Pitch Reviewfrom Sure Valley Ventures
Contents

The standard · version 1.1.0 · 15 November 2026

Seed deck template

A seed deck for the first meeting. Twelve slides, one message each: the title states the message and everything else on the slide proves it.

For each slide the template gives what investors read there, with a test for each, and the two mistakes they see most. Each read is marked on what the deck says and how plainly it says it, not on how strong the business is: that is read separately. A read is met wherever in the deck its content sits, so folding two of these slides into one is not a fault.

A mistake is marked only where it can be pointed at on a slide, never because something is missing.

Download the template

PowerPoint file (.pptx), 46 KB. Opens in PowerPoint, Google Slides and Keynote. Each slide's notes say what investors read there.

See a worked example: a fictional companyA finished seed deck for a company that does not exist. PDF, 17 pages, 292 KB.

A line marked for AI, not AI, a sales team or self-serve applies only to those companies.

1. Title and one-liner

Says what you do, for whom and to what end, in one sentence a stranger can repeat.

What investors read here

  1. Whether they can say what you do in one sentence after reading it once.

    The test for each mark
    Done
    One of the first two slides has a sentence that says what the product does, in words a stranger would understand.
    Partly
    Anything between: the sentence leans on a term a stranger would need explained, or first comes after slide two.
    Not done
    No sentence in the deck says what the product does: a name, a logo or a tagline only.

  2. Which buyer pays, and for what outcome.

    The test for each mark
    Done
    The first two slides name the customer (a role or a kind of organisation) and the outcome they get.
    Partly
    Anything between: they name one of the two.
    Not done
    They name neither.

  3. Stage, amount and where you are, so the deck is filed in the right place.

    The test for each mark
    Done
    The first two slides give all three: the round, the amount being raised and the city or country.
    Partly
    Anything between: they give one or two of the three.
    Not done
    They give none of the three.

The mistakes investors see most

  • A tagline ("reimagining planning") instead of a sentence with a customer and an outcome in it.

  • A one-liner that names the technology and not the job it does.

2. The problem

Shows what the problem costs one customer today, in their numbers.

What investors read here

  1. Whether the problem is measured in the customer's units: hours, days, pounds, error rates.

    The test for each mark
    Done
    The deck gives at least one figure for the problem in the customer's own units (time, money, a count or an error rate), for one customer or a typical one.
    Partly
    Anything between: the only figures are for the market as a whole.
    Not done
    The problem is told in words only.

  2. Who has the problem, and how they cope today.

    The test for each mark
    Done
    The deck names who has the problem (a role or a kind of team) and says how they deal with it today.
    Partly
    Anything between: it gives one of the two.
    Not done
    It gives neither.

  3. What the problem costs one customer, set beside your price.

    The test for each mark
    Done
    A title or the main text of a slide gives what the problem costs one customer, in money or time, and the deck gives what you charge.
    Partly
    Anything between: the cost is given with no price anywhere in the deck, or the cost is only in a footnote or a source line.
    Not done
    No cost to one customer is given.

The mistakes investors see most

  • A market-level problem ("planning is broken") with no single customer's numbers.

  • Numbers with no source: say which customer, which pilot or which interview they came from.

3. What we built

Shows what the product does, in the customer's words and not as a list of features.

What investors read here

  1. Whether they understand what the product does from one screen, without a demo.

    The test for each mark
    Done
    The deck shows the product at work: a screen, a simple diagram of the steps a user or a piece of work goes through, or one worked example. It says what goes in and what comes out.
    Partly
    Anything between: the product is explained in a paragraph or bullets with no screen, diagram or example; or a screen or diagram that does not say what goes in and what comes out.
    Not done
    Only a list of features or a diagram of the system's parts; or nothing on what the product does.
    Could not see
    Judged from the slide as a picture, not from its text.

  2. Who the daily user is, and what they use it for.

    The test for each mark
    Done
    The deck names who uses the product day to day, by role, and says what they do with it.
    Partly
    Anything between: a user is named with nothing on what they do with it, or the use is described with no user named.
    Not done
    Nothing on who uses the product.

  3. The before-and-after in the customer's numbers, not in features.

    The test for each mark
    Done
    The deck gives at least one result in a customer's own numbers (time saved, errors caught, money saved or found) and says whose it is: a named customer, a pilot or a test.
    Partly
    Anything between: a result with nothing on whose it is, or a change claimed with no number.
    Not done
    No result is given: features only.

The mistakes investors see most

  • An architecture diagram or a feature grid where the product screen should be.

  • Describing what the model is instead of what the user gets.

