Customer & product adoption
Understanding needs, explaining options, and helping people move from uncertainty to confident use.
Daniel D’Jesus — Restaurant Technology, Customer Onboarding, Product Support & AI Workflows
Better context. Better decisions.
- Restaurant Technology
- Customer Onboarding
- Product Support
- AI Workflows
10+ years inside restaurants Creating systems that connect people, operations, and information.
Scroll to explore01 / The intersection
From service to systems
Technology works better
when you understand the people using it.
A decade on the floor reading intent, timing, and constraints under pressure. I now use that same judgment to shape restaurant technology and AI workflows.
Operational systems
Connecting people, timing, information, and constraints inside real-world environments.
AI-assisted building
Using agents, persistent context, and coding tools to research, plan, prototype, and document.
02 / Selected workIdeas, made tangible.
Two case studies.
Restaurant operations software · Agentic workflow
LINE.
The information exists.
The context is missing.
Ingredients in one place. Allergens in another. An 86 that hasn’t reached the floor.
LINE is a restaurant knowledge layer — the same dish data, for the person at the table and the person serving it.
- My role
- Product concept, discovery, workflow mapping, requirements, use cases, prototyping, positioning / GTM, and engineering collaboration.
- The build
- Architecture, GitHub, and code review with Carlos, a senior software engineer. AI-assisted research, specs, and coding with human review.
One question at the table.
Answer the question. Offer the solution. Keep service moving.
02 — Working prototype
See it in action.
One knowledge layer. Two views.
What the server needs during service.
Open full demo ↗LINE MENU HUB
Staff viewIllustrative full-service restaurant
Three fictional dishes and sample prices. Allergen filters are not food-safety clearance.
Tomahawk Ribeye
$200Sample price
Ingredients
45-day dry-aged tomahawk ribeye, salt. Grass-fed, grain-finished.
Allergens / risk
No dairy in the default preparation.
Shared grill. Large format; confirm doneness with the kitchen.
UnverifiedService notes
Sharing cut. Confirm doneness and rest time before sending.
Lobster L’Américaine
$48Sample price
Ingredients
Lobster, sauce Américaine, gnocchi, sunchoke, little gem.
Allergens / risk
Shellfish, dairy, gluten, egg.
Complex prep station. Confirm with the kitchen.
UnverifiedService notes
Half or whole. Confirm substitutions before describing the plate.
Mac n Cheese
$18Sample price
Ingredients
Pasta, béchamel, cheese.
Allergens / risk
Dairy, gluten.
Wheat in the sauce. Confirm with the kitchen.
UnverifiedService notes
86’d in this sample. Confirm availability before offering.
LINE MENU HUB
Guest viewSteakhouse
Public menu for the same three sample dishes. Kitchen notes stay in Staff view. Allergen filters are not food-safety clearance.
My allergies
No allergies selected
No dishes match these allergies. Clear the selection to see the menu.
Description
Ingredients
86’d — not available tonight
Explore the knowledge model
After the demo
The answer exists.
The server just can’t reach it
fast enough.
A guest asks about an allergen, a substitution, or a pairing.
The answer might live in a recipe, a training document, the POS, or a manager’s memory. The restaurant already has the knowledge.
The problem is getting the right context to the person at the table — without slowing down service.
LINE is how I’m exploring that problem.
Knowledge already exists
Menu & recipes
Ingredients, allergens, pairings
Training docs
Policies, procedures, guidance
POS data
Items, modifiers, availability
Manager knowledge
Experience, context, nuance
Context in
LINE
Answers out
Staff at the table
Fast. Confident. Accurate.
01 Built
Guest-facing menu + structured restaurant knowledge
02 My role
Product discovery · workflow design · GTM · prototyping
Engineering in collaboration with a senior software engineer
03 Next
The smallest authenticated staff experience that can answer: “What do I need to know right now?”
Technical notes ↓
Guest-facing fields and operational staff fields are separate in the data model. Unknown allergen information stays unverified rather than inferred. Operational staff information requires authenticated access. The current staff experience is still being validated; this portfolio uses local illustrative fixtures, not a live restaurant.
People · Context · Better service
02 — Context workflow
AGENTIC
WORKFLOW.
A personal OS for continuous thinking
A change of focus
shouldn’t mean
losing the thread.
A personal operating workflow that captures what happened, stores enough context to resume, and helps you decide one bounded next action.
- Memory
- Context
- Judgment
- Next action
Capture.
A loose thought from a day.
Get it out of your head. Capture ideas, decisions, questions, or observations before they fade.
Noticed a dip in onboarding completion after the copy change. Worth exploring.
Context.
Preserve what matters.
Store enough context to make sense of it later — related notes, assumptions, project state, and relevant history.
