It is good practice that for large business IT projects plans are made. Plans are made to control cost, resources and also to steer design and realization. Most of the time this is where things get complex. So you get complexity by design.
Most business IT plans are only reviewed and approved by direct stakeholders. E.g. directors, senior managers and a lead architect. Real critical questions and critical reviews to prevent IT disasters are filtered out. Too often engineers and software developers who must create the new IT systems and software systems are not involved in creating these often too detailed plans. Especially when IT development is done by other companies, early involvement of the right people who actually do the realization is exceptional.
When plans for new business IT projects are created without involving the real hands-on expertise of people who need to execute your plan into reality you hit the survivorship bias (https://en.wikipedia.org/wiki/Survivorship_bias ).
In the IT industry project plans for building complex systems are often only approved by senior managers and directors. Trivial risks are mentioned but in general no red flag is raised.
When launching a new large business IT project you should be aware that survivorship bias is present. Intentionally but often unintentionally. Direct stakeholders are often overly optimistic and failures from earlier experiences are ignored. So to prevent complexity problems when starting a new large IT project: put extra effort in the preliminary project phase.
Are really all stakeholders and subject matter experts involved when setting the project goals without taking into account formal roles and functions?