Guide · Working with us

Is it OK to use AI for pitch decks? Yes, if a person stays in charge

We use AI every day. It makes us faster. It doesn't decide your numbers, and nothing reaches you that a person hasn't read.

About 4 minutes to read · Updated 22 October 2026
A small team planning at a whiteboard

The short answer

Yes, we use AI. We say so every time.

Using AI for pitch decks and financial models is fine when a person stays in charge of the numbers and reads every draft before it goes out. That's how we work: AI does the legwork, and Devin or Adrian checks it and signs it off.

We use Claude, an AI model made by Anthropic, to research, draft and build faster. We'd rather tell you up front than have you wonder.

Every message we send you on Upwork ends with the same line:

"I use Claude to help me draft, but the approach and the claims are mine."On every Upwork message we send

That line is the whole policy in one sentence. The tool helps with the work. The judgement, and the responsibility for every number, stays with us.

Who does what

What AI does in our decks and models, and what a person does

Think of AI as a very fast junior analyst. It does the legwork. It never gets the final say.

AI helps us

  • Research comparable companies and market data
  • Write first drafts of slides, memos and messages
  • Build pages, charts and layouts
  • Read our work back and flag anything unclear

A person

  • Picks the method and sets the numbers
  • Checks every formula in your model
  • Checks every source behind a number
  • Reads every draft before it goes out, and signs it off

Investors

Why investors won't see "AI slop"

Founders worry an investor will spot AI and switch off. It's a fair worry. What investors actually reject is work that's generic or wrong: a template with the name changed, a statistic nobody can trace, a deck whose numbers don't match the model.

So that's what we guard against:

  1. Built from your numbers. Your model starts from your own drivers, like your prices, your costs and your hiring plan. It isn't a template.
  2. Every figure traces back. The numbers on your slides come straight from your model, so they always match.
  3. Every outside fact has a source. Comparable deals and market figures come with a link and a date, so an investor can check them.
  4. Every guess is labelled. Where you don't have a number yet, we use an assumption and mark it clearly, so you can correct it.

Our checks

What happens before you see anything

A first draft goes through checks before it reaches you. On a full Smart Deck, that means:

A 33-point audit

A fixed checklist run over every Smart Deck, from the cover to the last number.

Fresh readers

A panel that sees only your deck and nothing else, and flags any word or step an investor would trip on.

A logic check

Do the slides, the memo and the model tell the same story with the same numbers?

A person's read

Then Devin or Adrian reads it through before we send it to you.

Our checks catch errors in our own work. They can't tell us whether something you've told us is true, which is why we label your inputs and ask you to confirm them.

Working with us

More questions about working with us

Who is Hugo?

Hugo is the name of our AI assistant. It drafts, organises and keeps our notes in order. Devin and Adrian review what it produces and send it themselves. You'll always be dealing with a person.

What about my confidential files?

We're happy to sign your NDA before you share anything. If your NDA limits which AI tools can be used on your files, tell us and we'll follow it.

Does AI make it cheaper?

It makes us faster, so you get a first draft sooner and more time goes on checking and on your questions. You pay for the work and the judgement, not the typing.

Your questions

Questions founders ask

Is it OK to use AI for pitch decks?

Yes, as long as a person stays in charge. AI is good at research, first drafts and layouts, but it shouldn't decide your numbers. What investors reject is work that's generic or wrong, like a template with the name changed or a statistic nobody can trace. If every figure ties back to your model and every outside fact has a source, the tool that helped draft it matters far less.

Can AI build a financial model for investors?

AI can speed up the build, but it shouldn't set the numbers. A model investors trust starts from your own drivers, like your prices, your costs and your hiring plan. A person picks the method, sets the assumptions and checks every formula. Where you don't have a number yet, the assumption should be labelled so you can correct it. AI helps with research and layout, and the judgement stays with a person.

Can investors tell if a pitch deck was made with AI?

Sometimes, but what they really notice is generic work. Stock phrases, slide numbers that don't match the model and facts with no source are the giveaways. Investors judge the work. If the numbers hold up, match each other and trace back to a source, the tools behind them matter far less. That's why we label every guess and give every outside figure a link and a date.

Still unsure?

Ask us on our next call

We're happy to walk you through exactly how your project was built, and which parts a person did by hand.

Next, see how we value a company: the valuation guide.