NEIRA. AI in the Pizza Business. Volume 1. From the First Idea to Pizza Production: A Practical Guide for Beginners and Experienced Entrepreneurs

Day 7. Select no more than three priority hypotheses. For each, assign an owner, a safe test, a deadline, and a measurable success criterion.

The order can be adjusted to suit operating schedules. A week is enough for an initial diagnosis, not proof of seasonal patterns. Reliable conclusions usually require a longer comparable period.

3.10. What Is the Beginner Doing Meanwhile?

While the Veteran investigates existing losses, the Beginner learns an essential lesson: do not design a new pizzeria solely around someone else’s sales or the line at their counter.

Beginner: “I saw a pizzeria that’s always busy. Should I copy its format?”

Neira: “Not yet. We don’t know what remains after all its costs. Study demand, but don’t copy an invisible financial model. Your business plan should specify measurable indicators, product movement records, and accurate accounting for discounts and expenses from day one.”

The Beginner’s assignment is to extend the concept profile from Chapter 2 with planned control metrics: order count, average ticket, food cost, waste rate, dispatch time, refunds, and operating result. For each, identify the data source and the person responsible for its accuracy.

3.11. Who Makes the Decisions?

Neira may prepare calculations, options, instructions, and alerts about deviations. But pricing, staffing, purchasing, recipe changes, and commitments to suppliers must be approved by designated people. Any automated operation requires limits of authority, an audit trail, and a way to stop or reverse it.

Veteran: “So you won’t run my business instead of me?”

Neira: “I’ll help you manage it using evidence rather than guesses—and show whether each recommendation comes from a report, a measurement, or an untested hypothesis.”

Review Questions

1. Why does increasing revenue not necessarily increase profit?

2. Which data can come from a management system, and which must be measured in the kitchen?

3. Why should ingredient quantities not be reduced before checking quality and economics?

4. What distinguishes a confirmed fact from a hypothesis about the cause of losses?

5. What should the Veteran have at the end of the first diagnostic week?

6. Which decisions must remain under responsible human control?

Chapter Summary

The Veteran received no promise of miraculous profit growth. Instead, he has a workable plan: establish trustworthy baseline figures, follow the production process, identify measurable deviations, and test a small number of improvements. The Beginner has learned which mistakes to prevent before opening. The essential principle of the proposed independent Neira system is that digital records and deep production knowledge must reinforce one another.

The next chapter examines Neira’s responsibilities in more detail: what existing management software can do, what Neira adds, and which functions still need development and pilot testing.

Neira and Management-System Integration — The Veteran

The first step is to agree which indicators Neira should receive from an existing management system: item- and hour-level sales, waste, purchasing, inventory balances, and labor costs. Neira can then propose explanations for deviations and an improvement plan, but every recommendation must be supported by source data and evaluated against actual results. Integration remains a proposed technical capability until it is built and validated.

Workshop Deliverable and Acceptance Criterion

The Veteran produces a seven-day diagnostic table with baseline revenue, waste, utilization, and complaints; three priority potential causes of losses; assigned owners; and dates for review. Unknown figures are explicitly marked as data to be collected.

Chapter 4. What Neira Can Do Today and What Still Needs to Be Built

Editorial Clarification: The Limits of Current Capabilities

Today Neira can help prepare assignments, analyse spreadsheets supplied by users, explain approved production specifications, and draft business plans and training materials. She does not automatically have access to the point-of-sale system, inventory, equipment or bank transactions. Continuous operations management requires authorised connections, a dedicated data store, access permissions, event logs and verifiable calculation services. Throughout this book we distinguish three states: available now; possible after configuration; and requiring separate development.

4.1. A Conversation That Began with an Uncomfortable Question

After meeting the Beginner and the Veteran, Neira returned to her digital workspace. On her virtual desk were two documents: a profile of a proposed pizzeria and a diagnosis of an operating business. Both raised many questions. The hardest was this: who would actually solve each problem?

Neira: “My task is to connect production technology and operational records into one sequence of actions without confusing verified facts with assumptions. I can analyse data you provide, draft instructions and check calculations. I can obtain sales, stock balances and orders automatically only after the appropriate connections have been developed and tested.”

Neira: “I also know what to ask a pizza maker, technologist or owner. But that knowledge must be turned into verifiable instructions, working modules and secure data exchange.”

They agreed that in this book “Neira: AI in the Pizza Business” denotes a proposed industry-specific system. Connecting it to external software is a development and pilot-testing project, not an already functioning official integration.

