IT project success: A Computer Weekly Downtime Upload podcast

In spite of what seems like the best laid out plans, IT projects are prone to over-runs and sometimes failure. Jason Olkowski, chief strategy officer at Creatio recalls that in the past, IT projects fell over causing major issues in the business such as the inability to ship a product because it was unaware of what was in the supply chain and inventory. But he says: “These things are now fewer and far between.” However, he says 75% of all IT projects often fail to deliver a business impact. Unpacking the reason behind these failures, he says: “You go live with something, data's moving and you are monitoring the system with dashboards, but [the application] never really solves the needs of the business.”

In his experience, in cases where IT projects fail to deliver a business impact, often it is because the workflows are not right. What this means, according to Olkowski, is that over time, work starts to happen outside of the IT systems that the business has spent time and money implementing.

He says: “I'm sure you've seen that? You have the temporary workaround that becomes permanent, such as the spreadsheet that people start to rely on.” For Olkowski, such workarounds represent indicators of failure within an organisation, where work eventually happens outside the IT system.

Why this happens, according to Olkowski is because project teams get bogged down in the implementation. He says they spend so long iterating requirements with a group of people that they end up straying from the original vision. In addition to losing sight of the core project focus, he says: “You can't keep pace and speed with what's happening and changing with the business while you're doing all of this heavy lifting.”

Olkowski recommends having the right people engaged in the project steering process as well as strong governance for the implementation process. But even with the right mix of people on the project team and strong governance, he says: “One of the things I see is that only when people actually see it and they get their hands on what it is that you're going to deliver, do they suddenly realise that they need something that's a little different, which doesn't actually mirror exactly how the they work.”

He urges business IT teams to work towards delivery of a minimum viable product, which initially is focused on tackling the biggest pain point in the business. This, he says, offers the best opportunity to deliver a business impact. Starting with this base, he recommends that the project team should focus on iterating and adding continuous value to the business over time.

While the big bang or waterfall approach to delivery an IT project in one go has largely been superceded by an agile approach to project delivery,  Olkowski says: “One of the things that a lot of us still see in the industry is that the initial implementation is frankly a bit  like Wagile - a bit of waterfall because you have to have some set of scope and requirements and a goal that you want to meet, then agile iterations.”

After the first phase of the project implementation he says an agile approach then takes over, iterating around cycles of business requirements and delivery.

Even when adopting an agile methodology, Olkowski  urges IT teams to consider delivering prototypes to business users. “You can prototype in a way that allows users to touch and feel [the application] and see how the workflow and user interface works for their teams and their business,” he adds.

As businesses start reengineering workflows to take advantage of artificial intelligence (AI) and the use of agentic AI, Olkowski believes prototyping can help the project team and project sponsors uncover gaps that might not have ordinarily been uncovered until after the project goes live, where the gaps then impact the business.

He urges IT and business leaders to use agentic AI pilots to understand which workflows make sense from a business perspective, and which ones offer the biggest opportunity to achieve the goal of enabling the human and the AI agent to work together. The challenge he sees is that project sponsors may be tempted to move beyond the minimum viable product for an agentic AI workflow. “I think that one of the traps people fall into is maybe going too large with their scope,” he says.