AI applied without the right foundations accelerates failure as readily as it prevents it. Two thirds of transformation programmes in financial services will fall short this year. That number will not improve until the fundamentals do.
This piece sets out where AI is producing measurable impact in project management, where the risks of poor implementation sit and what it takes to get it right.
Accelerating Change Management
Change requests on a project are inevitable. What matters is how quickly the team identifying them, assesses their impact and adapts. Historically, change control processes have been slow, largely because the analysis required to evaluate downstream impact is time-consuming and manual.
AI tools integrated into project environments can now perform impact analysis in minutes. When a requirement changes, a well-configured AI layer can cross-reference the project schedule, the dependency map, the resource plan and the risk register to generate a first-cut impact assessment automatically. The project manager still makes the decision, but the inputs arrive faster and with greater coverage than a manual review would typically produce.
Consider a regulatory change programme at a mid-sized bank. A late amendment to reporting requirements lands six weeks before go-live. Manually, the team would spend two to three days mapping the downstream impact across workstreams. With AI-assisted analysis, that same assessment takes hours, giving the sponsor enough time to make a meaningful decision rather than simply react.
Automated Reporting
AI can pull live data from task management systems, financial trackers and risk logs, then generate draft status reports in seconds. The project manager's role shifts from data collector to editor. The practical benefit is consistency and frequency: when reporting is cheap to produce, issues surface earlier and sponsors have better visibility.
Requirements Gathering
This is where the research is clearest. PMI's 2024 Pulse of the Profession found that nearly half of unsuccessful projects fail to meet goals due to inaccurate requirements management. These failures compound each other and typically originate early, sitting behind scope creep, poor communication and inadequate sponsor support.
AI tools can support requirements workshops by transcribing conversations, extracting requirement statements and flagging ambiguity, prompting teams to define measurable acceptance criteria rather than accepting phrases like "the system should be fast." For example, a Business Analyst uploads the output of a discovery workshop to an AI tool, which categorises statements into functional requirements, non-functional requirements, assumptions and open questions. Two hours of post-workshop consolidation becomes thirty minutes.
Predictive Modelling for Risk and Resource Allocation
This is where AI's potential is significant but the practical reality, for most organisations, falls short of the marketing. Reliable predictive risk models require substantial historical project data, consistently structured. The majority of financial services firms do not yet have project records at that level of maturity as data is incomplete, terminology is inconsistent and records are scattered across systems.
Until that foundation exists, predictive modelling outputs should be treated with caution. Monte Carlo simulation and portfolio-level resource optimisation are genuinely valuable techniques, but feeding them poor data produces false confidence, which is more dangerous than no model at all. For most organisations, the priority should be building the data discipline now so that predictive tooling delivers on its promise in two to three years.
The Real Constraint: Data and People
The single biggest barrier to AI adoption in project management is data quality. Before investing in tooling, any Project Manager or PMO leader should audit what data is captured consistently, whether it is accurate, and whether project taxonomy is standardised across the organisation.
The second challenge is change management within the team. Project managers and analysts who feel their judgement is being replaced rather than supported will resist adoption. The framing matters, AI handles the data work; the project manager handles the decisions.
A Final Thought
The organisations seeing the greatest return from AI in project management are not necessarily those with the most sophisticated tools. They are the ones that adopted it on clean foundations, with teams that understood how to use it. The question worth asking is not which AI tools to buy, but whether the fundamentals are in place to use them well.
Delta Capita works with financial institutions and complex organisations to close the gap between AI ambition and delivery reality.
We support organisations in identifying where AI can genuinely accelerate their project outcomes, building the data and process foundations that make adoption sustainable and upskilling project teams to work effectively alongside AI-assisted workflows.