Northfield Applied AI Software LLC

Words in [brackets] are blanks you fill in with your own details when you adopt this sample in Supreme Plan AI.

Executive Summary

This company in Texas is seeking a grant of $150,000 from the Frontier Software Grant Fund to finish and deploy Northfield Permit Assist, a software tool that helps county and municipal permit offices review applications faster and more consistently. The tool reads submitted plans and forms, checks them against the local code, flags missing items and conflicts, and drafts a review checklist that a human reviewer then confirms. A reviewer always makes the decision. The grant pays for the research, testing, security work and first pilot deployments that turn a working prototype into a product that counties can trust.

Permit review is slow. In many offices, an application can wait weeks for a first review, and applicants often resubmit because a missing detail was not caught early. Staff are stretched thin, and inconsistent review across reviewers creates frustration and cost for builders, homeowners and the office itself. Our software is designed to reduce that waiting time and rework while keeping people in charge of every decision.

Our Year 1 plan is to complete testing on public, anonymized permit records, run two paid pilots with partner counties and certify our security controls. Revenue is planned at about $650,000 from software subscriptions and about $90,000 from implementation and support in Year 1, growing to about $2,000,000 and $440,000 by Year 5. Our Year 1 budget is about $1,527,500. We have 9 paid staff in Year 1 and 19 by Year 5. Grants are planned to fall from about 52% of revenue in Year 1 to about 20% in Year 5 as paying customers grow.

We are not promising any county a particular speed, accuracy or savings. Our pilots will measure those things with the county and we will publish the results, including where the tool did not help.

I am Dana Whitfield, the Chief Executive Officer, a former county planning director and a software founder. The numbers in this proposal are estimates based on the assumptions listed at the end, not promises.

Company Description

Northfield Applied AI Software LLC is a limited liability company (LLC) in Texas, formed to build and sell software for local government permit review. Northfield is a for-profit small business with fewer than 25 employees. We will provide our formation documents, a good standing certificate, our insurance certificates, our federal and state tax registrations and our most recent financial statements with this proposal. We meet the Frontier Software Grant Fund's size rule for emerging technology companies, which we will confirm with the Fund's staff before we accept any award.

Our mission. To make public permits faster, fairer and easier to understand, without taking decisions away from the people who are accountable for them.

What we believe. Software should serve public servants and the people they serve. A tool that reviews applications should explain what it found, show its sources and make it easy for a human to agree or disagree. Local governments are accountable to residents, and we design for that accountability. We will never build a tool that makes a final decision about a person's permit, benefit or rights on its own, and we will say so in every contract.

How we started. Our founders worked in county planning offices and built a prototype after watching reviewers spend hours on missing-form problems. In an early test with retired reviewers on public records, the prototype flagged most common omissions (replace this sentence with your own measured results before you submit). Our first two pilot conversations, with a county and a regional council of governments, led to a letter of intent for paid pilots in Year 1, which we describe in the solution section.

Who we serve. Local permit and code offices, initially counties and mid-sized cities that process residential and small commercial permits. We will confirm each customer's procurement rules, records law and data-sharing requirements before any pilot begins.

What we are not. We are not a legal or code authority, a consultant who writes permits for applicants, or a surveillance or enforcement tool. Our product does not identify people, track residents, or score their likelihood of anything, and we will not sell or share resident data. We are not a registered design professional or engineer, and we say so in our product documents.

What we offer. Plan and form checking that finds missing items, conflicts and common code issues. A review checklist for each application that a reviewer confirms. A plain-language summary for the applicant, with the reasons behind each comment. A dashboard that shows wait times and rework rates, so the office can see whether the tool is helping. Training and support for reviewers.

How we are run. A three-person founding leadership team, a board of advisors that includes a retired county planning director and a building official, and an outside accountant who reviews our books each year. We will hold a formal responsible AI review at the start of each major release, and we will publish our model documentation and the results of our bias and accuracy tests for each customer deployment.

Where things stand today. We have a working prototype, a letter of intent for paid pilots with one county and a memorandum of understanding with a regional council of governments. We have an outside security firm under contract for a readiness assessment. We have not yet taken any outside equity, and we are seeking grant funding to finish product work before we raise our first equity round.