4. Why now

Shows what changed, and why the window is open now and was not two years ago.

What investors read here

  1. A dated, external change, not an internal milestone.

    The test for each mark
    Done
    The deck names a change outside the company (in regulation, cost, technology or behaviour) and says when it happened.
    Partly
    Anything between: an outside change is named with no date.
    Not done
    The deck names no change outside the company: the argument rests on the company's own progress, or there is none.

  2. Evidence a stranger can check: a regulation, a price curve, a published target.

    The test for each mark
    Done
    The deck gives a number or names a document that shows the change, with its source.
    Partly
    Anything between: a number or a document with no source.
    Not done
    No number and no document is given.

The mistakes investors see most

  • "AI is here" as the whole argument.

  • A trend with no date and no number.

5. Traction

Shows the headline number, its growth, and what counts in it.

What investors read here

  1. The headline number, and whether it is billed and collected recurring revenue.

    The test for each mark
    Done
    The deck gives one headline revenue figure with its date and says what it counts: recurring revenue from customers who pay, under contract or, for a product bought self-serve, by card.
    Partly
    Anything between: a dated figure with nothing on what it counts, or one that says what it counts with no date.
    Not done
    No revenue figure, or one with neither a date nor a word on what it counts; or the deck's own words show that the headline includes pilot or trial fees, one-off fees or contracts not yet signed.

  2. Growth over six months, and whether the recent months are faster than the early ones.

    The test for each mark
    Done
    Revenue is shown for each of at least six months, each month's own figure, so that the recent months can be set against the early ones. Where there is no revenue yet, paying customers are shown the same way.
    Partly
    Anything between: one growth rate or two end points with no monthly figures, fewer than six months, or a running total only.
    Not done
    No growth figure.

  3. Usage and customers staying: how often customers use the product, and whether they renew or leave.

    The test for each mark
    Done
    The deck gives a figure, with its period, for how much customers use the product or for how many have renewed, stayed or left; or it says plainly that no customer has yet come up for renewal.
    Partly
    Anything between: a figure for use or for customers staying with no period, or a claim that customers stay with no figure.
    Not done
    Nothing on use and nothing on customers staying.

The mistakes investors see most

  • Adding pilots, set-up fees and signed-but-unbilled contracts into "ARR".

  • A cumulative chart, which would hide a flat recent quarter.

6. How we sell

Shows who buys, how they buy, and the proof that it repeats.

What investors read here

  1. A written ideal customer: the kind of organisation, and who in it buys.

    The test for each mark
    Done
    The deck describes the customer you sell to by kind and size of organisation and names who buys: a role or, for a product bought self-serve, who signs up.
    Partly
    Anything between: the customer is described with no size, or with nothing on who buys.
    Not done
    No customer is described, or it is "any company".

  2. A process with named stages, and a number that shows it at work: deals at each stage, conversion or the time a sale takes.Sold by a sales team

    The test for each mark
    Done
    The deck sets out the stages a sale goes through, in order, and gives a number for them: the deals at each stage, the conversion between stages or how long a sale takes.
    Partly
    Anything between: stages with no number, or a sales cycle or a conversion rate with no stages set out.
    Not done
    Nothing on how a sale is made.

  3. On what terms customers start: whether pilots are paid and time-boxed, and how long contracts run.Sold by a sales team

    The test for each mark
    Done
    The deck says on what terms a customer starts: for a pilot or trial, whether it is paid and how long it runs; for a contract, its length.
    Partly
    Anything between: a pilot is shown with only one of the two, or customers are shown with nothing on the terms they buy on.
    Not done
    No pilot and no customer is shown.

  4. The path from sign-up to paying, with a count or a conversion rate at each step.Bought self-serve

    The test for each mark
    Done
    The deck shows the steps from visitor or sign-up to paying, with a count or a conversion rate at each.
    Partly
    Anything between: steps with no figures, or figures for some steps only.
    Not done
    No path from sign-up to paying is shown.

  5. Where sign-ups or new customers come from, by source.Bought self-serve

    The test for each mark
    Done
    Sign-ups or new customers for a stated period are split by source (for example search, referral or paid advertising), with a share or a count for each.
    Partly
    Anything between: sources named with no figures, or figures with no period.
    Not done
    Nothing on where sign-ups come from.
    The bar it leans on
    Where demand comes from

  6. What you charge, and what for.

    The test for each mark
    Done
    The deck gives the price and the unit it is charged on: a seat, a user, a unit of work or a flat fee.
    Partly
    Anything between: a contract value or revenue per customer with nothing on what the price is charged on, or the unit with no figure.
    Not done
    No price is given.

