You’ve heard it before: AI has arrived. It’s revolutionizing business. (“OMG, obviously,” as our kids might say.)
But what does that mean for IT services leaders who must juggle near-term business demands while preparing for an AI-enabled future? How do they separate the hype from reality and understand where AI is delivering real value today?
In this article, we discuss practical ways AI is being used today across each phase of the delivery lifecycle today. How forward-looking companies are using a mix of people, AI tools and agents to accelerate the pace of delivery, free up time for delivery teams to do higher level work, and improve adoption.
As former consultants who now work inside software vendors, we offer a unique perspective on how AI is affecting not only our own companies’ product development, but also how it’s influencing the way our service partners leverage and deliver solutions around our products.
AI Across the Service Delivery Life Cycle
It helps to start by thinking of AI as a highly efficient team member. One that is exceptional at repeatable, mundane tasks and infinitely scalable, but still needs oversight. It’s possible today to weave AI-enabled tools and processes across every aspect of the service delivery life cycle (SDLC), from planning through execution and support.
Discovery
Discovery is about fully understanding the scope of a client’s needs. It starts in the sales (or ideation) process, and continues through requirements development. This is when teams alongside clients establish the plan of action, and define features and user stories.
Traditionally, discovery was accomplished manually. Handoffs from sales and delivery were choppy and often incomplete. Delivery teams would then conduct dozens of interviews (in person or long virtual sessions), collect data, map current systems, identify gaps, write requirement documents, and so on. The process relied heavily on human labor. It was time consuming and error prone.
AI can automate much of the tedious grunt work enabling teams to focus on the needs behind the requirements and higher-impact tasks.
Consultants will still manage the overall process and the customer interactions and relationships that are required to do this phase well. However, AI can automate much of the tedious grunt work enabling teams to focus on the needs behind the requirements and higher-impact tasks.
Solution development
This phase entails architecting, developing and building a solution. Traditionally, teams would brainstorm and manually design the architecture. They would interpret business needs and develop diagrams, infrastructure plans, integrations, workflows and proofs-of-concept.
Many delivery teams have already become adept at leveraging existing code or IP to accelerate delivery, but AI can take it a step further, fully automating some of an architect or developer’s day-to-day work.
Many delivery teams have already become adept at leveraging existing code or IP to accelerate delivery, but AI can take it a step further, fully automating some of an architect or developer’s day-to-day work. Popular coding co-pilots, AI-enabled migration utilities, digital business analysts (BAs), and intelligent search capabilities for asset and solution libraries are all tools that solution teams can bring into the mix.
Test & Deploy
In this phase, the solution development team hands the project off to a testing or release-management team. As that team confirms that components of the solution are functioning properly, the code is deployed via continuous integration and continuous delivery (CI/CD) to various environments.
Increasingly, AI can act as the tester and release manager. It can automate deployments, help identify deployment errors and automatically sync environments, all of which results in greater stability and reliability.
Increasingly, AI can act as the tester and release manager.
For example, Copado’s DevOps platform includes products like Robotic Testing, CI/CD, and Value Stream Mapping. VSM can intelligently analyze the cycle time for each phase of SDLC and identify bottlenecks, ultimately enabling the firm to assess whether AI-based process improvements reduced cycle time. It can also guide services firm on DORA metrics (key DevOps metrics measuring speed and reliability). Other more general DevOps vendors like GitHub and GitLab can also play a role in automating deployments to ensure seamless and efficient releases.
Thrive
In the Thrive phase, the solution has gone live in a production environment. During this phase, focus shifts to keeping the solution running smoothly, scaling it effectively, and evolving it based on user needs and business outcomes. This phase is no less critical than the build. In fact, for many organizations, Thrive is where the long-term value of a system is proven or lost.
AI can act as both a watchful system guardian and a continuous improvement coach.
In this phase AI can act as both a watchful system guardian and a continuous improvement coach. Intelligent agents can monitor production environments to proactively detect anomalies, predict outages and recommend optimizations. AI can also be used to develop training and change management plans and documentation to increase and expand system adoption.
And last, but certainly not least, AI can also act as a supplemental customer support and success team. It can assist with customer communications, create knowledge articles or fully handle support issues. More and more, digital agents are able to solve a customer’s query before a case is even logged. If human intervention is required, AI-driven search greatly improves support efficiency.
Accelerating AI Usage within the Delivery Lifecycle
The advent of AI requires a cultural transition, a radical shift in how teams operate. Naturally, this demands change management. Some people are incredibly comfortable with change and will demand these tools to improve performance and stay relevant in their careers. Others may resist them.
Ultimately, it is up to leaders to ensure that the game-changing productivity gains of AI make their way into the organization. Some leaders, like the CEO of Shopify, will mandate it, while others take a lighter touch. Whichever tact you take, the first step is to give the organization encouragement to use these tools, and budget to make them available.
AI agents and tools need to be introduced, trained and continuously developed to address whatever tasks they will undertake, and their performance must be routinely evaluated and adjusted.
AI agents and tools need to be introduced, trained and continuously developed to address whatever tasks they will undertake, and their performance must be routinely evaluated and adjusted.
As these tools and agents are introduced into the flow, governance and measuring success is critical. Delivery leaders should set key performance indicators (KPIs) to help them measure what impact AI is having on teams, customers and overall solution quality to make sure things are improving. Not going the wrong way. Some of these KPIs might be:
- CSAT (customer satisfaction): How satisfied are customers with our service delivery?
- Time to value acceleration: How much faster are deployments and solutions delivered?
- Customer adoption and growth: Is it increasing and/or are more features and projects being started?
- ESAT (employee satisfaction): How do employees feel about how their jobs are evolving and the quality of the tools we introduce?
- Talent: Are the tools helping us attract and retain the best talent?
- Project profitability: Is AI driving delivery cost down and margins up, increasing profitability despite reduced billable hours? Are assets being created and used for repeatable processes?
- New customer growth: Are we acquiring customers more efficiently?
Next Steps
It’s clear that AI is reshaping how software and systems are being delivered both inside client organizations and services firms. AI offers unprecedented opportunities for efficiency, speed, and customer satisfaction in IT services. If you’re not already thinking about how AI can become more embedded in your SDLC, we can guarantee many of your competitors are.
You can start by working with your teams to think through high-impact use cases. Talk to peers to understand what they are doing, and invest in a few tools with proven AI capabilities in every part of the lifecycle. From there, spend time with teams to understand what’s working and what’s not. AI is evolving so fast that sometimes the best way to learn is to just jump right in, learn and adapt.
About the Authors
Katie Brown and Gloria Ramchandani worked together at Appirio, one of the earliest cloud consultancies known for their high-quality service delivery that could be done far faster, and often at a much lower cost, than traditional service providers. Gloria, currently SVP of product at Copado, consulted with Fortune 100 customers at Appirio and served as product manager of the firm’s DevOps-inspired Cloud Management Center. Katie is the VP of Expert Acceleration at Okta. Previously, she held delivery leadership roles at Appirio, managing solutions, strategic programs, global methodology and operations.