4.2. Three Distinct Concepts: Management Software, the Neira Assistant, and the Neira Platform

First is the management system: a set of software tools for automating food-service operations. Depending on the package and configuration, it may support POS transactions, order handling, inventory and production records, reports, delivery and workforce management.

Second is Neira as the digital assistant depicted in this textbook. She helps people discuss pizzeria data and operations. The reports, requests and automatic actions actually available in a particular management system must be verified against its documentation and a demonstration. A feature existing in the software does not mean Neira can use it independently.

Third is Neira as a proposed specialised AI platform for the pizza industry. Its scope goes beyond report figures to the entire life cycle of a pizzeria: idea, design, production technology, staff training, procurement, quality, profitability and chain development.

Veteran: “So Neira is not the whole management system, and she is not another cash register?”

Neira: “Exactly. To find daily sales, we first check what reports and data are available. To find out why a pizza burns at the edge, a sales report is not enough. And to determine how many dough balls to prepare for the evening rush, we need both a demand forecast and the dough’s production constraints.”

4.3. Responsibility Map: Who Does What

The following matrix describes a proposed division of responsibilities, not a promise that every function is available in a ready-made integration. Before a pilot, classify each function as available now, available through configuration or an interface, requiring development, or unsupported.

Task

Existing management software / available records

Neira: proposed contribution

Sales and orders

Available data, reports and order navigation; access must be verified

Translate demand into a production plan and assess execution quality

Recipes and costs

Recipe records, prices, stock movements and available cost indicators

Check technological feasibility, actual portions and production tests

Flour and dough

Receipts, stock and write-offs where configured

Calculate maturation, mixing and dough-ball requirements

Cheese and sauce

Purchase and inventory records

Verify portioning, defects and ingredient performance

Staff and shifts

Available schedules, roles and sales statistics

Skills training, workstation instructions and quality checks

Supply

Supplier records and purchase documents

Demand planning, order proposals and confirmation workflows

Management

Financial and operating reports

Investigate root causes and prioritise verified improvements

4.4. The Beginner: Digital Assistance Before the First Sale

The Beginner has not yet rented premises. There is no POS system, sales history or stockroom. Analysing past revenue would therefore make no sense. Neira proposes a profile of the future business: district, intended customers, service format, menu, floor area, budget, electrical capacity limits, opening hours and sales channels.

Beginner: “They want to sell me automation software when I open. Isn’t it too early to discuss it?”

Neira: “Quite the opposite. If you do not decide in advance how to record dough, prepared ingredients, waste and shifts, you will have to correct mistakes while the business is already running. But before opening we must not present invented figures as real ones. We will list the required data and settings and label forecasts as assumptions.”

Task one: map the journey of a single order—customer → order → production ticket → dough preparation → assembly → baking → packing → handover → payment → ingredient and result records. For each transition identify the actor, the record and the possible error.

Task two: list future control indicators: orders by time interval, average ticket, discount share, flour and cheese use, dough waste, handover time, remakes, cost of goods and operating profit. Where a measure cannot yet be obtained, specify the equipment, procedure or record needed.

4.5. The Veteran: Why Good Reports Cannot Replace a Technologist

The Veteran brings a spreadsheet: 1,200 pizzas sold in a week, revenue up, and cheese write-offs also up. He wants to punish the stockroom employee immediately. Neira stops him: the increase may result from inventory errors, incorrect recipe quantities, spoilage, stock mix-ups, a changing sales mix or unrecorded movements.

Veteran: “The report says cheese was written off. Does that mean the stockroom is to blame?”

Neira: “The report records a transaction. It does not establish its cause or who is responsible. First we check source documents, batches and stock accuracy. Then we weigh actual cheese portions at the assembly station and compare them with approved recipe specifications. Without observing production, the diagnosis is incomplete.”

Procedure: select one period; export sales of cheese-containing pizzas and cheese movements; calculate standard consumption for the actual product mix; inspect stock balances and adjustments; measure portions over several shifts; document discrepancies; and separately assess spoilage, returns and remakes. Only then decide what to change.

4.6. First Practical Calculation: How Many Dough Balls?

Consider a hypothetical example. The evening forecast is 180 pizzas. Each requires an average of one dough ball. At the start of the period, 35 usable, ready-to-use balls are available. An additional reserve of 8% of forecast demand, rounded up, is 15 balls.

Total requirement = 180 + 15 = 195 balls. New preparation required = 195 − 35 = 160 balls. This is an arithmetic result, not an instruction to mix dough immediately: ball weight, maturation time, shelf life, refrigerator capacity and acceptable production windows must also be checked.

