The weekly report
Someone pulls numbers out of the POS, a spreadsheet, and two inboxes every week and retypes them.
The report assembles itself in the format you already read, checked by a person for the first few weeks before it runs on a schedule.
For restaurants, hotels, and travel brands
Every hospitality business runs on software. None of it was built just for you.
I help you close the gap between what your software does and what your business actually needs. Sometimes that's a quick fix with AI. Sometimes it's connecting the tools you already pay for. Sometimes it's building something new. Either way, it starts with a conversation, not a proposal.
A few of the companies I've worked with
The Problem
Why your POS, booking, and marketing tools don't work together
Your tools are good. They just don't know your business. Each one does its own job well, but none of them was built to talk to the others, so the gaps between them get filled by hand, usually by you.
What you're probably running
Square
Toast
Clover
GoTab
Popmenu
SpotOn
Lightspeed
DoorDash
UberEats
Postmates
Stripe
QuickBooks
OpenTable
Resy
Tock
Cloudbeds
Mews
MarginEdge
Craftable
Restaurant365
MarketMan
Mailchimp
Constant Contact
Klaviyo
Buffer
Google Sheets
Slack
Twilio
Gmail
Zapier
Anything with an API or an inbox
These are good tools. I use several of them myself. But each one is built for ten thousand businesses, and you only have the one.
Where it starts to hurt
- The Friday report that takes 20 minutes and three inboxes.
- The Saturday that double-books because the reservation system and the event calendar don't talk.
- The loyalty program your POS almost supports.
- The PMS export nobody opens because it lands as a CSV.
- Menu costing in a spreadsheet nobody trusts.
- Guest data in five systems, none of them the one you're looking at.
- The feature your vendor says is on the roadmap. It has been on the roadmap for two years.
None of it is a crisis. All of it adds up.
How I Help
What AI and automation can take off a restaurant's plate
You're not getting an agent or an app. You're getting a Friday afternoon back: the weekly report, bookings that stay in sync, guest follow-up, online ordering, and costing, handled without the copying and pasting.
Owners and operators lose hours every week wrestling with apps to get the numbers they need to make a decision. Sales are in one system, labor in another, inventory in a third, and the answer you need is somewhere in between. So you export, retype, and guess, and the week goes to working in the business instead of on it.
I make that work go away. You get clear numbers without the digging, routine jobs that run themselves, and time to think about where the business goes next. The same jobs come up again and again. Two of the ones below have a finished, running example on this page. The rest show the shape the fix usually takes, not a promise about yours.
Bookings and calendars that don't talk
The reservation system, the events calendar, and the room block each think they own Saturday.
The calendars are kept in step, and a person gets a text when they disagree. Nothing writes to a reservation on its own, and if an integration fails, it fails loud: you get told, and you fall back to the manual process you already know.
Guest follow-up and marketing
Three tools, one person playing designer, copywriter, and scheduler, and a brand that drifts.
Posts, emails, and texts in minutes, on brand, from one place. This is what MessageKit does.
Direct online ordering
A marketplace takes a cut of every order and keeps the customer's contact.
Your own storefront on your own payments, with the guest's details staying yours. This is PizzaCart.
Costing and inventory
Menu costing lives in a spreadsheet nobody trusts, and the count never matches the order.
Costs come from the systems you already run, and what doesn't add up gets flagged, with a person checking it before anyone acts on it.
Not sure which one is yours?
Every business has a job its software won't do that isn't on anyone's list.
That's what the Software Gut Check is for.
How we can work together
How much does it cost to work with me?
A conversation first. The first 20-minute phone call is free, and it's how we figure out whether we should work together. After that, a Software Gut Check is $500, Advisory is $500 a month, and custom builds start around $2,500, with a fixed price agreed before work starts.
A friendly conversation about what would help you most. You'll leave knowing whether we're a fit, and whether it's worth doing at all.
Based on our initial conversation, we might decide the right answer is simply advice. Or we might decide to build something. If we do, it'll be one of three things, and I always start with the simplest one that works.
- No pitch, no deck, no homework
- If a $30 subscription already does the job, I'll tell you. That advice is free.
You're probably paying for more than you use. In a couple of hours with you and your tools, I find what your current software can already do that nobody turned on, what it will never do, and what's not worth touching. You get a short written report: keep this, fix that, drop the other, in that order.
Sometimes the answer is that your software already does what you want. You just didn't know, or it wasn't switched on. That's a good outcome, and it's still $500.
Book a Gut Check Starts with the free call.A person to call before you sign anything. Vendor demos, contracts, integrations, and AI pitches, read by someone who has built the other side. Month to month.
