We build AI systems for founder-led companies.
You started this company, and you still answer the email. We build AI systems that take the repeating work off you and your people: replies drafted, documents assembled, reports that write themselves, with a person approving anything that goes out.
At a glance
- What it is
- AI systems designed, built and installed in the tools your company already uses, with a person approving anything that leaves.
- What it costs
- $10,000 to $20,000 for the AI readiness assessment, credited in full if you go on. $25,000 to $75,000 to build. Or start with one workflow, $7,500 fixed. Support after launch is optional, $3,000 to $10,000 a month.
- How long
- Two to three weeks for the assessment. Six to twelve weeks to build. About three weeks for one workflow.
- What you keep
- The workflows, prompts, configuration and documentation, in your own accounts.
- Who leads it
- Anthony Clemenza leads every engagement.
- The legal side
We build systems. We do not tell you what any law or rule requires of you; that is for your own counsel.
Anthony Clemenza also owns Clemenza Law Group, a separate New York law firm that advises companies on the legal side of AI, like AI policies and vendor contracts. You hire it separately, under its own agreement and fees: clemenzalaw.com, (212) 365-4875.
What we build for a founder-led company.
- 01
Inquiries that get read, sorted and answered in draft
Customer and partner emails are read, sorted by what they need, and answered in a draft you edit and send. Nothing goes out on its own.
- 02
Documents from your own templates
Proposals, quotes, onboarding packets and the paperwork you send over and over, assembled from templates you supply and own.
- 03
A weekly report that writes itself
A plain summary of what happened this week, drawn from your own records, in your inbox before Monday's meeting.
- 04
An assistant that knows your company
Your staff ask a question and get an answer drawn from your own documents, with the source named. When your documents do not settle it, it hands the question to a person.
- 05
Update drafts for customers, partners and investors
First drafts of the regular updates you send, built from your own numbers and notes, for you to correct and approve.
- 06
Operating controls
Who may use which tool, what each tool can see, a log of what happened, and access that ends the day someone leaves.
Your decisions stay with you. The systems draft, sort and assemble; your people decide what is said, what is sent and what is signed. We do not make business decisions for you, and we do not tell you what any law or rule requires of you. That is for your own counsel.
What we have already built.
These run in companies our principal owns or leads. They show what we can build, not results for anyone else.
A client portal you can tour
Each matter's stages in plain language, documents, messages and invoices, on sample data, with no sign-in: clemenzalaw.com/tour.
Tamper-evident signing
A document is fingerprinted when it is issued, and a signature is refused if a word has changed.
Draft-and-approve workflows
A model drafts, a person approves before anything reaches a customer, and every draft is logged.
Questions people ask before they begin.
How much does AI implementation cost for a founder-led company?
$10,000 to $20,000 for the readiness assessment, then $25,000 to $75,000 to build, fixed in writing once the scope is agreed. The assessment is credited in full against the build. If you would rather start with one job, the First Workflow is $7,500 fixed, and that fee is credited against Implementation if you continue within six months.
Where should a founder start: tools, a policy, or a process map?
Start with one repeating job, not a tool. Pick the piece of work that eats the most of your week and look at how it runs today. A tool follows from that. A process map is what the The Diagnostic produces. A written policy is a separate question for your own counsel.
What is an AI readiness assessment for a founder-led company, and what do I get?
It is two to three weeks of study and a written report. We interview you, walk through how your company really runs day to day, and write down where AI could take work off your people and where it should stay out. You also get a ninety-minute review with Anthony Clemenza. The report is yours to use with us or without us.
How long does it take to get AI working in your company?
About three weeks for one workflow, and six to twelve weeks for a full implementation. For a full build, the two-to-three-week assessment comes first. Your part is a few conversations, access to the people and records we need to see, and your approval at each step. We confirm the dates in writing before work starts.
What does it mean that you design, build and run the system?
We do three jobs. We design the system with you, build it into the tools you already pay for, and then run it with you for as long as you want us to, with a person approving anything that matters. You are not buying a lecture about AI or a piece of software. You are buying a working system that does a named job.
What does the first 30 days look like?
In week one we talk through how your business runs and pick the job. In weeks two and three we build it and you try it, with a person approving every output. After about three weeks you have one workflow running, a written guide, and a choice about what comes next. That is the First Workflow, $7,500 fixed.
How do we avoid a pilot that never ships?
Agree in writing what working means before anything is built, and choose a job you can finish. The First Workflow, $7,500 fixed, is built that way: one job, about three weeks, a person approving, and a written guide at the end. A fixed fee and a fixed scope keep a project from drifting.
Can you work inside the tools we already use?
Where the tool allows it, yes. During the assessment we list what you use today, such as email, shared documents, your customer list and your accounting software, and say which can be connected safely and which cannot. We build around what you already have before suggesting you replace anything, so your people are not learning a new system.
What happens to our confidential and customer data when we use ChatGPT, Claude or Copilot at work?
It depends on which plan and settings you use, and each provider's terms change, so we read the current terms for the plan you use before we build on it. We set your system up on business terms and tell you in writing what each tool can see. A personal chat account is not a good place for customer records.
Will our data be used to train anyone's model?
No, not under the business terms we set model access up under. OpenAI's API data policy says data sent to its API is not used to train its models unless you opt in, and Anthropic's commercial terms say it may not train models on your content. Terms can change, so ask us which providers a build uses before you sign.
Who owns what you build?
You do. The workflows, prompts, configuration and documentation are yours at handover, in your own accounts, and you can run them without us. Operate, our monthly support, is optional and ends on notice, so leaving us never means losing what we built. Your data stays in your own accounts too.
Do we need to hire someone technical to keep it running?
No. We build inside tools your people already use and write down how each system works, so a technical hire is not a condition of using it. If you would rather not look after it yourself, Operate is $3,000 to $10,000 a month, covers monitoring and monthly tuning, and ends on notice.
What should we ask an AI vendor before signing?
Ask what happens to your data, who can see it, whether it is used to train models, what it costs as you grow, and how you leave with your data. Those are business questions we can work through with you. Reading and negotiating the contract itself is for your own counsel.
Is a fractional AI lead worth it, or is a project enough?
That depends on how much you want to change. For one job, a project is enough, and the First Workflow is $7,500 fixed. For several jobs, Implementation is $25,000 to $75,000, and Operate, $3,000 to $10,000 a month, is there if you want ongoing care. We will tell you which fits, even if it is the smaller one.
Do we need an AI policy before we start, and do you write it?
We do not write AI policies or tell you what the rules require of you. That belongs to your own counsel. Anthony Clemenza also owns Clemenza Law Group, a separate New York law firm that advises companies on the legal side of AI, like AI policies and vendor contracts. You hire it separately, under its own agreement and fees: clemenzalaw.com, (212) 365-4875.
Start with an assessment, or with one workflow.
Based in New York and South Florida. We take engagements anywhere; travel outside those areas is billed at cost.
- The Diagnostic
$10,000 to $20,000, two to three weeks
- Implementation
$25,000 to $75,000, six to twelve weeks
- Operate
$3,000 to $10,000 a month, ongoing
- The First Workflow
$7,500 fixed, about three weeks