- Related contextThread around this note
- Previous decisionChanged onboarding copy (May 12)
- Key assumptionSimpler copy would improve conversion
- Related projectOnboarding Experience
- Referenceanalytics / notion / figma
Action.
One bounded next step.
Turn context into momentum. Choose one clear next action, optionally get agent help, and make it easy to return.
Analyze onboarding drop-off
Pull last 30 days of conversion data and compare cohorts before and after copy change.
Ask an agentDraft analysis, surface insights
Return pathBack to this context anytime
Preserve the thread. Reduce restart cost. Return with clarity.
Small captures.
A more continuous you.
How I use it
Private journaling. Better decisions. Less rework. A human–AI workflow for real life, not a perfect system. Examples above are reconstructed and sanitized — not a live vault, not published journal text. Optional agent help; human judgment stays in the loop. LINE is one project this workflow can support.
03 / Putting the thinking to work
How I would
approach the work.
I’m building toward software onboarding and support. These simulated exercises show how I would clarify the work, help people learn, and keep the next step clear.
From setup
to confident use.Restaurant software onboarding
Practice scenario: a restaurant is preparing to introduce a new staff knowledge tool. This is an approach exercise, not a completed customer implementation.
Discovery
Ask who uses the tool, which decisions it supports, and what a successful first week would look like. Confirm an owner, the current workflow, and staff availability.
Configuration
Map roles and permissions. Confirm locations, menu ownership, and required content. Log missing information and agree who will supply it.
Testing
Use sample content to test a routine update, an exception, and a role without permission. Review expected versus actual results with the owner and retest fixes before staff rely on the tool.
Staff training
Show a short task, let staff try it, and ask them to explain the next step. Provide a concise reference and a clear route for questions.
Go-live
Confirm the owner’s approval, access, readiness, and support contact. Agree a fallback to the existing workflow if service is disrupted.
First-week review
Ask where people hesitated or worked around the tool. Review unresolved issues with the owner and agree the next changes and follow-up date.
From an issue
to a clear next step.Structured product troubleshooting
Practice scenario: a manager reports that a menu update appears on one device but not another. No customer incident is being represented.
Receive & clarify
Acknowledge the impact. Ask which update, location, user role, and device are affected, when it started, and whether service can continue.
Reproduce
Record the exact steps and compare the affected view with a working one using appropriate access. Capture the observed result without collecting unnecessary guest data.
Separate symptom from cause
“The menu is wrong” is a symptom. Check the selected location, publish state, permissions, and freshness of the view before deciding why it happened.
Check documentation & configuration
Compare the behavior with documented publishing rules and known limitations. Use a test environment where available, check one explanation at a time, and preserve the original settings.
Resolve or escalate
Apply an authorized, documented fix and verify it with the user. If escalation is needed, include impact, reproduction steps, evidence, and attempted checks. Confirm who owns the issue and when the user will hear back.
Document & follow up
Summarize what was found, the resolution or current owner, and the next update. Add a reusable note when it could help prevent a repeat.
The aim in both exercises: a clear owner, a safe next step, and a useful follow-up.
04 / Where it startedSan Francisco, California
Before I thought about
software systems,
I worked inside
human ones.
On the floor, timing and judgment matter together: a guest’s expectations, the kitchen’s capacity, a reservation change, and what the team needs to know next.
Service taught me to explain the menu, set expectations, handle exceptions calmly, and coordinate across the room. POS and reservation tools are part of that work; people still need to interpret the situation.
I bring that floor judgment to learning software onboarding and support: listen carefully, check what is understood, and follow through. It is a foundation to build on as I learn the technical work.
People & timingProduct knowledgeTeam coordinationEnglish & Spanish
Atmospheric illustration; not a photograph of a résumé employer.
The question I keep coming back to
How can computers and AI give humans better context to make better decisions in the physical world?
Hospitality made this question visible: people act on information that is incomplete and keeps changing. AI interests me when it preserves useful context, reduces the effort of finding it, and leaves judgment with the person doing the work.
05 / The foundationDownload full resume
Experience,
in perspective.
Hospitality experience in San Francisco.
Product and AI projects alongside it.
LINE
Founder / Product & GTM Lead
Personal AI workflow
AI Workflow Builder
EPIC Steak
Server / Customer-Facing Sales
Bon Délire
Server / Customer-Facing Sales
La Mar Cebichería Peruana
Server / Customer-Facing Sales
Cañada College
Associate Degree · Interdisciplinary Studies
Social & Behavioral Sciences
06 / What’s nextOpen to the right conversation
I’m looking for teams building software for real people, real operations, and real decisions.
Let’s talk
Where I want to contribute
Customer onboarding · Implementation support · Product support · Product specialist · Restaurant technology · Customer experience · Selective consultative sales