Neira: “I can help prepare the demand forecast when suitable data and authorised access exist. Otherwise a person enters the agreed sales plan manually. We must also establish how many of the 160 balls can be ready on time, the mixing schedule and storage conditions.”

Veteran: “So ordering flour begins with expected sales, not an empty flour bag?”

Neira: “With expected sales, verified recipes, available stock and actual production constraints.”

A flour order needs a separate calculation using dough weight, flour percentage in the recipe, process losses, available stock and supplier lead time. Pizza counts cannot simply be converted into kilograms of flour without these inputs.

4.7. How Neira Should Communicate with Employees

A pizza maker needs a short working instruction, not a lecture on AI. A manager needs a prioritised list of deviations. An owner needs a proposal whose financial effect can be checked. A supplier needs an accurate order specifying items, quantities, timing and confirmation.

Neira must never replace a production specification with an attractive but vague suggestion. Each instruction must refer to a particular recipe, equipment, measurable tolerance and accountable person. If those facts are missing, she asks a clarifying question instead of inventing a standard.

4.8. Data, Permissions and the Limits of Automation

Neira should receive only the information necessary for the task. Access to employee and customer personal data, financial information and supplier terms must be role-based. Data exchange between systems requires a lawful basis, secure channels and an audit trail.

Every action belongs to one of four levels: (1) display information; (2) recommend an action; (3) prepare a document or command for approval; (4) execute an action automatically within previously authorised limits. Level four is permissible only after testing, permission configuration and a rollback mechanism.

Supplier orders, price changes, recipe changes, staffing decisions and financial commitments should require human approval by default. Food safety, allergens, storage limits and hygiene are especially important: no AI recommendation overrides applicable requirements or approved procedures.

4.9. What Neira Still Needs to Learn

To become an industry platform, Neira needs more than a library of pizza books. She needs structured and verified recipes and process specifications, defect reference materials, training scenarios, workstation standards, demand-calculation models and a system for validating recommendations in real production.

Development block one, Beginner: concept profile, format selection, capacity calculations, premises, equipment and accounting requirements. Block two, Veteran: diagnosing an operating pizzeria, identifying losses, improvement plans and measuring results. Block three, Production: dough, crusts, sauces, cheese, assembly, baking and finished-product control. Block four, Personnel: training, practical assignments, digital skills records and assessment. Block five, Supply: demand forecasts, inventory counts, orders, confirmations and links with ingredient manufacturers.

For each block define inputs, expected outputs, autonomy limits, quality criteria and the person responsible for verification. Replace “AI will do everything” with a testable process: “The system prepares a calculation, an employee verifies the inputs, and a manager approves the action.”

4.10. Pilot: Test Results, Not Promises

The first pilot should cover a limited set of tasks at a participating pizzeria. Baseline indicators, observation period, comparison conditions and the method for calculating benefits must be defined beforehand. Otherwise seasonality, demand shifts or random variation may be mistaken for an improvement.

A pilot result is confirmed only after checking data quality, comparability of periods and the absence of unacceptable deterioration in quality or safety. Net economic benefit must deduct implementation, support, equipment and employee labour costs.

4.11. Practical Assignments for the Beginner and the Veteran

Beginner: draw a map of ten key operations in the proposed pizzeria. For each, record who performs it, required data, controls, potential errors, the role of existing management software, the role of Neira and what still needs development. List five indicators to measure from day one.

Veteran: choose one recurring problem—for example, dough waste, excessive cheese portions or delivery delays. Collect data for a comparable period, plan observations at the workstation and formulate one testable hypothesis. Do not change several variables at once, or it will be difficult to attribute the result.

Joint assignment: create a scenario card with the objective, inputs, source of each number, Neira’s tasks, permitted automated actions, human tasks, authority limits, outcome metric and method for reversing an erroneous action.

Acceptance criterion: for the chosen scenario, specify inputs, access boundaries, control indicator, independent verification method and reversal procedure. If integration is not implemented, label the exchange as proposed rather than operational.

Review Questions

1. Why must an entire management-software ecosystem not be equated with Neira’s capabilities?

2. Which tasks require direct observation of pizza production?

3. How does a forecast differ from a verified fact?

4. Why does the dough-ball calculation not replace a mixing and maturation schedule?

5. Which operations require approval by a responsible employee?

6. How can an existing function be distinguished from one that is only proposed?

7. Which metrics should be used to evaluate a pilot of the independent platform?

Workshop Deliverable and Acceptance Criterion

Produce a matrix of functions available in existing management software and functions proposed for Neira, without claiming that a working integration already exists. Every function must be marked as available, requiring verification, or planned.

