help from process experts external to the company. Yet working in this manner led to scalability issues as there lacked outside variations in this approach.
Salesforce’s Agentforce 2.0
Salesforce unveiled Agentforce 2.0, an AI agent to automate customer support s. The company concluded that developing this solution internally without professional input suffered
Table 1: Comparison of Implementation Approaches
A Holistic Framework for Agentic-AI Integration
In order to enable the economic benefits from Agentic-AI and diminish the risks, businesses need to pursue a structured and balanced approach. The S.M.A.R.T.E.R.-Strategic, Measurable, Adaptive, Reliable, Transparent, Ethical, Resilient © provides an extensive framework for effective Agentic-AI adoption.
Strategic
The implementation of Agentic-AI should be done in the context of a business needs, making sure that automation does not disrupt your main business.
Measurable
These could be anything from cost savings to increased efficiency of processes but It requires to define some key performance indicators (KPIs) to measure your Agentic-AI impact.
Adaptive
Agentic-AI should be able to be developed alongside organizational requirements, enabling organizations to iteratively modify and optimize implementations as time progresses.
Reliable
We need to thorough and constantly testing Agentic-AI systems to ensure that they are retaining cognitive accuracy, operational efficiency, and more importantly, consistent outputs.
Transparent
Agentic-AI decision-making processes should be explainable to core stakeholders: avoid black-box operations that erode trust.
Ethical
Ensure AI compliance: Agentic-AI deployments should comply with legal frameworks for fairness, bias, and impact on the workforce.
Resilient
Having a human-AI team, able to take charge when a system fails or when there is unexpected operational disruption.
These principles help ensure that such integration is sustainable and responsible, minimizing risks while maximizing long-term benefits for businesses and society.
In the new project for my current client, I have won a “resistance” from the provider who claimed to be a leader due to some enterprise solutions in some industries working alone Now we define and discuss to build things together.
With a focus on #impact and #intensity, I have wrote and communicated what I expect over a pragmatic path: a roadmap to pragmatically smooth out the Agentic-AI adoption process.
Table 2: impact and intensity
PS: Did I speak about the technical solution so far?
About the Author
Luca Collina is a transformational and AI Business consultant at TRANSFORAGE TCA LTD. Awarded by York St John University with Business –Postgraduate Programme Prize and by CMCE (Centre for Management Consulting Excellence-UK) for his paper in Technology and Consulting . Published Academic author. Thought leader with THINKERS360 in GEN-AI, Business Continuity, and Education.