The mistakes investors see most

  • Three segments and two geographies at seed, with no beachhead.

  • Free pilots with no end date and no conversion price agreed up front.Sold by a sales team

  • Free users or trials counted as customers.Bought self-serve

7. Market

Shows the market from the bottom up: how many buyers, at what price, and how many you can reach.

What investors read here

  1. A bottom-up number: counts × price, with the assumptions on the slide.

    The test for each mark
    Done
    The market figure is built as a count of buyers times a price, and both the count and the price are in the deck.
    Partly
    Anything between: a bottom-up figure with the count or the price behind it missing; or, for a product bought self-serve, demand shown as counts for a stated period (visitors, sign-ups, paying accounts) with no market figure built from them.
    Not done
    No market figure, or a top-down figure only.

  2. The beachhead: the one segment you sell to first, named and sized, so that a reader can judge whether it supports the next two years.

    The test for each mark
    Done
    The deck names the segment it sells to first and gives its size as a count of buyers or a value.
    Partly
    Anything between: a first segment is named but not sized.
    Not done
    No first segment is named: the whole market only.

  3. Whether the counts have a source, or came from a report's top-down figure divided by something.

    The test for each mark
    Done
    Every figure the market size is built from names where it comes from, so that a stranger could check it: a publication, a register, or your own count or prices.
    Partly
    Anything between: some do and others do not, or the source is not named ("an industry report").
    Not done
    No figure says where it comes from.

The mistakes investors see most

  • A £30bn market with no path from here to there.

  • Counting every organisation as reachable at the same price, unless that price is given as an average.

8. Competition

Shows where you win against every alternative, including the in-house build and doing nothing.

What investors read here

  1. Whether doing nothing and building it internally are named beside the other products; they are the alternatives that win most deals.

    The test for each mark
    Done
    The deck names at least one other product or company and at least one way a customer copes without buying: doing it by hand, building it themselves or doing nothing.
    Partly
    Anything between: it names products only, or ways of coping only.
    Not done
    No alternative is named.

  2. Where each alternative wins and where you win, said for each one.

    The test for each mark
    Done
    For each alternative it names, the deck says what that alternative does well and where you win.
    Partly
    Anything between: where you win is said once for all of them, or with nothing on what the alternatives do well.
    Not done
    Where you win is not said.

  3. Whether the incumbent could add your feature in a quarter, and what stops them.

    The test for each mark
    Done
    The deck names an incumbent, or a supplier the customer already uses, and gives a reason specific to it why it has not done this or could not quickly.
    Partly
    Anything between: an incumbent is named with no reason, or with a reason that would fit any incumbent.
    Not done
    No incumbent is named.

The mistakes investors see most

  • "We have no competitors."

  • A matrix where you have every tick and everyone else has none.

9. The AI and the moat

Shows how you know the product works, what it costs to deliver, and why it gets harder to copy.

For software companies that are not AI companies, the slide is titled "The product and the moat".

What investors read here

  1. A measured accuracy number, with what it was measured on and what it is compared with.AI only

    The test for each mark
    Done
    The deck gives a measured figure for how well the AI does its job, says what it was measured on, and says what it is compared with: people doing the job, a baseline or the method used before.
    Partly
    Anything between: a measured figure with nothing on what it was measured on, or with nothing to compare it with.
    Not done
    No measured figure: "accurate" or "proprietary AI" only.

  2. The cost per unit of work, and gross margin with what its costs include. There is no pass mark: what is published for AI companies is reported with the gross margin measure.AI only

    The test for each mark
    Done
    The deck gives gross margin, or the cost of one unit of work beside its price, and says what the cost includes: the model, hosting and support.
    Partly
    Anything between: a gross margin or a unit cost with nothing on what it includes.
    Not done
    No gross margin and no unit cost.
    The bar it leans on
    Gross margin

  3. A measured advantage over what the customer used before, on a number the customer cares about.Not AI

    The test for each mark
    Done
    The deck gives a figure, says what it is compared with, and says whose it is: a customer, a pilot or a test.
    Partly
    Anything between: a figure with nothing to compare it with, or with nothing on whose it is.
    Not done
    No measured advantage.

  4. Gross margin, with what its costs include. There is no pass mark: what is published is reported with the gross margin measure.Not AI

    The test for each mark
    Done
    The deck gives gross margin and says what its costs include: hosting, support and any services.
    Partly
    Anything between: a gross margin with nothing on what its costs include.
    Not done
    No gross margin.
    The bar it leans on
    Gross margin

  5. What gets harder to copy with every customer: data, integrations, a workflow the customer depends on.

    The test for each mark
    Done
    The deck names what the company holds that others would have to rebuild (which data, which integrations or which workflow) and says how it grows with each customer.
    Partly
    Anything between: an asset is named with nothing on how it grows, or the deck says "data" or "network effects" without saying which.
    Not done
    No asset is named.