Chapter Summary

Neira establishes the central rule: do not present future capabilities as existing ones. She can already help people work with information they provide and with accessible digital tools; an independent Neira platform would translate industry expertise into verifiable production and management actions. The Beginner now has a basis for a digital pizzeria plan, and the Veteran has a way to connect reports with what actually happens in the kitchen. The next chapter brings these tasks together into the modules of an industry-specific system.

Additional Clarification: What Is Available Without Building a Platform

With an existing AI assistant, an entrepreneur can draft business plans, analyse manually uploaded spreadsheets, prepare checklists, compare supplier proposals and develop training materials. Input figures, legal requirements and technical decisions still require verification.

Additional Clarification: What Requires Custom Development

Automatic order retrieval, inventory management, purchasing approval, equipment connections and continuous quality monitoring require databases, calculation services, interfaces, access controls and testing. A language model by itself is not a POS system, inventory system or production-safety system.

Additional Workshop: Neira’s Task Map

The Beginner lists ten opening tasks and records their inputs, outputs and verification methods. The Veteran selects three recurring operations and measures their time, errors and cost before automation. For each process, identify manual, semi-automated and automated modes.

Chapter 5. Six Modules of the Independent Neira System

5.1. Six Modules, One Pizzeria

The independent Neira system is designed around six interconnected modules: (1) business creation; (2) production and quality control; (3) personnel, economics and management; (4) sales, marketing and delivery; (5) procurement and cooperation with manufacturers; and (6) network development. They share reference data on ingredients, recipes, employees, locations and operational events. Each module has defined inputs, an action, a responsible person and a verification criterion.

“We can turn this into a working system,” Neira said, “if we divide the work into clear areas. Every area needs a purpose, reliable inputs, an accountable employee and a measurable result.”

The six-module structure is a proposed industry-specific architecture, not a list of functions already implemented in a particular accounting or point-of-sale system. The platform may exchange data with different systems through documented, authorised interfaces, without depending on a single vendor.

General rule: order, stock, sales and financial data may be used only where access is authorised and data quality has been checked. Neira’s technical advice must be grounded in the company’s approved standards rather than replacing them.

5.2. Module One: From an Idea to a Sustainable Business

The Beginner arrived with a notebook. On its first page he had written, “Open a pizzeria.” On the second was a list of ovens he had admired at a trade exhibition.

“It is too early to buy an oven,” Neira stopped him. “First identify your customers, format, menu, expected demand and production method. Only then can we determine the capacity the equipment must provide.”

Neira proposed a project profile: city and neighbourhood; service format (dine-in, delivery, takeaway); target customers; available capital; expected floor area; opening hours; proposed menu; lease, electrical and ventilation constraints; and desired opening date. Unknown figures are recorded as assumptions, not facts.

Example: the Beginner expects to sell 90 pizzas a day. If 45% of orders fall within a three-hour evening peak, about 41 pizzas will need to be made during that period—an average of approximately 14 per hour. This is only an initial estimate. Demand spikes within individual hours, the mix of products, baking times and reserve capacity all require separate checks.

For the Veteran, this module begins with a profile of an operating business: format, area, seats, equipment, menu, revenue by channel, production bottlenecks and customer complaints. The objective is not to open another business immediately but to identify what prevents the existing one from earning more.

Deliverable: an agreed concept for the Beginner or a map of constraints and a development plan for the Veteran. The owner makes the investment decision.

5.3. Module Two: Pizza Production and Quality

“A beautiful chart cannot improve a pizza on its own,” Neira said. “We must understand what happens to flour, dough, sauce, cheese and crust at every workstation.”

Here Neira acts as a production mentor. She helps prepare recipe specifications, control points, storage instructions, dough-forming procedures, portioning standards and baking instructions. Operational records can support analysis when accessible and verified. If a required quality-control function does not exist, it is documented as a development requirement.

Example: a pizzeria makes 120 pizzas a day, with a standard cheese portion of 110 g per pizza. The theoretical requirement is 13.2 kg. Actual consumption is 14.7 kg. The difference is 1.5 kg, or approximately 11.4% of the standard. Before declaring overuse, check sales records, pizza sizes, recipes, opening and closing stock, write-offs and trial batches.

Neira recommends checking dispensers, scales, staff instruction and actual portions in a sample of pizzas. She also identifies which records are needed. Employees confirm measurements, and managers approve corrective actions.

For the Beginner, the priority is to establish standards before opening. For the Veteran, it is to compare approved recipes with actual practice without changing either until the production evidence has been reviewed.

5.4. Module Three: People, Training, Economics and Management

“I will hire good pizza makers,” the Beginner said.