How System Integrators Build AI-Agent Practices That Actually Make Money
A Survival Guide for Enterprise Software System Integrators
If you're someone who works in the professional services domain for enterprise software, you've had your share of endless commentary on the extinction of SaaS in an AI Agent world. While timelines for such extinction events are debatable, one thing we must all agree on: change is inevitable. It's coming. When? We don't know, but my experience working in this space for over a decade tells me it's coming faster than we thought.
I don't want to dwell further on this debate. My intention in writing this is to provide insights on how enterprise software system integrators can prepare themselves for what's ahead. I've spent over a decade in this space watching transformations unfold. I've seen on-premises ERP and CRM evolve into B2B SaaS. I've watched machine learning use cases embed into workflows, and now fully autonomous agents emerge as the next frontier. I've witnessed it all. I got my first lessons in enterprise software as a Business Analyst for CGI's made-for-government Advantage ERP. I spent time in technology consulting as a Microsoft and Oracle implementation partner in the USA. I worked for Oracle as part of their Fusion Applications go-to-market team in the Middle East, and now I'm associated with Microsoft as a member of the GTM team for AI Business Solutions, helping enterprises and public sector scale AI. The perspective I bring comes from watching these shifts happen from multiple vantage points within the ecosystem.
This will be a 7-part series where I'll break this playbook into easy to digest bite size learning, so make sure you read all of them.
Note: I am writing this in my personal capacity. Views expressed are my own and not of my employer.
Chapter One - The Agent-Native World Order
This new world order will be established by agents. Not AI assistants like ChatGPT, Gemini, Claude, or Copilot. Don't confuse the two. Think of AI assistants as your silent partner for all the work you do, quietly augmenting you. Agents are like team members capable of automating entire business processes. Think of them as the task doers.
Every business process will have an agent. Finance. Supply chain. Human resources. Sales. Service. Marketing. The list goes on. In this new world, we'll end up with vastly more agents than we have applications today. IDC projects there will be 1.3 billion agents deployed by 2028, with over 230,000 organizations already using platforms like Microsoft Copilot Studio to create and customize them.[1] Agents will function like the internet as we know it today: millions of sites, clever indexing systems, and interfaces to help you find what you need. Similarly, new interfaces will help us discover and orchestrate these agents. Essentially, the AI assistant will become the go-to place to look for agents. It's a fundamental restructuring of how work gets organized inside enterprises.
Professional services firms need to evolve their business and operating models for this agent-native world. Leaders who win won't be the fastest to adopt AI. They'll be those who define and defend the human value they sell. That's the core insight. That's what separates the firms that thrive from the ones that fade.
You need to see this as an opportunity to reinvent your organization, not as a threat. We've all done pivots in our lives. It's never easy, and we must start with hard questions: What customer value do I deliver that cannot be replaced by AI? Your answer to that question will determine your future. You need to double down on the subtle judgment you've cultivated over years, leverage the contextual expertise your team has built through thousands of man-hours on complex projects, and nurture the trust-based relationships you've sustained. These three things - judgment, expertise, and trust - cannot be commoditized by AI.
What Dies, What Survives
Before we dive into execution, let's be brutally honest about what's coming.
What will be commoditized?
Basic configuration and setup work
Standard data migration scripts
Generic training materials and documentation
Routine project management tasks
First-level troubleshooting and support
What remains defensible?
Deep domain judgment forged in the fire of real projects
Complex orchestration across multiple systems and stakeholders
Trust-based relationships built over years
The ability to navigate organizational politics and change
Strategic advisory that requires reading between the lines
McKinsey reported that approximately 40% of its projects in 2024 were AI-related, with nearly 500 clients requesting AI support.[2] Yet here's the paradox: while firms like McKinsey were cutting overall staff, they expanded resources in their digital divisions to meet AI demand. The message is clear - the work isn't disappearing; it's transforming. Firms that survive will be those that transform with it.
But here's the uncomfortable truth: 42% of companies abandoned most of their AI initiatives in 2025, up from just 17% in 2024.[3] The average enterprise scrapped 46% of AI pilots before they ever reached production. That's an 88% failure rate. Why? Because firms set overly aggressive timelines and underestimated complexity. This playbook won't promise you full transformation in 90 days. That's nonsense. What I will give you is a realistic roadmap: 90 days to get your first agent into production, then a 9-month journey to real transformation.
That's all for chapter one. Make sure you stop by again to read the Chapter Two - Building your Agent-Native Practice.
Notes:
[1] IDC. (2025, May). "1.3 Billion AI Agents by 2028." IDC Info Snapshot (US53361825). IDC projects over 1.3 billion AI agents will be deployed by 2028, with 230,000+ organizations already using Copilot Studio including 90% of Fortune 500.
[2] McKinsey & Company. (2024). "The State of AI in Early 2024" survey reporting 72% of organizations use AI in at least one business function (up from 55% in 2023), with 65% using generative AI regularly.
[3] S&P Global Market Intelligence. (2025). Survey of 1,000+ enterprises across North America and Europe showing 42% of companies abandoned most AI initiatives in 2025 (up from 17% in 2024), with 46% of AI proof-of-concepts abandoned before production.
