LIVE CASE STUDIES
Ten live systems, one standard. Webshop, restaurant kitchen, construction ERP, fleet, display network, project management, SEO portal, asset management, an auto workshop and a company operating system. Where operations run, the system runs — live.
Plantapont · multi-store ecommerce
Plantapont is a multi-store packaging-free retail network where customers see the same stock and prices online and in-store. The full digital stack — webshop, central admin and store till — runs on one database with live traffic and day-to-day operations.
AI supports product and blog optimization behind the scenes; the brand is visible in search and AI answers. Real AI agents can already search the catalogue, request pricing and place orders. Multi-channel delivery: courier, pickup points and in-store collection share the same inventory logic.
One click from the editor to search-ready product and blog copy: suggestions, preview, archive — less manual work, faster publishing.
We build the site so ChatGPT, Perplexity and Gemini can cite your store, products and prices as a trusted source — not your competitor.
Real AI agents can connect: they search the catalogue, request pricing, and place orders — like a trusted assistant shopping for you.
Not just keywords: it understands shopper intent and leads people to where revenue happens — your own webshop.
Complex multi-channel system: web + store + admin from one team
Real stores, real stock and orders — day-to-day business operations, not a showcase page
Measurable presence in search and AI answers — the brand shows up where buyers ask
The same capability we bring next to ERP, into manufacturing and B2B trade
Da Mauro · restaurant / pizzeria · kitchen ops
Da Mauro’s full digital backbone is ours: guest ordering, admin, and counter–kitchen–handoff–courier operations. Where pizza bakes, there is no room for half-finished software — and AI agents can place orders too.
Guests order online for delivery or pickup with scheduled slots and reliable pricing. The kitchen runs a three-column bake queue; the daily menu follows a separate lane. Counter, oven and courier are traceable in one system — daily operations, not a demo screen. AI assistants work from a live menu and opening hours, and can place real orders today.
Confirm incoming orders on a tablet. Large buttons, live refresh — whoever is there acts immediately.
Three-column bake queue, daily summary (“needed today”), printable tickets — chefs see what is on the fire.
Ready pizzas per order, daily menu by quantity — handoff → out for delivery → delivered. Traceable end to end.
Manager / kitchen / counter / courier: everyone sees only their job. Fast login, designed for phone and tablet.
Menu, prices, opening hours, location — AI assistants treat the brand as a restaurant, not “some webpage”.
AI agents work from a live menu, opening hours and timeslots — they don’t guess from yesterday’s screenshots.
A real AI agent can place an order today; the guest pays on site. Not a demo — the order lands in the kitchen queue.
Live opening hours, clear guidance, validated cart — ordering is a reliable process, not hope.
They see the kitchen actually runs on it — not a showcase screen
Counter → oven → courier tracked end-to-end in one system
Guest site, kitchen operations and AI agent ordering from one team
The same operational rigor we bring to manufacturing, logistics and service businesses
Gaia · construction ERP + public sales portal
Gaia is not “CRM plus Excel”. A full construction ERP and public sales portal share the same data. Web enquiries, quotes, schedules, AI invoice reading and marketing results meet in one system — measurable and traceable end to end.
The internal ERP manages the full project lifecycle: plots, house plans, schedule, materials, finance, subcontractors, handover. On the public portal buyers browse the catalogue, configure, request quotes — every enquiry lands straight in the ERP. From advertising to contract and revenue, everything lives in one system — zero data islands, zero Excel exports.
Custom-built, live ERP
Internal ERP + public sales
Ad → enquiry → contract → revenue
Live production — not a presentation
Executive overview: active projects, cash-flow snapshot, overdue tasks, margin alerts — one screen.
Construction project lifecycle: status, owners, milestones, documents — CRM logic with construction semantics.
Customers, contacts, contracts, communication history — all tied to projects.
Plot registry, buildability calculation, zoning rules, analytics — foundation for sales and design.
Plan catalogue, variants, m²-based pricing, linked material lists — ERP and public catalogue in sync.
Scheduling, task assignment, subcontractor deadlines — trackable on site, including what is slipping and why.
Material lists, stock, receipts, project allocation, planned vs actual cost — protecting the margin.
Invoices, payments, cash-flow, project margin, budget vs actual — executive financial control.
Daily log, photo documentation, approved plan changes — everything is retrievable for the full project.