Talk about Advisory Services Starts with the free call.An agent, a workflow, or an app. The price is based on the value to your business, not my hours, and it's agreed before work starts. You'll never get an invoice you didn't see coming.
As a guide, not a quote: custom builds start around $2,500. Something the size of MessageKit would be around $30,000. You'll have a written scope and a fixed price before you decide, and it's fine to say no.
Then we build what closes the gap.
-
Agent
“Someone does this by hand every week.”
I set up an AI agent around your real examples for the job someone does by hand every week: the Friday report, the caption, the purchasing list, the follow-up. A person still reads the result before it goes anywhere.
Days, not weeks. Priced on what it's worth to you, agreed before work starts. -
Workflow
“It works, but only when someone remembers to run it.”
I connect the tools you already pay for so nobody has to carry data between them by hand. Orders pulled on a schedule, run through the agent, dropped where they belong, and a text to someone if anything looks off. I test it against real cases before it runs on its own.
Two to four weeks. Priced on what it's worth to you, agreed before work starts. -
App
“More than one person needs it, and the records have to last.”
When nothing off the shelf does the job, I build it. Ordering, operations, guest data, tools for the trade. Logins for your staff, records that last, and connections a workflow tool can't guarantee. You own the code.
Weeks, not quarters. Priced on what it's worth to you, agreed before work starts.
What to Expect
You own what I build for you
For custom work, the code, accounts, API keys, and documentation are in your name from day one. If I disappeared tomorrow, any competent developer could pick it up. If the answer is one of my own products instead, like MessageKit, that's a subscription: I host and maintain it, and your data and your accounts stay yours to take with you.
Your team is trained
Before anything goes live during service, the people who'll run it know what it does, what it looks like when it's wrong, and what to do about it without calling me.
Plain-English documentation
Every step written down in words, not code. What runs, when, and where the results land.
Support when you want it
A support and maintenance agreement is available for a fee. It puts the response times in writing, including whether evenings and weekends are covered, and it covers things like a vendor changing an integration. Without one, you're not stranded: the documentation and your team cover the ordinary day.
Examples
Case study: one app for email, SMS, and social across two restaurant brands
Case study
MessageKit
The mess
At Lucky Hare and Night Train Pizza, a Tuesday special had to go out on Instagram, Facebook, and the email list before dinner service, looking like the same restaurant each time. One person did the job of designer, copywriter, and scheduler across three tools: build the image in Canva, upload it to Buffer, write the caption by hand. A post took about twenty minutes. An email took up to an hour. It broke down as often as it worked.
The systems
Square for the point of sale and the customer list. Canva for images. Buffer for scheduling. Square wanted $125 a month for email and text marketing alone; Klaviyo was closer to $500.
What I did
I climbed the ladder in order. First, a saved prompt that drafted a caption from a photo and a few details. Then a process anyone on staff could run the same way. Once both brands were running the same four steps, I built it into a tool: a content library that turns out branded social posts, emails, and texts. That tool is MessageKit, and other restaurants use it now.
What changed
- Under five minutes to publish a post, down from about twenty.
- Under ten minutes to send an email, down from as long as an hour.
- Under five minutes for a text, down from about twenty.
- Staff run it, not the owner. That's the part that made it stick.
The math
- Before
- Square marketing $125 + Canva $30 + about 17 hours a month of staff time at $20 an hour ($345). About $500 a month.
- After
- MessageKit $79 + about 4 hours a month of staff time ($80). About $160 a month.
- Saved
- About $340 a month, and 13 hours. Six posts, one email, and three texts a week, at the old and new task times.
Also built for my own businesses
Donna
Restaurant operations
The problem. The legacy tools were overly complicated. Recipes and ingredients were a chore to track, and ordering and receiving never went smoothly. Nobody on the team wanted to use them.
What I built. Donna listens to Square sales and adds the reporting the POS doesn't give you. Ordering, receiving, and inventory tracking are simple, and now it tracks tasks too.
After. It's just easier to use, so the team actually uses it. $99 a month. For comparison, Craftable lists at $450 and MarginEdge at $600.
donnasays.com →
PizzaCart
Online ordering
The problem. Third-party ordering took a cut of every pie and kept the customer. Our own site couldn't take an order.
What I built. Our own storefront with Square payments and live order tracking. We keep the margin and the guest.
After. No marketplace commission on orders through the storefront, and every guest's contact stays ours.
pizzacart.app →
Winote
Wine trade
The problem. I taste 40 wines at a sitting and was logging them in Apple Notes. Every wine app is built for collectors, not people who work in wine.