Problem & Opportunity

Permit offices are a bottleneck that nearly everyone touches. A homeowner who wants a deck, a small business that wants a sign, a contractor who wants to replace a roof, and a developer who wants to build homes all wait on the same overloaded reviewers. Many applications come back for corrections that could have been caught on the first pass, such as a missing site plan, an unsigned form, a setback that does not match the zoning map or a fee that was not calculated. Each round trip adds weeks.

Local governments have trouble hiring and keeping experienced reviewers. When a senior reviewer retires, their knowledge of the local code often leaves with them. Less experienced staff are asked to decide questions that once took a decade to learn, and different reviewers can reach different answers on the same question. That inconsistency is unfair to applicants and creates legal and political risk for the office.

Existing tools are mostly electronic submission portals and workflow systems. They move paper faster, but they do not help a reviewer check the content of a plan against the code. Some vendors have promised automatic approval, which is not a responsible goal, because a permit decision carries legal responsibility that has to remain with an accountable official.

The opportunity is narrower and more useful: a tool that helps reviewers find problems sooner, explains its reasoning and leaves the decision with the person who is accountable. If that tool is accurate and well-tested, it can reduce rework and wait times without weakening oversight.

We also know the risks. Software that reads plans can be wrong, and a wrong flag can slow an honest applicant or miss a real problem. Models can be biased against certain kinds of projects or certain neighborhoods if they are trained or tested on unrepresentative records. Data about applicants is sensitive. These risks are why our first year is about testing, measurement and security, not sales.

We do not know yet how much time a county will save, and we do not promise any saving. Our pilots are designed to measure it honestly, with control periods and reviewer feedback, and we will publish the results whether they are good or not.

Products & Services

Northfield Applied AI Software LLC will complete and deploy Northfield Permit Assist in three phases during the grant period. Phase one is product hardening and testing. Phase two is supervised pilots with partner offices. Phase three is paid deployment with ongoing monitoring.

How it works, step by step:
1. An applicant submits a permit application and plans through the county's existing portal. Our software receives the file only with the office's written approval, and only the files needed for review.
2. The tool reads the forms and drawings, checks required items against the local code checklist, looks for conflicts between documents (for instance, a site plan that does not match the application) and drafts a list of findings with the code section and page reference for each one.
3. A reviewer opens the findings. The reviewer can accept, edit or reject each finding, and can add their own. Nothing is sent to the applicant until a human has reviewed it.
4. The applicant receives a plain-language summary of the confirmed findings, with the reason and the code citation, and a way to ask a question or submit a correction.
5. The office dashboard records each decision, the time to first review, the number of rounds and the reviewer's changes, so the office can measure results and spot patterns that need attention.

What the grant will pay for. Part of the salary of one Applied Research Scientist, a large share of the cost of the Phase One testing program, including compute, licensed evaluation datasets and a bias and accuracy audit by an outside evaluator, a security readiness assessment and penetration test, and the legal work on pilot agreements and data-use terms. The grant does not pay for marketing or general sales expenses.

Human decisions. Every finding that reaches an applicant has been confirmed by a reviewer. Our contracts require the county to keep final decision authority and to keep a record of every reviewer change. We will not offer any feature that issues a permit, denies an application or sends an automatic notice of decision.

Testing before deployment. Before any pilot, we measure the tool's false positive and false negative rates on a held-out set of anonymized permit files from several jurisdictions, broken down by project type, applicant type and neighborhood where the data allows. A pilot starts only when the results meet thresholds agreed with the county, and the thresholds are published in our model documentation.

Market Analysis

Who benefits: local permit offices and the residents, builders, homeowners and small businesses who wait on them.

Our direct customers are county and municipal building, planning and zoning departments that process residential and small commercial applications. We are starting with offices that handle between 5,000 and 40,000 applications a year and have fewer than 30 reviewers, because they feel the bottleneck most and have the most to gain. Our first pilot county processes about 18,000 applications a year. Our planning assumes two paid pilots in Year 1, growing to a steady base of subscription customers by Year 3, with about 19 paid staff in Year 5.

Market size. We estimate there are several thousand local permit offices in the United States in the size range we target. We will use the most recent Census of Governments and public procurement data to size this market in our final application, and we will not overstate it. Adoption will be slower than the market size suggests, because offices buy after a pilot, after a security review and after an agreed measurement plan.

Our secondary customers are contractors and applicants, who gain fewer rounds of correction; reviewers, who gain a checklist and a second reader; and residents, who gain clearer explanations of what is needed.

