· Solveion · Perspectives · 4 min read
The best AI projects are boring
The AI work that pays for itself rarely looks like the keynote. It looks like an unglamorous task, done thousands of times a year, finally taken seriously. A case for choosing the dull project over the dazzling one.
There is a moment in AI strategy discussions that we have learned to watch for. Someone senior, often sincerely, proposes the visionary project: the AI that will reinvent the customer experience or leapfrog the competition.
Meanwhile, three floors down, a team spends every Thursday afternoon manually reconciling two systems that don’t talk to each other. A shared inbox absorbs four hours of skilled attention a day, answering questions whose answers already exist in a document nobody can find.
We have come to believe that the second category is where AI earns its keep. The instinct to reach past it toward the visionary project may be the most expensive habit in technology adoption today.
The case for boring
Boring problems are well defined. The Thursday reconciliation has inputs you can point to, an output everyone agrees on, and years of examples of the job done correctly. That definition is exactly what an AI system needs in order to be built and, more importantly, evaluated. Visionary projects tend to have success criteria nobody can pin down, which means nobody can ever say whether they are working. A project that can’t be shown to work eventually gets treated as if it doesn’t.
Boring problems repeat. The economics of AI are the economics of repetition: effort invested once, value collected on every occurrence. A task performed two thousand times a year justifies real care in building and maintaining a system around it. A dazzling capability used twice a quarter does not, however good it looks on a slide.
Boring problems are forgiving in a useful way. They usually sit inside a process where a human already checks the work, so the AI can start as a drafting assistant instead of an unsupervised decision-maker. That is the configuration in which trust gets built honestly. People watch the system be right, case after case, and there is an easy way to catch it when it slips.
And boring problems have grateful owners. Nobody’s identity is wrapped up in the reconciliation task. Automate the keynote-worthy thing and you may discover you have wandered into someone’s sense of self. Automate Thursday afternoons and you have made an ally who will champion the next project.
The part nobody budgets for
The strongest argument for starting dull has less to do with the task than with what it does to the organization.
A small, successful AI system teaches a team things no workshop can. What these tools feel like when they’re right. What their failures look like. How to phrase things. When to double-check. It also creates internal proof that the technology works here, on our problems, which changes the quality of every idea that follows. The second project gets proposed by someone who has used the first one, and that second proposal is nearly always better chosen than anything in the original strategy deck.
Organizations that open with the moonshot get the opposite effect. A stalled flagship, a quietly disillusioned team, and a reputation that outlives the project by years: we tried AI, it didn’t really work.
A filter for choosing where to start
If you’re picking a first project, we’d suggest a deliberately unromantic test. Look for work that comes up at least weekly, that a competent new hire could learn from existing examples, that someone can check quickly, and that the people currently doing it would happily give up.
Work passing all four parts of that test is rarely scarce. Most organizations have a dozen candidates hiding in plain sight, and any one of them, done well, will beat the visionary project on every measure that survives a finance review.
The irony is that the boring path is how the visionary outcomes actually arrive. A business that has spent two years steadily automating its repetitive cognition, building up skills and evaluation habits and trust along the way, is the one positioned to attempt something genuinely ambitious when the moment fits. The keynote project turns out to be downstream of the boring ones.