What I built. Structured tasting notes, a catalog built from public state records, and a version for distributors.
After. In daily use by working sommeliers, with a trade version for distributors.
winote.app →Real results from using custom apps in my own businesses
$340/mo savedon email and SMS marketing software and staff time
13 hrs a month backon creating and sending social posts, emails, and texts
79% fewer 86sfrom better PAR visibility and just-in-time ordering
$1,170/mo COGS savedon wasted product due to ordering and receiving errors
97% increasein opening/closing checklist compliance across the entire team
“J has helped us deliver amazing value to our customers time and time again. His ability to take nebulous concepts and drill down until they become quality products is second to none.”
“J shows you how to ask the right questions, dig into the core of the challenge, and drive towards results-based solutions.”
“For J, nothing is impossible. He knows how to get things done, even on ridiculously short notice. A great eye for small details but never forgets the high-level strategy.”
“The concepts he teaches can be applied to any type of product or business idea. A clear and concise way to make sense of the chaos and build something people will actually buy.”
Fit
Who I work with, and who I don't
People who know their business and are tired of fighting with software.
Travel and hospitality brands between $1MM and $20MM in revenue
- Independent and multi-unit restaurants
- Boutique hotels, inns, and resorts on Cloudbeds, Mews, or Opera
- Tour operators and travel brands with a booking platform and a spreadsheet behind it
- Wine and food distributors, importers, and brands going direct
- Local and regional hospitality groups
The hotel and travel work earlier in my career was for large brands. At your size the jobs are the same ones on this page: the report, the double booking, the follow-up, the order that should have been yours.
Who I'm probably not right for
- Big companies that need an RFP, a procurement portal, and a security questionnaire before we can talk.
- Anyone who wants a dev shop to build the next marketplace or super-app.
- A problem a $30-a-month subscription already solves. I'll tell you that on the first call, for free.
Hi, I'm J.
For most of my career I helped giant companies solve their software problems: Delta, IHG, UPS, Coca-Cola. Then I opened a wine bar, then a restaurant, then added a second brand to it, and found out firsthand how badly operators are served by the tools they pay for.
Now I run SparrowSoft, an AI-driven software company, and I own and operate a dual-concept restaurant, Lucky Hare and Night Train Pizza. I'd rather work with the person who owns the problem than a committee that manages it. AI is what makes that possible: work that once took a team and a six-figure budget, one experienced person can now do for a fraction of that.
I'm the person you call when the software you already have stops one feature short.
More about me →Questions I get asked first
Do I have to replace my POS or PMS?
No. I'd much rather keep what works and build the missing piece around it. If it has an API, an export, or an inbox, I can connect it.
How is this different from hiring a dev shop?
A dev shop builds what you tell it to build. The expensive mistakes happen in the telling. I'm an advisor first: I start with the job, not the spec, and about half the time the answer is a small workflow, not an app. A shop has no reason to say that. I do, because I'd rather you come back.
Who owns what you build?
If I build it for you, you do: code, accounts, API keys, documentation. If I got hit by a bus tomorrow, your tool would keep running and any competent developer could pick it up. If the right answer is one of my products, like MessageKit or PizzaCart, that's a subscription I host and maintain, and your data stays yours.
What does it actually cost?
The first call is free. A Software Gut Check is $500, one time, and you get a written report on what your current tools can already do, what they can't, and what to do about it. Advisory is $500 a month for about four hours of my time on your software and technology questions. Anything I build is priced on what it's worth to your business, agreed before work starts. As a guide, custom builds start around $2,500, and something the size of MessageKit would be around $30,000. I don't quote anything until we've talked.
Do you work with hotels and travel companies, or just restaurants?
Both. Earlier in my career I did product work for an airline, a global hotel group, and a travel technology company. That work is under NDA, so the specifics stay private. The restaurant is mine, and the numbers on this page come from it.
What happens if it breaks at 7 p.m. on a Saturday?
Everything I build comes with plain-English documentation, and I train your team on what to do if something goes wrong during service. If you want me to keep looking after it, a support and maintenance agreement is available for a fee. It puts the response times in writing, including whether evenings and weekends are covered, so you know what you're paying for before launch. Nothing depends on one person's laptop, including mine.
Can I try this myself before calling you?
Yes, and you should. Download a starter prompt. It interviews you one question at a time and writes a one-page brief you could hand to me, another developer, or another AI. Free, no email required.
Tell me the one thing your software won't do.
Twenty minutes. You'll leave knowing whether we're a fit and whether it's worth doing at all.