Tips to prevent GPT initiatives from failing

If strategy is part of your job description, then GPT must be part of your roadmap. Here are a few hints to prevent initiatives from failing.

1. Start projects with a maximum of 3 people who can gather the following 3 fundamental skills:

- A domain expert, someone who understands the output and can spot issues immediately.

- A prototyper, a builder who can prototype things in hours, if not in minutes.

- A hands-on and yet tangible visionary, someone with a clear idea of what processes should be changed and where.

These people should be assigned to deliver proof of concepts in weeks, with significant progression every day. They should show cross-checkable results that their approach is bringing: higher quality and/or faster outputs.

The timeline for this kind of project should be no more than 2 months.

2. If halfway through, results are promising, think change management. 

Some integrations on popular social networks can be useless since they bring no value or are unusable because the way they were integrated creates more issues than value: this is failed change management.

Change management must be realistic; if the vision is to change a core functionality that will have severe side effects, this is unlikely the right path. On the other hand, if it replaces hours of copy-pasting within a spreadsheet with a button to click, it is a no-brainer.

3. Remove all overhead. 

I have seen teams who preferred building software from scratch using hype programming languages rather than leveraging desktop tools such as Word or Excel (or Google's equivalent) and struggling to justify the time and cost involved.

It is much less understandable today than a decade ago. 

Start prototyping and even releasing products on top of existing ones. This is the only way to be ahead of the market.