How to use this checklist
This checklist is for companies building products on top of AI models — whether third-party, open-source or their own fine-tuned versions.
Work through each list before you contact investors. Anything you cannot tick is either a task to finish or a risk to explain openly in your pitch.
1. Company and legal basics
These apply to every company raising money, but gaps here slow rounds more than almost anything else.
- Up-to-date cap table showing every shareholder, option and convertible.
- Signed IP assignments from all founders, staff and contractors.
- Articles of association and any shareholder agreement to hand.
- SEIS or EIS advance assurance applied for, if you are raising from UK angels.
- Company filings at Companies House up to date.
2. Applied AI proof points
The evidence investors in this sector look for first.
- A named buyer and budget for the problem you solve.
- Paying customers, pilots with agreed conversion terms or signed letters of intent.
- An evaluation set and quality metrics you track over time.
- A written note on what makes the product hard to copy beyond the model.
3. Numbers and financial model
Your model should show how the money you raise gets you to the next milestone.
- Gross margin per customer after model and hosting costs.
- Monthly model spend and how it changes with usage.
- A plan for when cloud or model credits expire.
- Runway to the next milestone with the amount you are raising.
4. Data room documents
Have these organised in one shared folder before the first meeting.
- Data source inventory with licences and consent terms.
- Customer contracts, especially data-use clauses.
- Model provider terms and any usage limits.
- Security and data protection policies.
5. Pitch and investor readiness
Readiness is also about how you run the process.
- A short deck that states the problem, customer, traction and ask in the first few slides.
- A clear amount to raise and a list of what it pays for.
- A target list of investors who back your sector and stage.
- A one-paragraph answer to 'why now?'.
Red flags to fix before you pitch
Investors often stop at these issues:
- Training data scraped without clear rights.
- Margins that only work on free credits.
- No way of catching incorrect outputs.
- A product a model provider could ship as a feature.
Applied AI readiness: ready versus common gap
| Area | Ready when | Common gap |
|---|---|---|
| Data rights | Every source documented with its licence | Unknown or scraped sources |
| Quality | Tracked evaluation metrics | Anecdotal demos only |
| Economics | Margin after inference known | Model costs ignored |
| Defensibility | Moat visible in data or workflow | Relies on prompts alone |
Questions to prepare before you pitch
- Can we prove the right to use every dataset we rely on?
- What is our margin per customer after model costs?
- How do we detect and fix bad outputs?
- What would we do if our model provider changed pricing?
How KJ Enterprises evaluates applied ai businesses
KJ Enterprises reviews AI applications against these same points — data rights, quality control, margins and defensibility — informed by the AI products we build through AutoThink Group.
If that describes your company, you can apply for investment. Related reading: the complete funding guide and the step-by-step how-to and costs and terms and the machine learning checklist.