A barbershop customer wants to know what a haircut will cost and how long it will take. The answer gets more complicated when you add a beard trim, extra services, and different pricing rules. A simple tool that lets customers select their services and see the total could solve that problem.
It probably doesn’t need its own booking system. The shop can buy one that already works.
That example, outlined in Amin Rabinia’s presentation, captures a useful lesson for business owners exploring AI: understand the job you need done before deciding what to build.
On November 4, Startup Folsom’s AI for Business workshop at the Roseville Venture Lab featured Rabinia, founder of Glissando AI. His presentation, “From Idea to Application,” focused on how to turn a business problem into a useful product.
AI has made it easier to get a working prototype on a screen. Deciding what belongs on that screen still takes judgment. Here are the practical takeaways from Amin’s presentation.
1. Your business already has ideas worth investigating
Amin encouraged business owners to look at what’s slowing them down. The opportunity might be a task someone repeats by hand every day, unanswered customer questions, or software that doesn’t fit the way the business operates.
Consider a team that copies information from customer emails into a spreadsheet, then uses that spreadsheet to prepare estimates. Before building anything, map out where the time goes and where mistakes happen. That gives you a specific problem to evaluate.
I think this is a useful starting point for entrepreneurs. You don’t need to begin with an ambitious idea for an AI startup. Begin with a problem you understand well enough to explain to another person.
2. Buy what’s proven and build what’s missing
One of Amin’s clearest recommendations was to use existing tools when they already do the job well. Booking calendars and payment systems were among his examples.
Custom software makes more sense when the available options don’t match your process, cost too much for what you actually use, or don’t address the problem at all.
Return to the barbershop example. The shop might benefit from a custom service menu that calculates price and appointment length using its particular rules. That doesn’t mean it needs to rebuild everything around the menu.
My suggestion is to identify the smallest part of your operation that needs something different. Keep the scope there until you have a reason to expand it.
3. Put a cost on the problem
Before spending money on a product, Amin suggested asking what the problem costs today. That could mean staff hours, money spent, or customers missed.
For example, if a task takes five hours each week, write down who performs it, what that time costs, and how much of the work a new tool could realistically remove. Include the time someone will spend checking its output.
That exercise helps separate an interesting experiment from a worthwhile business investment.
There’s also a potential startup opportunity: do other businesses have the same problem, and would they pay to solve it? Amin’s advice was straightforward: ask them.
An internal tool can become the starting point for a product. Customer conversations help establish whether that opportunity extends beyond your own business.
4. Separate the function from the features
Amin used a car to explain the distinction. Its function is getting someone from one place to another. Heated seats and an entertainment system are features around that function.
The same discipline applies to software. For the barbershop menu, the core function is helping customers understand price and time before they sit down. Extra features should earn their place by supporting that purpose.
With AI, it’s tempting to keep asking for additions because each one looks easy to generate. Before long, a simple tool has acquired enough features to qualify for its own organizational chart.
Write down the one job the product must do. Use that sentence to evaluate every proposed addition.
5. Write down the process before asking AI to build it
Amin’s presentation identified a common source of fragile prototypes: missing process details. Give AI a vague request, and it has to fill the gaps with guesses.
A business owner may know the rules so well that they forget to explain them. Which services can customers combine? When does an extra charge apply? What happens when a request falls outside the standard options?
Amin broke the process into four parts:
Steps: What happens, and in what order?
Choices: What decisions happen along the way?
Rules: What must always apply?
Exceptions: Which situations need different handling?
Writing those details down gives both AI and a human developer something concrete to work from. It also gives you a way to check whether the result behaves correctly.
6. Use the six Ps to explain what you’re building
Amin offered a product brief built around six Ps. Think of it as a set of instructions for the product before anyone starts building.
The six PsWhat to describeProfileYour business and the market it servesPeopleWho will use the product and whenPurposeThe one job the product must doPrioritiesWhat it must, should, could, and won’t includeProcessThe steps, decisions, rules, and exceptionsPayoffWhat a successful experience gives the user
The brief also helps diagnose problems. Wrong prices point back to the process. Unnecessary extras suggest a priorities problem. A product that looks right but misses the customer’s need calls for another look at its purpose.
That’s much more useful than repeatedly asking AI to “make it better” and hoping it guesses what you mean.
What to do before you start building
Personally, I would start with one recurring problem and write a short brief using the six Ps. Estimate what the problem costs, check the existing software options, and define how you’ll know a first version works.
Then test it against the situations your business actually encounters, including the awkward exceptions. A successful demonstration is a starting point. Your staff and customers need something that holds up during an ordinary workday.
Thank you to Amin Rabinia for sharing his approach, and to the City of Roseville for supporting the monthly AI for Business series at the Roseville Venture Lab.
Find upcoming workshops and opportunities to connect with local entrepreneurs on the
About Amin Rabinia
Amin Rabinia is an AI product builder in Sacramento and the founder of Glissando AI, where he designs and builds custom AI products for founders and business owners. He has spent six years in machine learning, from computer vision and prediction models to the multi-agent AI systems he builds today. His work includes an automated procurement platform, an AI design platform for a furniture company, and two live consumer products he built and runs himself. He holds three master's degrees in computer science and philosophy of science.