Handover checklist, warranty tickets, repair workflow — the customer journey doesn’t stop at keys.
Contracts, permits, drawings, invoices — versioned, role-based, project-linked.
Partner master, contracts, performance tracking — plus dedicated subcontractor portal for their tasks.
Supplier relations, quotes, orders, delivery status — procurement in one place.
Detailed budget builder, variants, approval workflow — from quote to contract.
Upload invoice → AI extracts line items → auto booking — minutes, not hours.
Internal design workflow, revisions, approvals — plan and project data linked.
Utility connections, applications, deadlines — plot- and project-level registry.
Production sequence, module assembly, delivery scheduling — industrial modular build support.
Closed-loop tracking: ad → enquiry → contract → revenue. Every dollar traceable.
Public site content, catalogue, blog — editable from the ERP, live instantly.
An internal AI assistant that learns from the user manual — it truly knows the system, it doesn’t guess.
House plans, plots, projects, tiny houses — searchable catalogue that also performs well in search and AI answers.
Lightweight frame vs brick pricing side by side — buyers see construction options instantly.
Web form straight into ERP enquiries — no Excel export, no data loss.
Cart, payment, order — a webshop experience with construction products.
Progress transparency: photos, percentage, milestones — trust-building for buyers.
Interactive house configurator: variants, extras, instant price estimate — enquiry into ERP.
The portal is built so ChatGPT and peers can cite GAIA house plans and plots — and AI-driven enquiries are measurable.
Every campaign is traceable in the marketing module — return on spend, not a guess.
Invoice image → extracted line items → booking. Human review can stay; the repetitive work is automated.
The internal AI assistant answers from the handbook — in context, grounded in how the system really works.
When someone asks ChatGPT about home building, GAIA data can be the source — not the competition. AI agents can also search the live catalogue and request an indicative quote.
Ad → click → enquiry → quote → contract → invoice → revenue. One system, one truth.
Modular yet integrated: not 15 separate tools, but one coherent platform
Public sales and internal operations live from the same data
AI invoice reading, internal AI assistant, AI visibility — built in, not bolted on later
Closed-loop marketing: every ad dollar traceable to revenue
Jógarázs · fleet management · company & private cars
Jógarázs warns you before MOT, oil change and insurance expire, keeps each vehicle’s full history and costs in one place, and flags suspicious fueling. From a handful of cars to multi-site fleets — drivers log on their phone, managers get decision support.
The fleet manager sees to-dos on one screen: what expires tomorrow, what is slipping, what needs executive action. Drivers log fueling from their phone — even offline — and AI reads receipt photos. Cost per km is measurable per vehicle; suspicious fueling and odd consumption surface early. Company and private cars on one platform: employees’ private vehicles stay hidden from admins.
Proactive deadline alerts
Vehicle 360° + cost per km
Driver mobile app, offline-ready
AI receipt reading + executive alerts
MOT, oil (km + time), tachograph, fire extinguisher, warranty, insurance — the system warns before trouble hits.
The day starts with to-dos: done, postponed, escalated — not with searching.
A car’s full life on one screen: service, fuel, damage, trips — faster decisions on replacement and liability.
When service is done, the task clears itself — less manual admin.
Manual, mobile, card import or receipt photo — fresher odometer, fuller fuel picture.
Severity, cost, photos, repair — transparent claims handling, not a lost spreadsheet row.
Simpler trip admin: place resolution, distance from odometer — backing for projects and costs.
Pool-car calendar with approval — conflict-free use, fewer arguments.
Fuel + service + damage + other — you see which vehicle drains the budget.
Suspicious fueling, odd consumption — fraud or technical issues surface earlier.
MOL, OMV, Shell and peers via CSV — no more typing transactions by hand.
Who sees and edits what — fleet manager, site, driver: order and accountability.
Registration, invoices, damage photos in one place with permissions — everything retrievable.
Employees manage their private car too; admins don’t see the details — only the count.
The driver photos the receipt; AI fills amount, litres, date — you only approve.
Installable on the phone. Fueling can be logged offline and synced later.
Critical issues reach the right person immediately — with a daily digest for leadership.
One weekly executive glance: cost, fuel, most expensive vehicles, anomalies.
