Give me the annoying thing.
You bring one recurring workflow and access to the people who know it. I bring a concentrated build period to understand it, simplify it and build the better version.
The deal
You bring: one recurring workflow, examples of the real work, and access to the people who actually know it.
I bring: concentrated diagnosis, design and building.
We finish with: a working replacement your people have tried, plus documentation for how it works.
The build week starts once I have the access, examples and people we agreed I need. Waiting three days for a software login does not count as three days of building.
A deliverable, not an automation vibe.
The technology depends on the problem. The finish line does not: the workflow works better.
Show me yours →What happens
What makes something worth fixing?
Hours matter. So do dropped handoffs, mistakes, slow customer response, delayed work, missing information and the constant little investigations people conduct because the system can't answer a normal question.
Sometimes the payoff is simple: a salesperson spends more time selling, an estimator estimates, an administrator stops rebuilding the same report every Friday.
If I don't see a plausible return on a project starting at $5,000, I won't recommend it.
A good project has edges.
A lead arrives and needs to become a real sales opportunity. A signed customer needs to become an active project. Documents arrive and need to be read, named, filed, routed or acted on. A job finishes and needs to become an invoice. A weekly report needs to exist without a weekly scavenger hunt.
“This thing is driving everyone nuts” is enough to start.
What doesn't fit
A complete replacement of every system in your company. A vague request to “implement AI.” A project where I can't talk to the people doing the work. Something so mission-critical that a one-week intervention would be irresponsible. A business case whose only argument is eliminating somebody's job.
Questions I would have too.
What counts as one workflow?
Something with a recognizable beginning, end and outcome. A lead comes in and gets routed. A customer signs and gets onboarded. A job finishes and gets invoiced. If we need a conspiracy board to establish where it ends, it's probably too large.
Do you only work with AI?
No. AI is one tool. Sometimes a normal automation is better. Sometimes a database is better. Sometimes your process is the problem. I have no quota for how much AI I put in your company.
Are you trying to automate jobs?
No. I build for people to do their best work. If the entire case depends on eliminating someone's job, I'm probably the wrong person for it.
What software do you use?
Whatever earns its place: automation platforms, APIs, databases, AI models, Google Workspace, Notion, existing CRMs and business software, and custom code when custom code is warranted.
What if our problem is bigger?
Then I'll tell you. We can find a valuable piece with a clean finish line or scope something larger separately. I won't turn a bounded engagement into an accidental ERP implementation.
Do you provide ongoing support?
The engagement includes documentation and handoff. If ongoing work makes sense afterward, we can discuss it afterward. The first engagement does not require a retainer.
As little as one week, starting at $5,000.
Don't diagnose it for me. Tell me what happens.
Show me the workflow →