Pricing. We charge an annual subscription based on application volume, plus a one-time implementation fee and optional training. We do not charge applicants, and we do not sell resident data to anyone.

We do not guarantee any customer a result. If a pilot does not show a benefit, we will say so, and we will stop or adjust the deployment.

Competitive Landscape

Other choices local governments have for permit review, and what makes us different.

- Electronic permitting portals and workflow vendors. These are established and move applications through steps. Most do not check plan content against code, and they are the systems our tool is designed to work alongside, not replace. We integrate with them where the office allows.
- Staff reviewers and code consultants. Experienced staff are the best judges of a plan, and consultants can help with overload. Both are expensive and in short supply. We support them and do not replace them.
- Plan review services. Some firms review plans for a fee. They can be effective, and their costs are passed on to applicants. We are a tool for the office, not a service that charges applicants.
- Other AI document tools. General-purpose tools can read documents, but they are not built around local code, reviewer decisions or public records rules. We differ by building the checklist, the audit trail and the measurement around the office.
- Doing nothing. This is the real competitor. Many offices have accepted long waits as normal. We aim to show, with measured pilots, that a better process is possible without giving up oversight.

What makes us different.
- Built around human decisions and a clear audit trail for every finding.
- Local code checklists built with former reviewers and confirmed by each office.
- Published accuracy and bias measurement, with thresholds agreed before a pilot starts.
- A dashboard that measures time and rework, so the office can see the benefit or the lack of it.
- No resident tracking, no resident scoring and no sale of data.

What we do not claim. We are early. Our prototype has been tested on a limited set of records, and no outside evaluation has been completed yet. That is what the grant is for.

Marketing & Sales Strategy

How we will reach customers, partners and the people we serve once the project is funded.

Customers. We will sell through a structured pilot. A county signs a memorandum of understanding that sets out the pilot's scope, data handling, success measures and an exit point. We will meet the county's procurement office, its building official and its IT and records staff, and we will ask for a short written agreement before any file is shared. After a successful pilot, the county may purchase a subscription.

Partners. We will work with regional councils of governments, state building officials' associations, professional associations for permit staff and code consultants. We will present our measurement approach at their meetings and invite their feedback before we make any claims.

Channel partners. We may partner with existing permitting software providers to integrate our review tool into their systems. Any partnership will include the same human review and data protection terms as our direct contracts, and we will not pay any partner for a referral that carries a conflict of interest.

Applicants and contractors. We do not market to applicants. We explain the process to them only through the office that uses the tool.

Investors and funders. We will report our measured results honestly, including failures, to funders and to prospective investors. We will not show pilot results we have not yet checked independently.

Public. We will publish our model documentation, our evaluation methods and a plain-language explanation of what the tool does and does not do. We believe public trust depends on that openness.

Operations Plan

How the work will be carried out.

Timeline for Year 1:
- Months 1 to 3: Hire the Applied Research Scientist and the first machine learning and software engineers. Complete pilot agreements with one county and a regional council. Engage an outside evaluator and a security firm. Set up secure development and data handling environments.
- Months 4 to 6: Build the evaluation set from anonymized public permit records with the county's approval. Measure baseline accuracy. Finish the first code checklist for the pilot county with a former reviewer.
- Months 7 to 9: Run the supervised pilot with a reviewer team, with control periods and feedback. Complete the penetration test and fix findings before any live resident data is used.
- Months 10 to 12: Review results with the pilot county, publish model documentation and the first evaluation summary, and decide on paid deployment only if the agreed thresholds are met.

Products and infrastructure. We run on cloud infrastructure with encryption at rest and in transit, role-based access and audit logs. Each customer's data is kept separate, is deleted on the schedule in its contract and is never used to train models for other customers without written permission. Cloud computing and model hosting cost about $90,000 in Year 1, and model licenses, data labeling and evaluation tools cost about $40,000.

Responsible AI controls. We keep model cards for each release. Each release is tested for accuracy, false positives and false negatives by project type and applicant type, and for performance gaps across neighborhoods where we have the data and the county agrees. A release that does not meet agreed thresholds is not deployed. Customers can see the results. Every finding carries a source reference, and a reviewer can reject it with a reason, which we use to improve the system.