Proactive alerts: critical deadlines fire immediately; neglected ones escalate to leadership
Real vehicle cost: full cost picture and cost-per-km — you see which car is expensive
Less paper, fewer errors: drivers log on the phone; AI reads the receipt photo
Order and trust: role-based access; employees’ private cars stay hidden from admins
Display Stream · digital signage · any number of screens
Display Stream’s greatest strength: no upper limit. One screen, ten or a hundred — the same system drives them all. Video, images and messages in reception, on the shop floor, in the waiting room or the shop window — from one admin, instantly.
New TV? Enter a PIN, and it is on the network. A shared loop for the whole company, different content per site, or a playlist for a single screen. An urgent message overrides everything. Content changes appear on the wall in minutes — no USB sticks, no walking from site to site.
Any number of displays
One central control
Video, image, formatted message
Urgent override + live preview
One reception TV or a full network: connect as many screens as you need. No hardware upper limit.
YouTube, photos or formatted text — the wall shows what the brand needs, not a rigid template.
Slide order is drag-and-drop. What is on the list is what plays — predictably, on repeat.
A shared loop for everyone, different content per store, or a loop for one display — as operations require.
Closing, promotion, alert: urgent content overrides everything while it is valid — on the wall, not in email.
New display gets a PIN. Type it on the TV, and it joins the network. No on-site software install.
See which screens are live and which went quiet. Reload remotely if one is stuck — without driving there.
Day, time of day, validity: mornings show one thing, evenings another — automatically, not by hand.
The player continues the already loaded loop. Shoppers do not stare at a black screen during a short outage.
See what the wall will show before it is live. Fewer surprises, faster approval.
Growth does not mean new software: you add a screen, and the same centre drives it.
Reception, workshop, shop window, waiting room — one interface, one media library.
What is urgent is on the display at the next refresh — not tomorrow, when someone walks a USB stick over.
Preview and status: you know what the network is showing before staff or guests do.
Scalable: as many screens as the network needs — no “maximum 8 TVs” cap
One person, one interface: content reaches every site without USB or on-site installs
Urgent messages hit the wall immediately — closing time, promotions, alerts, guest info
You see what the screens show: preview, status, remote reload
DevOps · project management · board, sprint, your own workflow
DevOps project management’s greatest strength: you do not rent the board per head. Tasks, board, sprint, your own workflow, time and reports — the whole operation runs on the company’s own system. Where the work happens, the system is.
Drag a card from column to column. Start a sprint, close it, see how much work is left. Shape the workflow to the company: who may move what, what is required. Comments, files, mentions and alerts stay on the task. Leaders do not assemble progress from a spreadsheet.
Board + sprint in one
Your own workflow
Time, reports, access
Yours — not a rental
Cards in columns, drag and drop. A load limit per column so the team is not drowning in parallel work.
What the team commits to this period, how much is left, what is slipping. Start, close, roll over — in one view.
Statuses, transitions, required fields. New, in progress, done — as many steps as the company needs.
Hierarchy, labels, owner, deadline. What is on the wall is the work — not a parallel spreadsheet.
Who spent what on which task. Remaining work can be counted in four ways — as leadership asks.
Comments, files, mentions, watching. The discussion stays on the task, not scattered across inboxes.
Pace, progress curve, cycle time, readiness by theme. A leadership picture in minutes — CSV if needed.
What matters today is one click tomorrow. Favourites and a dashboard: the team sees what they are working on.
System and project level. A guest or supplier sees only what they are involved in.
The company’s data, not a generic template. Typed, scoped to a project — what operations need is on the card.
Data, process and access in-house. No vendor lock-in, no monthly fee per head.
As many steps as operations need. Required fields, blocked moves — the company’s rule stands.
What they committed to, what shipped, what is slipping. One view for the team and for leadership.
Progress, pace, time spent — not a Monday-morning spreadsheet assembly.
No monthly fee per head: the system is yours, the data is yours
The workflow fits how you actually work — the software does not dictate the process
Board, sprint, time and leadership reports on one surface — not three separate tools
Access per project: outsiders see only what they are involved in
Békéskert · SEO-focused service site · with content engine
Békéskert’s strongest point is ranking first on priority Google keywords. This is not built on a single campaign: service pages, local landing pages, structured data and deliberate internal linking work together for sustained visibility.
The built-in blog engine creates additional growth headroom: every new article can target new keyword clusters while strong articles route intent back to service pages and quote requests. Content is not decoration here; it is a measurable acquisition channel.
Top rankings on priority keywords
SEO-first site architecture
Built-in blog engine
Organic lead acquisition
Page-level titles, descriptions, Open Graph and keyword logic managed as landing assets, not as boilerplate.