If you're someone who works in the professional services domain for enterprise software, you've had your share of endless commentary on the extinction of SaaS in an AI Agent world. While timelines for such extinction events are debatable, one thing we must all agree on: change is inevitable. It's coming. When? We don't know, but my experience working in this space for over a decade tells me it's coming faster than we thought.
I don't want to dwell further on this debate. My intention in writing this is to provide insights on how enterprise software system integrators can prepare themselves for what's ahead. I've spent over a decade in this space watching transformations unfold. I've seen on-premises ERP and CRM evolve into B2B SaaS. I've watched machine learning use cases embed into workflows, and now fully autonomous agents emerge as the next frontier. I've witnessed it all. I got my first lessons in enterprise software as a Business Analyst for CGI's made-for-government Advantage ERP. I spent time in technology consulting as a Microsoft and Oracle implementation partner in the USA. I worked for Oracle as part of their Fusion Applications go-to-market team in the Middle East, and now I'm associated with Microsoft as a member of the GTM team for AI Business Solutions, helping enterprises and public sector scale AI. The perspective I bring comes from watching these shifts happen from multiple vantage points within the ecosystem.
This will be a 7-part series where I'll break this playbook into easy to digest bite size learning, so make sure you read all of them.
Note: I am writing this in my personal capacity. Views expressed are my own and not of my employer.
Chapter One - The Agent-Native World Order
This new world order will be established by agents. Not AI assistants like ChatGPT, Gemini, Claude, or Copilot. Don't confuse the two. Think of AI assistants as your silent partner for all the work you do, quietly augmenting you. Agents are like team members capable of automating entire business processes. Think of them as the task doers.
Every business process will have an agent. Finance. Supply chain. Human resources. Sales. Service. Marketing. The list goes on. In this new world, we'll end up with vastly more agents than we have applications today. IDC projects there will be 1.3 billion agents deployed by 2028, with over 230,000 organizations already using platforms like Microsoft Copilot Studio to create and customize them.[1] Agents will function like the internet as we know it today: millions of sites, clever indexing systems, and interfaces to help you find what you need. Similarly, new interfaces will help us discover and orchestrate these agents. Essentially, the AI assistant will become the go-to place to look for agents. It's a fundamental restructuring of how work gets organized inside enterprises.
Professional services firms need to evolve their business and operating models for this agent-native world. Leaders who win won't be the fastest to adopt AI. They'll be those who define and defend the human value they sell. That's the core insight. That's what separates the firms that thrive from the ones that fade.
You need to see this as an opportunity to reinvent your organization, not as a threat. We've all done pivots in our lives. It's never easy, and we must start with hard questions: What customer value do I deliver that cannot be replaced by AI? Your answer to that question will determine your future. You need to double down on the subtle judgment you've cultivated over years, leverage the contextual expertise your team has built through thousands of man-hours on complex projects, and nurture the trust-based relationships you've sustained. These three things - judgment, expertise, and trust - cannot be commoditized by AI.
What Dies, What Survives
Before we dive into execution, let's be brutally honest about what's coming.
What will be commoditized?
Basic configuration and setup work
Standard data migration scripts
Generic training materials and documentation
Routine project management tasks
First-level troubleshooting and support
What remains defensible?
Deep domain judgment forged in the fire of real projects
Complex orchestration across multiple systems and stakeholders
Trust-based relationships built over years
The ability to navigate organizational politics and change
Strategic advisory that requires reading between the lines
McKinsey reported that approximately 40% of its projects in 2024 were AI-related, with nearly 500 clients requesting AI support.[2] Yet here's the paradox: while firms like McKinsey were cutting overall staff, they expanded resources in their digital divisions to meet AI demand. The message is clear - the work isn't disappearing; it's transforming. Firms that survive will be those that transform with it.
But here's the uncomfortable truth: 42% of companies abandoned most of their AI initiatives in 2025, up from just 17% in 2024.[3] The average enterprise scrapped 46% of AI pilots before they ever reached production. That's an 88% failure rate. Why? Because firms set overly aggressive timelines and underestimated complexity. This playbook won't promise you full transformation in 90 days. That's nonsense. What I will give you is a realistic roadmap: 90 days to get your first agent into production, then a 9-month journey to real transformation.
That's all for chapter one. Make sure you stop by again to read the Chapter Two - Building your Agent-Native Practice.
Notes:
[1] IDC. (2025, May). "1.3 Billion AI Agents by 2028." IDC Info Snapshot (US53361825). IDC projects over 1.3 billion AI agents will be deployed by 2028, with 230,000+ organizations already using Copilot Studio including 90% of Fortune 500.
[2] McKinsey & Company. (2024). "The State of AI in Early 2024" survey reporting 72% of organizations use AI in at least one business function (up from 55% in 2023), with 65% using generative AI regularly.
[3] S&P Global Market Intelligence. (2025). Survey of 1,000+ enterprises across North America and Europe showing 42% of companies abandoned most AI initiatives in 2025 (up from 17% in 2024), with 46% of AI proof-of-concepts abandoned before production.