Privacy and security. We collect only the files needed for review. We do not identify people, do not perform facial recognition, do not track residents and do not build profiles. Access is limited to named staff, with multi-factor authentication. We follow the security controls in our customers' contracts and an industry framework that we will confirm with the outside assessment. We will notify each customer of a security incident as its contract requires.

Customer success. Each customer has a named implementation lead for the first 90 days. Training is delivered in person or by video, and reviewers can flag any finding that seems wrong, which we review within five business days.

Space and equipment. We lease a small office and buy research workstations, a shared GPU server and laptops. Our equipment and cloud spending are planned in the startup and operating budgets, and we set aside a replacement reserve for hardware after Year 1.

Legal and intellectual property. Our attorney reviews our county data-use agreements, privacy terms, open-source licenses and patent filings before we sign. We will file patent applications where our counsel advises that they protect the product, and we will disclose any open-source components we use.

Insurance and records. We carry general liability, professional liability (errors and omissions), cyber liability and workers' compensation once we hire. Insurance is about $18,000 a year in Year 1. Our outside accountant reviews our books every year, and we keep grant money separate so each funder can see how it was spent.

Management & Organization

The people who will carry out the project.

Dana Whitfield, Chief Executive Officer. A former county planning director who led a permit office through a backlog and a software replacement. She leads sales, partnerships, finance and hiring, and she is the main contact for funders. (Replace this paragraph with your own leadership background.)

Chief Technology Officer. A software architect who has built secure, data-sensitive systems for public agencies and has led engineering teams through security reviews. Owns the architecture, the security program and the release process.

Machine Learning Engineers (two in Year 1, three in Year 3, four in Year 4, five in Year 5). Build and maintain the document reading and checking models, run evaluations and maintain the model documentation.

Software Engineers (two in Year 1 and three from Year 2). Build the reviewer interface, the applicant summary, the dashboard and the integrations with customer systems.

Applied Research Scientist. Designs the evaluation program, the thresholds and the bias and accuracy audits, and works with the outside evaluator. A doctorate or equivalent experience in machine learning evaluation is expected.

Customer Success and Implementation Lead (one in Year 1, two in Year 2). Runs pilots, trains reviewers and keeps a record of every reviewer change and customer concern.

Bookkeeper (part-time). Keeps the books and works with our outside accountant.

Advisors. A retired county building official, a former reviewer and an attorney experienced in public procurement and data protection advise on products, pilots and contracts. Advisors are paid only in equity or cash that we disclose to the board and to funders, and they sign a conflict-of-interest agreement.

Staff pay in this plan includes payroll taxes and benefits. Salaries are CEO $150,000, CTO $170,000, Machine Learning Engineer $145,000, Software Engineer $130,000, Applied Research Scientist $150,000, Customer Success Lead $90,000 and part-time Bookkeeper $36,000. Replace them with the pay you can offer where you live.

Milestones Narrative

Our goals for each year and how we will measure them.

- Year 1: Complete the product hardening, sign one paid pilot and one memorandum of understanding, pass the security assessment and publish our first model documentation and evaluation summary. A revenue of about $1,527,500 in revenue, with profit after income tax of about $42,400. 9 paid staff.
- Year 2: Run a supervised pilot with reviewer feedback and a control period, complete the penetration test fixes and begin the second pilot. A revenue of about $1,801,500 in revenue, with profit after income tax of about $45,200. 11 paid staff.
- Year 3: Convert successful pilots to paid subscriptions, add a second and third customer and launch the dashboard and integrations with two permitting software systems. A revenue of about $2,160,500 in revenue, with profit after income tax of about $49,000. 13 paid staff.
- Year 4: Grow to about ten customer offices, add reviewer training programs and begin the second product line for code compliance follow-up. A revenue of about $2,602,500 in revenue, with profit after income tax of about $54,900. 16 paid staff.
- Year 5: Expand to about twenty customer offices, with published results for each, and reduce the share of revenue from grants as subscriptions grow. A revenue of about $3,058,000 in revenue, with profit after income tax of about $61,000. 19 paid staff.

What we measure:
- Customer offices live, and subscriptions renewed.
- Time from submission to first review, and number of correction rounds, compared with a control period.
- Accuracy of findings, including false positives and false negatives, by project type and applicant type.
- Reviewer acceptance, rejection and editing rates, and reasons given.
- Security incidents, audit results and customer complaints, reported in full.
- Revenue, gross margin and cash runway.
- Applicant satisfaction, collected by the office and reported in totals.