LocalBusiness, Service, Blog, Article, FAQ and Gallery schemas for stronger search interpretation.
Town-targeted entry points that turn local demand into quote-ready intent.
Keyword-focused but readable service content with clean CTAs and smart internal routing.
Category-driven publishing and search-friendly listing for continuous new entry points.
Article + Breadcrumb schema, OG image and metadata for stronger shareability and indexing.
Service ↔ blog ↔ contact connections to increase organic flow quality.
Crawler-oriented signals and sitemap structure for faster and cleaner discovery.
Optimized imagery, lazy load and gallery markup to support visual search surfaces.
Analytics-backed tuning: what drives traffic and what drives qualified contact.
The brand shows up at the top exactly when buyers are making a choice.
Every new article becomes a new landing asset: more keywords, more organic entries, more quote opportunities.
Location-specific content ensures visibility where demand is created.
The model scales: new service, new article, new keyword territory on the same framework.
Not just a nice website: structure, metadata and content produce ranking
The blog engine keeps opening new keyword territories and refreshes visibility
Location-focused landing pages convert local searches into quote requests
SEO works as a business channel here: traffic, contact, revenue
Asset Manager · EAM / CMMS · scan, maintenance, cost
Asset Manager’s greatest strength: from the QR on the wall to the leadership cost picture — one system. Machine, tool, building, room — the same engine. Field staff scan, the workshop opens a work order, leadership sees due dates, downtime and spend — without Excel and “we’ll do it later”.
Preventive maintenance by date or runtime. Fault → work order → parts issue. Inventory, loans, certificates, fixed-asset accounting. AI speeds capture with the camera. Decision-makers do not hunt spreadsheets: they see what is due, what is slipping, what it costs.
One scan = full picture
Assets + buildings, one engine
Maintenance + work orders
Cost and compliance
A label on the wall or machine: one tap to the record, the fault, the meter, the work order.
Machine, tool, vehicle, building, room — types, location tree, owner. No twin clone apps.
Request → approve → assign → labour + parts → close with cost. The workshop runs on it.
By date or runtime. Lead time, checklist, alerts — before it breaks.
Warehouses, moves, reorder hints. Issued from the work order — stock stays honest.
Certificates, mandatory inspections, expiry watch. Something to show at audit time.
Walk-through with missing-item report. Loans with due dates — you know who has the tool.
Capitalise, depreciate, dispose, TCO. Accounting and the workshop are not separate worlds.
Downtime, availability, PM completion, cost per asset — a leadership view, not a guess.
Camera-assisted fill-in, less typing. Faster field capture, fewer bad records.
No menu hunting: scan and you know. Fewer clicks, more real work.
One system, one access model, one timeline — the facility is not “another spreadsheet”.
Maintenance and inspections speak at due date — before the machine or the authority does.
Labour, parts, downtime, fixed assets — decisions need numbers, not feeling.
Field and office see the same data — no “Excel vs reality” gap
Expiring inspections and warranties warn early — not after the fine
Work order + parts + time: maintenance cost is measurable, not a guess
One product from workshop to facility — not two separate tools for the same job
Auto Service · workshop system · job card, stock, AI
Auto Service carries a car’s full life in the workshop: who brought it, what was wrong, what was done, what it cost, and when it left. The counter searches by plate, the technician fills the job card by voice, and AI reads the registration document and the supplier invoice. An Agentic RAG knowledge base answers from the shop’s own manual and from live numbers.
One screen holds the car: plate, VIN, inspection date, owner, custom fields and the full service history. The job card moves from the yard to close-out: labour templates, parts, hourly rate, discount, printed quote and handover sheet. The Agentic RAG knowledge base is strong because the assistant chooses the source. A “how do I do this” question is answered from the workshop manual. A “how many cars are in progress” or “who owns this plate” question is answered from live data, within the person’s access. A new colleague is productive within minutes. The owner sees the day’s picture in one sentence.
Plate → full history
Voice-filled job card
AI registration and invoice
Agentic RAG knowledge base
Plate, model or name. One screen for the car, the owner, the inspection date and every previous job.
Not here, in the yard, in progress, finished, closed. Live counts, instant save, invoice number at close-out.
Oil service, brakes, diagnostics: hours and price attached. One click, and the line is on the job card.
