An AI consultant in Orange County works out which of your workflows is worth automating, proves the dollar value, then builds the thing. Portillo Technologies is an AI consulting, organization-optimization, and custom software practice in Orange County, California. We work out how a company should run, then build the software that runs it.
What does an AI consultant in Orange County actually do?
Most of the job is subtraction. A company shows up with twelve things AI could theoretically do for it. The work is picking the one workflow where the math is biggest and the test for "did it work" is writable, then shelving the other eleven.
After that it turns into ordinary engineering. Where does the data live. Who touches it. What does a correct answer look like, and how will anyone know we got one.
Here the consulting half and the building half are the same person. That's deliberate. A recommendation that can't survive contact with your codebase isn't worth the deck it's printed on.
Companies in Irvine, Newport Beach, Costa Mesa, and Santa Ana get an in-person version of this. Everyone else gets the identical thing over video.
Why do most enterprise AI pilots stall?
Because the pilot was never tied to a number. MIT's NANDA report found that 95% of enterprise AI pilots deliver no measurable P&L return, and the Institute of Project Management puts 88% of them as never reaching production at all. Those are mostly not technology failures. A model that drafts contract summaries works fine in a demo.
It stalls because nobody wrote down what "working" meant in dollars before the demo, so there is nothing to point at when the budget meeting arrives. Gartner attributes 85% of failed AI projects to data quality. IDC finds 65% of companies name data readiness as the blocker. Both are boring problems, and boring problems get solved before you touch a model.
The fix is to shrink the pilot until one person's week visibly changes. Read the AI Pilot Autopsy if yours is already stuck.
What does it cost to work with Portillo Technologies?
Builds are priced at 10 to 20 percent of the measured annual value, never by the hour. That number comes out of discovery, so by the time anyone quotes a build, both sides are looking at the same arithmetic. Every build carries the same promise: we work it until it passes your acceptance test.
| Engagement | Price | What it produces | Guarantee |
|---|---|---|---|
| AI Pilot Autopsy | Free | Two pages on why your pilot stalled and the two changes that unstick it, as a PDF attached to an email reply | Never gated, nothing to sign up for, and no follow-up sequence |
| Discovery | $2,500, about two weeks | The quantified annual dollar value of your best workflow, plus the written acceptance test for it | If discovery doesn't surface at least 10x the fee in quantified annual value, you don't pay |
| Custom web app or line-of-business system | MVP $15k–75k, complex $250k+ | A working system your team uses on Monday | We work it until it passes your acceptance test |
| AI / LLM integration | Strategy or pilot $5k–25k, mid-size $25k–100k | A model wired into a real workflow with an evaluation harness around it | We work it until it passes your acceptance test |
| RAG / knowledge copilot | $8k–45k build, plus $650–1,750/month operations | Answers drawn from your own documents, with citations back to the source | We work it until it passes your acceptance test |
| Fractional AI-CTO / technical advisory | $4k–15k/month | A senior technical read on your roadmap, vendors, and hiring | Cancel any month |
| AIOps / automation retainer | $2k–15k/month | Ongoing operation and improvement of what's already running | Cancel any month |
You own the code outright. It is never resold, and there is no license to renew.
How does an engagement start?
Five steps, in this order, and step one costs nothing.
- Autopsy. Email me what stalled. You get two pages back on why, and the two changes that unstick it. Free and ungated. If you don't reply, I don't chase you.
- Discovery. $2,500, about two weeks. I sit with the workflow, count what it actually costs you per year, and write down the number.
- Acceptance test. Before any code exists, we agree in writing on what "done" looks like. Specific inputs. Specific outputs. A threshold you can check without me in the room.
- Build. Priced at 10 to 20 percent of that measured annual value. It runs until it passes the test we already agreed on.
- Own it. The repository, the infrastructure, and the credentials are yours. Keep me for operations or don't.
Any step can be the last one. Stopping after discovery and building it yourself is a fine outcome, and one reason the documents are written to be handed to somebody else.
Who is doing the work?
Jason Portillo. Sole principal, and the person who writes the code. There's no bench behind me and no junior consultant who inherits your account after the sale.
That means one build at a time. If a build is already running, you get a start date instead of a slot, and I'll tell you that on the first call rather than after you've signed something. The cap is a qualifier, not an apology. It's what keeps the attention on one thing.
The person who maps the workflow is the person who writes the code. Nothing gets lost in a handoff document because there is no handoff. If you want a team of consultants and a separate delivery organization, that firm exists and it isn't this one.
More on how I work: the FAQ.
How is this different from a Big-4 firm, an offshore shop, or hiring in-house?
