The most common mistake: starting with technology
Many superintelligence initiatives start with the wrong question: where can we use a language model? The useful question is different: which process costs us the most time, money or opportunity today, and how much of that cost can be removed?
When the starting point is a business problem with an owner and a metric, the conversation changes. You are no longer evaluating a demo; you are evaluating whether a specific workflow improves.
The five prioritization criteria
Score each candidate process from 1 to 5 on these criteria. The highest totals make good first cases.
- Volume: how often it happens each week. More repetitions mean more return for every improvement.
- Known rules: the team can explain how decisions are made. If nobody can describe the criteria, superintelligence will not learn them either.
- Available data: the information needed exists and is accessible, even if scattered.
- Cost of error: a failure can be detected and corrected without serious harm to the customer or company.
- Clear owner: a business person is accountable for the result and can make decisions.
Candidates that usually work well
Some workflows almost always meet all five criteria: answering frequent customer questions, qualifying and routing sales opportunities, preparing proposals from a catalog, extracting data from documents and reconciling information across systems.
They share something: they are repetitive, the team already knows how to do them well and the output can be reviewed before it reaches the customer.
Measure before you build
Without a baseline there is no way to prove improvement. Before the first sprint, record how long the process takes today, how many people are involved, how many cases are lost or delayed and what the error rate is.
Two weeks of measurement is enough in most cases. Those numbers become the project goal and the argument to scale.
Define what done means
A well-framed first case has a scope that fits in a few weeks, a quantitative success criterion and an explicit decision at the end: scale, adjust or stop.
That discipline protects the budget and builds internal confidence for the next process.