From stock or as a new line. Part number, quantity, and a return suggestion for unused material.
Hourly rate, labour and parts discounted separately. The promised figure and the final total sit side by side.
Photo condition report on arrival. Evidence of how the car came in, before any repair starts.
A natural sentence from the bay. Phase, part, quantity, hours, status and discount land on the card, each line reversible.
Invoice and delivery note from a photo or PDF. AI reads the lines; the shop approves before they hit stock.
Plate, model and VIN from a photo of the registration document. If the plate already exists, the system says so before history splits.
The agent opens the manual when the question is how the shop works, and live workshop data when someone asks for a number or a specific car.
Two sentences from the bay: what was replaced, how many, how long. The change list can be undone with one tap.
Registration, invoice, delivery note. AI reads, the shop approves. Less typing at the counter, fewer mistyped part numbers.
The assistant chooses the source. A “how” question is answered from the workshop manual; a “how many” or “whose car” question is answered from live data. A new hire is productive within minutes, the owner sees open work in one sentence, and the answer comes from the shop’s own knowledge and its own numbers.
A plate spoken on the phone, and the owner, the open job and every previous repair are already on screen. History is the best argument for the next service.
A live workshop flow in one system: arrival, repair, price, handover
AI sits in the daily work: voice, registration, supplier invoice, with human approval
Agentic RAG: the manual and the live number come from the same shop, by role
The same workshop discipline travels to other service bays and vehicle operations
DMA · company operating system · PDCA, margin, cash
DMA puts a Deming PDCA loop on the company: compass, strategy, 90 days, weekly rhythm, measurement, then the lesson back into the next goal. The goal does not stay in the slide deck. The decision does not evaporate after the meeting.
Eight dimensions, filled in by the leader and by the team: the perception gap is the signal. Twenty-one dysfunctions, asked as concrete situations, with at most two or three treatments at once. Every lagging metric needs a leading one under it. Contribution margin is revenue minus direct cost. Cash runway speaks in weeks; the bank balance speaks late. A project starts with expected monthly benefit and payback in months, and closes with actual benefit and the variance. The AI advisor answers from the method library and from the live numbers, within role, with the source shown.
PDCA: 6 layers, one loop
21 dysfunctions, max. 3 treatments
90 days, at most 3 goals
Margin and cash runway
Plan is the compass, the strategy and the 90 days. Do is the weekly rhythm. Check is the metric tree: a leading indicator today, a lagging one after the fact. Act is the review, and the arrow points back, because the lesson is an input to the next goal.
Revenue is a lagging number. It cannot be moved directly, so a leading indicator has to sit under it or the system warns. Overhead and profit come out of contribution margin. Cash runway is a separate leading number: it is about liquidity, not a child of operating profit.
Leading · movable today · Lagging · says what happened
Vision, annual focus, one printable page. Everyone sees the same direction.
Twenty-one stalls, asked as situations. The leader says it. Two or three treatments at a time.
At most three goals, with key results, a weekly update and a quarter-close review.
Agenda, issues list, decision log. The meeting leaves a trace.
A lagging metric needs a leading one under it. Revenue says what happened; customer count and basket can be moved today. Above them sit margin and operating result.
At the start: expected monthly benefit, cost, payback in months. At the close: the actual and the variance. The measurement is a starting point, not proof.
Role map, knowledge, one-to-ones, onboarding. The key person is not the process.
Complaint, opinion, value. The company hears why it is chosen, and why not.
The key step is written down. On holiday, the company still knows how it is done.
It answers from the method and the live number. It shows the source and keeps the role.
Direction, breakdown, the week, measurement, learning. The lesson returns to the next goal.
A concrete situation, not an abstract question. The treatment order comes from that recognition.
If revenue rises and the margin rate falls, you are buying dearer or selling cheaper. Post-calculation sets expected monthly benefit next to the actual.
“Why is the quarter slipping?” The answer joins the key result and the blockage, from the company’s own data. A new leader uses the method within minutes, and the number lives in the system.
Strategy, the week and today’s task live in one loop
The company names its own stall and gets a treatment order
The lagging metric reaches contribution margin; cash runway speaks ahead, in weeks
AI answers from the company’s method and live numbers, by role
Whether ecommerce, restaurant, construction, fleet, displays, project management, SEO, asset management, an auto workshop or company leadership — we ship live systems. Let’s talk about your project.
future@454garage.com