The mistakes investors see most

  • "Proprietary AI" with no evaluation set and no number.AI only

  • A moat that is a model provider's moat, not yours.AI only

  • "Proprietary technology" with no measured advantage behind it.Not AI

  • A moat that is only a head start: features an incumbent could add in a quarter.Not AI

10. Team

Shows who builds, who sells and who knows the customer, all of them full-time.

What investors read here

  1. Build, sell and domain each owned by a named founder, and whether any of the three is missing.

    The test for each mark
    Done
    The deck names a founder against each of build, sell and domain, in so many words or by what it says each does. For a product bought self-serve, sell means whoever runs growth.
    Partly
    Anything between: one or two of the three have a named founder against them.
    Not done
    Founders are named with no roles, or no team is shown.

  2. Why these people for this problem: what each has done that bears on it.

    The test for each mark
    Done
    For each founder the deck gives something they have done that bears on this problem: years in the customer's world, a system of this kind built, or customers of this kind sold to.
    Partly
    Anything between: it does so for some founders and not others, or it gives employers, titles and degrees with no link to this problem.
    Not done
    Names and titles only.

  3. The next hires, by role, and when they start.

    The test for each mark
    Done
    The deck names the next hires by role and ties them to the round or to a date.
    Partly
    Anything between: hires are named with nothing on when, or a number of hires with no roles.
    Not done
    No hires are named.

The mistakes investors see most

  • Advisors listed as if they were the team.

  • A founder shown as part-time, or as still holding another job.

11. The plan: twelve months to the next round

Shows what the next round needs to see and when, and how each quarter gets there.

What investors read here

  1. The numbers the next round needs, stated now: ARR, customers, growth rate, and the date.

    The test for each mark
    Done
    The deck states where it aims to be before the next round as at least one number (ARR, customers, growth or retention) and gives the month or year it is due.
    Partly
    Anything between: a target with no date, or a date with no number.
    Not done
    No target is stated.

  2. What the plan assumes, as numbers: customers won, price, spend or hires.

    The test for each mark
    Done
    The deck states at least one assumption its targets rest on as a number: customers won in a period, contract value, spend a month or hires.
    Partly
    Anything between: assumptions are given in words with no number.
    Not done
    Targets are given with nothing on what they assume, or there are no targets.

  3. Whether each quarter has a milestone a stranger could verify.

    The test for each mark
    Done
    Each quarter or half-year of the plan shows at least one milestone a stranger could verify: a number, or a named event with its date.
    Partly
    Anything between: one dated target only, or milestones with no date.
    Not done
    No milestone with a number or a date.

The mistakes investors see most

  • A plan that ends when the money runs out rather than six months before.

  • Milestones with no numbers ("expand to Europe").

12. The ask

States the amount, what it buys and how long it lasts.

What investors read here

  1. Amount stated plainly, as one figure. Where part of it is already committed, say how much.

    The test for each mark
    Done
    The deck states one amount for the round, not a range.
    Partly
    Anything between: the amount is a range.
    Not done
    No amount is stated.

  2. Use of funds by percentage, in a split that adds up.

    The test for each mark
    Done
    The deck splits the money by use, in percentages or amounts that add up to the whole.
    Partly
    Anything between: a split that does not add up, or uses listed with no split.
    Not done
    No use of funds.

  3. Runway from the close: how long the money lasts.

    The test for each mark
    Done
    The deck gives the runway the round buys, in months or to a date.
    Partly
    Anything between: spend a month is given with no runway.
    Not done
    Nothing on how long the money lasts.

The mistakes investors see most

  • "Raising £1–£2m" with no use of funds.

  • A valuation, a cap or other terms of the round printed in the deck.

Appendix

Slides kept behind the main deck, for the questions it will raise.

Revenue month by month
Twelve months of revenue and customers, with what counts in the headline and what is shown apart.
The market model
The counts, the prices and the sources behind the market slide.
The pipeline
Open deals by stage, marked as pipeline and not as customers.
The evaluation
For an AI product: the evaluation set, the baseline and how often it is re-run.
The cash plan
Month by month from the close to the next raise, with the assumptions it rests on.
Customer stories
One or two customers, each with its numbers before and after.
The cap table
Who holds what today, and the rounds so far.
Questions you have been asked
One slide for each question an investor has asked more than once.

Back to the top