How we review results. The Chief Technology Officer and the Applied Research Scientist review release quality and evaluation results every two weeks, and the leadership team reviews outcomes monthly. Each pilot ends with a joint review with the customer, and we change the product or stop a deployment when the agreed results are not met. The board reviews results and security quarterly. Every number in this plan is a goal or an estimate, and we will report what really happens, including when we fall short.

Funding Request

What we are asking for, and how the money will be used. Our request: $150,000 from the Frontier Software Grant Fund for one Applied Research Scientist, the Phase One testing and evaluation program (including compute, licensed evaluation data, the outside bias and accuracy audit and the security readiness assessment) and pilot legal work, for July 2027 through June 2028.

The request is spread across these budget lines:
- Personnel: Applied Research Scientist salary, $60,000 (other funds add $90,000).
- Other: Phase One testing and evaluation program (compute, licensed evaluation data and outside bias and accuracy audit), $60,000 (other funds add $20,000).
- Other: Security readiness assessment and penetration test, $20,000.
- Contractual: Legal work on pilot agreements, data-use terms and patent filings, $10,000 (other funds add $10,000).

Other money in the project. Our own revenue, other grants and awards, and investment we are raising pay the other $120,000 of those lines, plus the leadership and engineering salaries, cloud computing and model hosting, office rent, sales and marketing and insurance in the budget. Investment is not yet committed, and we do not count it as received in this plan.

Year 1 budget: about $1,527,500. Year 1 money coming in: grants $787,500 (this request plus other grants and awards we are applying for), software subscriptions (county and municipal customers) $650,000 and implementation and support services $90,000. Year 1 costs: Chief Executive Officer $150,000; Chief Technology Officer $170,000; two Machine Learning Engineers $290,000; two Software Engineers $260,000; Applied Research Scientist $150,000; Customer Success and Implementation Lead $90,000; Bookkeeper $36,000; cloud computing and model hosting about $90,000; model licenses, data labeling and evaluation tools about $40,000; customer onboarding and hosting support about $20,000; office rent and utilities about $36,000; sales and marketing about $40,000; software tools, security tooling and compliance audits about $24,000; insurance about $18,000; legal, ip filings and accounting about $30,000; travel and customer visits about $15,000; recruiting about $20,000; and about $18,000 a year for the wear on equipment.

How we planned the budget. Most of our Year 1 cost is pay, because building and evaluating this product takes skilled people. Cloud computing and model hosting, about $90,000 in Year 1, grows quickly as we add customers and test more, so we plan for a 25 percent yearly increase and tie it to subscription revenue. Customer revenue starts small and grows as pilots convert, which is why grants are a large share of Year 1 revenue and fall in later years. We plan for the company to reach a small profit after tax in each year, and that profit is a planning figure, not a promise. Grant money is a planning estimate: the grant line in the plan is the balance needed after commercial or other income, and it is not a promise from any funder.

Research and products in the Year 1 plan depend on the success of the work, and if an award is delayed or not made, we will reduce hiring and scope before we reduce safety, testing or review.

About other funding. We are also applying for small business research awards, state economic development funds and pilot contracts from county and municipal offices. Our plan counts pilot contracts only at the amounts in each signed agreement. These sources are competitive, are awarded for set periods and have strict rules about eligibility, reporting and intellectual property. The amounts in this plan are our planning estimates, not promises, and funders can change their budgets, rules and awards.

One-time startup costs. Startup costs of $102,000 are paid once from startup money, not from the yearly budget: $30,000 for equipment, $8,000 for leasehold improvements, $9,000 for deposits, $12,000 for opening supplies and inventory, $3,000 for licenses and permits, $25,000 for legal and professional help, $6,000 for signage, brochures and a website, $9,000 for background checks, training and first insurance.

How we got our startup number. The Funding Need Calculator in this plan adds the startup costs ($102,000), the cash we need to cover the wait for the first payments (about $134,900) and a cash cushion of about $369,800, which is about three months of Year 1 costs. That comes to $607,000. We start with $60,000 of our own cash and $560,000 in equity and loans from our founders and angel investors that we have committed in writing, $620,000 in all, which covers that amount.