Each of these is genuinely good at something. The question is what it costs you when the fit is wrong.
| Alternative | Good at | Where it breaks | The number |
|---|---|---|---|
| Big-4 or large consultancy | Board-level air cover, change management across thousands of employees, regulated-industry process | Strategy and delivery sit in different buildings, so the deck arrives and the software doesn't | MIT's NANDA report: 95% of enterprise AI pilots deliver no measurable P&L return |
| Offshore development shop | Headcount, low posted hourly rates, running a spec somebody else already wrote | You still have to write the spec, and the rework lands back on your team | SmartDev puts total cost of ownership at 25–150% over the headline rate, and QA rework at 15–26% of hours |
| Hiring an in-house AI team | Permanent capability, deep context, work that never leaves the building | The ramp is longer than most first projects, and you're hiring for skills you can't yet evaluate | GroovyWeb puts it at $1M–1.8M loaded in year one, on a 3–6 month ramp |
| No-code or low-code platform | Fast internal tools, forms, dashboards, and anything a business analyst can maintain | The ceiling arrives suddenly, usually the first time you need real logic or real volume | The Bright Byte puts 25–30% of no-code builds as rewritten in custom code within two years |
Where this practice fits: one senior person, a quantified target, and code you own. If the job really is ten engineers, take the job to a firm with ten engineers. If it's custom software that has to be right, that's this.
Do you work outside Orange County?
Yes. The practice is based in Orange County, California and the work is remote-friendly across the United States. Most of a build happens in a repository and a video call, and that part doesn't care where you sit.
What being local buys you is the first two weeks. Discovery goes faster when I can stand in the room while somebody does the job I'm about to measure, and companies in Irvine, Tustin, Anaheim, and Huntington Beach get that at no extra cost. Out of state, the arrangement is one on-site week and the rest remote.
Industry matters less than you'd think. The pattern is the same whether the workflow lives in a claims desk, a shop floor, or a clinic. Worth noting: healthcare has the lowest pilot-to-production rate of any sector, at 8%, so scope there gets cut harder.
Who this is not for
- Companies shopping for the cheapest bid. There is always someone cheaper. Pricing runs off measured value, and if the value isn't there I'd rather say so during discovery than win the work.
- Anyone who wants a body shop. Staff augmentation, seats billed by the hour, a developer who joins your standup and takes tickets. Not this. More on that here.
- Teams committed to a broad pilot. If the mandate is "AI across the company" and shrinking it to one workflow is off the table, we will disagree in week one and you'll have paid for the privilege.
- Anyone who needs ten people starting Monday. One builder, one build at a time. That is the whole capacity, and no amount of budget changes it.
Common questions
Is the AI Pilot Autopsy actually free?
Yes. Two pages, written about your company, sent as a PDF attached to an email reply. There is no form, no signup, and no drip sequence behind it. If you read it and never write back, you never hear from me again. That is the whole arrangement.
What happens if discovery doesn't find enough value?
If discovery doesn't surface at least 10x the fee in quantified annual value, you don't pay. You keep the documents either way, and you can take them to any other firm and have them build it. The $2,500 buys a decision, not a dependency.
Who owns the code?
You do, outright. It is never resold to another company and there is no license that expires. The repository, the infrastructure, and the credentials are yours from the first commit, so you can hire anyone else to maintain it without asking permission.
How many builds do you run at once?
One. One builder, one build at a time. If a build is already running, you get a start date instead of a slot, and I will say so on the first call. The cap is fixed, which is the main reason to talk early rather than late.
Do you charge by the hour?
No. Builds are priced at 10 to 20 percent of the measured annual value the discovery documented. Hourly billing pays me to be slow and pays you to argue about timesheets. A fixed number against a written acceptance test puts the risk where it belongs.
Do I need clean data before we start?
No, but you need to know how dirty it is. Gartner attributes 85% of failed AI projects to data quality and IDC finds 65% of companies name data readiness as the blocker. Discovery looks at the actual records early, because a data problem found in week one is cheap and the same problem found in month four is not.
Sources
- MIT NANDA report (2025), 95% of enterprise AI pilots deliver no measurable P&L return — softwareseni.com
- 88% of enterprise AI pilots never reach production — institutepm.com
- Gartner 85% data-quality failure rate; IDC 65% data readiness — pertamapartners.com
- Healthcare pilot-to-production rate of 8% — softwareseni.com
- No-code builds rewritten in custom code within two years, 25–30% — thebrightbyte.com
- Offshore total cost of ownership 25–150% over headline rate; QA rework 15–26% of hours — smartdev.com
- In-house AI team at $1M–1.8M loaded in year one, 3–6 month ramp — groovyweb.co
Send me the pilot that stalled
Two pages back on why it stalled and the two changes that unstick it. Free, no form, no follow-up sequence. If discovery is the obvious next step, we'll talk about it. If it isn't, you keep the pages.