For decades, IT services were built on a simple equation: more demand required more people. Application development and maintenance contracts were shaped around pyramids, rate cards, utilization, SLAs, and ticket volumes. Clients bought capacity, providers supplied skills, and improvement was negotiated as either cost reduction or productivity gain. AI challenges that logic. It does not merely make the same team faster; it changes how work is broken down, who or what performs it, where human review is required, how quality is evidenced, and how commercial value is shared.
In this context, product-oriented delivery (POD) becomes the natural container for this transition because it already includes a product, application, platform, or backlog end to end. In the AI era, it evolves into a “Human + Digital FTE POD,” where the digital FTE is the consistent term for AI copilots, task agents, agentic AI capabilities, and AI-automated workflows that perform bounded delivery activities. The goal is not role-for-role replacement. The better objective is activity-level redesign: repeatable, observable tasks move toward digital FTE execution, while people focus on intent, judgment, stakeholder trust, exception handling, and accountability for outcomes.
AI is changing IT service delivery from a people-led execution model to a coordinated operating system in which humans and digital FTEs share responsibility across the delivery life cycle. Instead of adding effort whenever demand rises, service teams are beginning to redesign work itself: digital FTEs prepare analysis, draft artifacts, detect patterns, accelerate testing, summarize incidents, and surface options, while humans decide what matters, manage trade-offs, validate quality, engage stakeholders, and remain accountable for business outcomes. This creates a more intelligent delivery rhythm in which judgment is applied where risk is high, and automation is used where work is repeatable and evidence can be checked.
In practice, digital FTEs execute by taking bounded delivery tasks from the POD backlog, applying approved prompts, workflow rules, knowledge repositories, and system integrations, and then producing a traceable output for human review or monitoring. For higher-risk work, Human-in-the-Loop approval is required before the action proceeds; for lower-risk and repeatable work, Human-on-the-Loop supervision allows the digital FTE to proceed within guardrails while exceptions are escalated. This execution model gives the POD speed and scale without disconnecting automation from ownership, evidence, and control.
The traditional service pyramid was designed for manual scale: many junior resources, fewer leads, and a small senior layer for governance. AI shifts the shape toward a flatter, more expert-led model. Routine documentation, code scanning, standard test generation, log summarization, duplicate-ticket detection, and first-level impact analysis can be handled by digital FTEs. Demand concentrates around experienced engineers, architects, product owners, domain specialists, reliability leads, data experts, prompt and knowledge engineers, and AI governance roles, with location decisions shaped by onshore, nearshore, and offshore requirements.
The pricing story changes when a POD combines human expertise with digital FTE capacity that produces output at speed. As reusable knowledge grows, automation coverage expands, and rework reduces, the same team can deliver more verified outcomes without a proportional increase in human staffing. Because digital FTEs can handle repeatable work at a lower unit cost than equivalent human effort, the cost per accepted deliverable falls over time. This creates a practical basis for year-on-year price reduction.
As PODs blend human expertise with digital FTE execution, contracts are moving beyond seat-based pricing. The contract now needs AI-specific clauses covering permitted use cases, data access, prompt governance, audit trails, IP ownership, security controls, human approval points, platform-cost treatment, and liability for AI-supported outputs. Commercial models can then combine a base POD fee for delivery continuity with output-based pricing for accepted deliverables and gain-share mechanisms where the provider demonstrates improvements. This gives clients greater price transparency and control, while allowing providers to be rewarded for verified performance improvement.
For example, an ADM contract can shift from paying only for a fixed support team to paying for a Human + Digital FTE POD that delivers measurable outcomes. The client may retain a predictable base fee to ensure capability continuity, add unit rates for completed change requests or validated test packs, and include a gain-share clause when AI automation reduces effort, improves cycle time, or prevents incidents without lowering quality.
The future POD will not be fully autonomous by default. Governance must define which activities can be executed by digital FTEs, which require Human-in-the-Loop approval, and which can operate under Human-on-the-Loop supervision. Architecture approval, personal data handling, regulatory interpretation, and model risk assessment should remain under direct human approval, while lower-risk operational tasks can gradually move to monitored digital FTE execution as controls mature.
Client proximity remains one of several design factors in the Human + Digital FTE POD, applied on a need basis rather than as a default constraint. Regulation sets the hard boundary around how far that flexibility extends: local laws, client data-residency rules, sector regulations, and privacy guidelines determine where digital FTEs can process data and where human reviewers must sit, while cross-geography access approvals govern how freely work can move between locations. This makes proximity a risk-driven choice—repeatable, low-sensitivity activities can run wherever data is masked, controls are embedded, and outputs are traceable. In contrast, regulated or production data often needs to remain within a controlled environment closer to the client’s operating context. A European insurance client, for instance, may allow digital FTEs to generate test scripts from synthetic data offshore, yet keep production-claims review and final exception decisions within an EU-based setup, matching flexibility to sensitivity.
For service providers and end clients, the Human + Digital FTE POD changes the relationship from a capacity transaction into a measurable value partnership. Providers need to prove that AI-enabled delivery improves output and cost efficiency, while clients need to define how those gains are accepted, governed, and commercially shared.
The future of IT services will not be determined by how many AI tools are deployed, but by how well each POD combines Digital FTE execution with human accountability, business context, and measurable client value.
By: Ankit Bohra, Senior Consultant, Avasant and Dipankar Dhariwal, Associate Director, Avasant
Avasant’s research and other publications are based on information from the best available sources and Avasant’s independent assessment and analysis at the time of publication. Avasant takes no responsibility and assumes no liability for any error/omission or the accuracy of information contained in its research publications. Avasant does not endorse any provider, product or service described in its RadarView™ publications or any other research publications that it makes available to its users, and does not advise users to select only those providers recognized in these publications. Avasant disclaims all warranties, expressed or implied, including any warranties of merchantability or fitness for a particular purpose. None of the graphics, descriptions, research, excerpts, samples or any other content provided in the report(s) or any of its research publications may be reprinted, reproduced, redistributed or used for any external commercial purpose without prior permission from Avasant, LLC. All rights are reserved by Avasant, LLC.
Login to get free content each month and build your personal library at Avasant.com