If you're someone who works in the professional services domain for enterprise software, you've had your share of endless commentary on the extinction of SaaS in an AI Agent world. While timelines for such extinction events are debatable, one thing we must all agree on: change is inevitable. It's coming. When? We don't know, but my experience working in this space for over a decade tells me it's coming faster than we thought.
I don't want to dwell further on this debate. My intention in writing this is to provide insights on how enterprise software system integrators can prepare themselves for what's ahead. I've spent over a decade in this space watching transformations unfold. I've seen on-premises ERP and CRM evolve into B2B SaaS. I've watched machine learning use cases embed into workflows, and now fully autonomous agents emerge as the next frontier. I've witnessed it all. I got my first lessons in enterprise software as a Business Analyst for CGI's made-for-government Advantage ERP. I spent time in technology consulting as a Microsoft and Oracle implementation partner in the USA. I worked for Oracle as part of their Fusion Applications go-to-market team in the Middle East, and now I'm associated with Microsoft as a member of the GTM team for AI Business Solutions, helping enterprises and public sector scale AI. The perspective I bring comes from watching these shifts happen from multiple vantage points within the ecosystem.
This will be a 7-part series where I'll break this playbook into easy to digest bite size learning, so make sure you read all of them.
Note: I am writing this in my personal capacity. Views expressed are my own and not of my employer.
Chapter One - The Agent-Native World Order
This new world order will be established by agents. Not AI assistants like ChatGPT, Gemini, Claude, or Copilot. Don't confuse the two. Think of AI assistants as your silent partner for all the work you do, quietly augmenting you. Agents are like team members capable of automating entire business processes. Think of them as the task doers.
Every business process will have an agent. Finance. Supply chain. Human resources. Sales. Service. Marketing. The list goes on. In this new world, we'll end up with vastly more agents than we have applications today. IDC projects there will be 1.3 billion agents deployed by 2028, with over 230,000 organizations already using platforms like Microsoft Copilot Studio to create and customize them.[1] Agents will function like the internet as we know it today: millions of sites, clever indexing systems, and interfaces to help you find what you need. Similarly, new interfaces will help us discover and orchestrate these agents. Essentially, the AI assistant will become the go-to place to look for agents. It's a fundamental restructuring of how work gets organized inside enterprises.
Professional services firms need to evolve their business and operating models for this agent-native world. Leaders who win won't be the fastest to adopt AI. They'll be those who define and defend the human value they sell. That's the core insight. That's what separates the firms that thrive from the ones that fade.
You need to see this as an opportunity to reinvent your organization, not as a threat. We've all done pivots in our lives. It's never easy, and we must start with hard questions: What customer value do I deliver that cannot be replaced by AI? Your answer to that question will determine your future. You need to double down on the subtle judgment you've cultivated over years, leverage the contextual expertise your team has built through thousands of man-hours on complex projects, and nurture the trust-based relationships you've sustained. These three things - judgment, expertise, and trust - cannot be commoditized by AI.
What Dies, What Survives
Before we dive into execution, let's be brutally honest about what's coming.
What will be commoditized?
Basic configuration and setup work
Standard data migration scripts
Generic training materials and documentation
Routine project management tasks
First-level troubleshooting and support
What remains defensible?
Deep domain judgment forged in the fire of real projects
Complex orchestration across multiple systems and stakeholders
Trust-based relationships built over years
The ability to navigate organizational politics and change
Strategic advisory that requires reading between the lines
McKinsey reported that approximately 40% of its projects in 2024 were AI-related, with nearly 500 clients requesting AI support.[2] Yet here's the paradox: while firms like McKinsey were cutting overall staff, they expanded resources in their digital divisions to meet AI demand. The message is clear - the work isn't disappearing; it's transforming. Firms that survive will be those that transform with it.
But here's the uncomfortable truth: 42% of companies abandoned most of their AI initiatives in 2025, up from just 17% in 2024.[3] The average enterprise scrapped 46% of AI pilots before they ever reached production. That's an 88% failure rate. Why? Because firms set overly aggressive timelines and underestimated complexity. This playbook won't promise you full transformation in 90 days. That's nonsense. What I will give you is a realistic roadmap: 90 days to get your first agent into production, then a 9-month journey to real transformation.
That's all for chapter one. Make sure you stop by again to read the Chapter Two - Building your Agent-Native Practice.
Notes:
[1] IDC. (2025, May). "1.3 Billion AI Agents by 2028." IDC Info Snapshot (US53361825). IDC projects over 1.3 billion AI agents will be deployed by 2028, with 230,000+ organizations already using Copilot Studio including 90% of Fortune 500.
[2] McKinsey & Company. (2024). "The State of AI in Early 2024" survey reporting 72% of organizations use AI in at least one business function (up from 55% in 2023), with 65% using generative AI regularly.
[3] S&P Global Market Intelligence. (2025). Survey of 1,000+ enterprises across North America and Europe showing 42% of companies abandoned most AI initiatives in 2025 (up from 17% in 2024), with 46% of AI proof-of-concepts abandoned before production.