The next five years. Revenue grows from about $1,527,500 in Year 1 to about $3,058,000 in Year 5. Grant money is planned at $787,500, $451,500, $360,500, $422,500, $618,000 in Years 1 to 5, and the grant share of revenue falls from about 52% in Year 1 to about 20% in Year 5 as paying customers grow. Our staff grows from 9 in Year 1 to 19 in Year 5.

Savings. We plan to keep several months of costs in cash. The lowest year-end balance in the plan is about $455,100, at the end of Year 1. Customers and funders often pay after we invoice or submit reports, so we plan for about $183,300 owed to us at the end of Year 1.

Risks and how we will handle them:
- A pilot shows the tool does not help or is not accurate enough. We agree thresholds before each pilot, publish results and stop or change the product when the thresholds are missed.
- A model makes a costly error or treats a group of applicants unfairly. Every finding is confirmed by a reviewer, releases are tested for performance gaps, and a release that fails is not deployed.
- A security incident or data breach. We follow the controls in each contract, complete an outside penetration test before live data is used, carry cyber insurance and have a written incident response plan.
- Procurement cycles are long. We plan with a cash cushion, use pilots that can start quickly and do not assume that a county will buy on a fixed date.
- A larger competitor copies the approach. We build trust through measured results, local checklists and reviewer relationships, and we protect our code and filings where counsel advises.
- Government and grant funding is delayed or not awarded. We plan with a cash cushion, stage hiring so costs follow funding and apply to other funders.

If you are just starting. A new company can begin with a founder-funded prototype, two pilot agreements and a smaller grant request for a narrower test plan, and can apply for accelerator or small business research programs before it seeks a larger grant. Strong pilot results make larger awards and investment easier to win.

Appendix Notes

Documents to attach with this proposal:

- Formation documents, a good standing certificate and the operating agreement.
- Your most recent financial statements, tax returns and current year budget.
- A list of owners, officers and advisors, with their roles and any conflicts of interest.
- Resumes for the CEO, CTO and Applied Research Scientist.
- Your letter of intent for pilots and memorandum of understanding with the council of governments.
- Your model documentation, evaluation plan and pilot thresholds.
- Your security policies and the outside readiness assessment or its engagement letter.
- Your proposed county data-use agreement and privacy terms.
- Your patent filings or invention disclosures, where they exist.
- Proof of general liability, professional liability and cyber insurance, or a quote.
If you don't have any of this yet, no worries - let's continue.

The assumptions behind our numbers (change them in the Financials section to match your own):
- Grant Revenue: $787,500 in Year 1, then $451,500, $360,500, $422,500, $618,000.
- Software Subscriptions (County and Municipal Customers): $650,000 in Year 1, growing to $2,000,000 in Year 5.
- Implementation and Support Services: $90,000 in Year 1, growing to $440,000 in Year 5.
- Pay (a full year, per person, including payroll taxes and benefits): Chief Executive Officer $150,000 (1, 1, 1, 1, 1 people in Years 1 to 5); Chief Technology Officer $170,000 (1, 1, 1, 1, 1 people in Years 1 to 5); Machine Learning Engineer $145,000 (2, 2, 3, 4, 5 people in Years 1 to 5); Software Engineer $130,000 (2, 3, 3, 4, 5 people in Years 1 to 5); Applied Research Scientist $150,000 (1, 1, 2, 2, 2 people in Years 1 to 5); Customer Success and Implementation Lead $90,000 (1, 2, 2, 3, 4 people in Years 1 to 5); Bookkeeper (part-time) $36,000 (1, 1, 1, 1, 1 people in Years 1 to 5).
- Other yearly costs in Year 1: Cloud Computing and Model Hosting $90,000, Model Licenses, Data Labeling and Evaluation Tools $40,000, Customer Onboarding and Hosting Support $20,000, Office Rent and Utilities $36,000, Sales and Marketing $40,000, Software Tools, Security Tooling and Compliance Audits $24,000, Insurance (General, Professional and Cyber) $18,000, Legal, IP Filings and Accounting $30,000, Travel and Customer Visits $15,000, Recruiting $20,000. The other costs grow between 4% and 25% a year.
- Equipment: $30,000 at the start, then $15,000 to $15,000 a year.
- Startup costs: $102,000, paid once from $620,000 in startup money.
- About 12% of each year's revenue is still owed to us at year end.
- Tax: 21 percent federal and state income tax on profit, with a standard rate planning assumption.
- Every dollar amount is an estimate. Replace them with your own quotes and real budget.

Startup Costs