Investor Update Template: Draft Monthly Updates From Metrics With AI

A startup can draft a monthly investor update by keeping a metrics sheet with the same rows each month, pasting this month and last into ChatGPT or Claude, and asking for a short update in a fixed shape: numbers, wins, risks, asks. The founder checks every figure against its source before sending, because AI will tidy a bad month.

A static template gets you the headings. The workflow below gets you the draft: the sheet does the arithmetic, the AI turns the arithmetic into prose, and a short review checks that the prose says what the numbers say. The result is an update you can send in about thirty minutes, written the same way every month so that investors learn where to look. The aim is not polish. Investors want to see the numbers, hear about problems early and know what you need from them.

Follow me on Instagram@sagnikteaches

Start with a sheet that does the sums

The most common failure in AI-drafted updates is arithmetic done in the chat. Language models handle words better than long calculations, and a wrong growth figure in an investor email is expensive in trust. Do the sums in a spreadsheet, and give the AI finished numbers only.

Connect on LinkedInSagnik Bhattacharya

Here is a metrics sheet for an illustrative software-subscription startup. All figures are invented for the example.

Subscribe on YouTube@sagnikteaches

The sums behind it, so you can see what the sheet should compute: MRR bridge is 42,000 + 4,200 + 900 − 1,740 = 45,360. Growth is (45,360 − 42,000) / 42,000 = 8.0%. Customers are 210 + 17 new − 6 cancelled = 221, which means 17 new customers at an average of about $247 each ($4,200 / 17). Customer churn is 6 / 210 = 2.9% for the month, and churned MRR is 1,740 / 42,000 = 4.1%. Runway is cash divided by net burn: 310,000 / 38,000 = 8.2 months. Put each of these in a formula cell, and give the AI the results. For help with the spreadsheet side, the ChatGPT startup metrics tutorial shows how to read the numbers before you write about them.

Add a second small table for qualitative items you always include: three wins, three problems, any hires or departures, and what you asked investors for last month with its status. A shape that stays fixed from month to month is what lets an investor skim.

The shape of the update

A reliable structure puts the most important things first and the same sections in the same order every month:

1. One-paragraph summary. Headline numbers and the single biggest thing that happened, including bad news.
2. Metrics table. The table above, trimmed to the five or six numbers you stand behind.
3. Wins. Two or three, each with a number or a named outcome.
4. Problems and risks. Honest, with what you are doing about each.
5. Asks. Specific and answerable: an introduction to a named kind of person, feedback on a decision, a recommendation for a hire.
6. Next month. Two or three commitments you can be held to.

The order matters. Problems come before asks because an ask without context looks like a favour, and an ask beside a problem looks like a plan.

The prompt that turns the sheet into a draft

The prompt does three jobs: it sets the voice, it forbids invention, and it makes the AI show its checks. Keep it in a saved Project so you paste only the new month's numbers.

You are helping a founder write a monthly investor update. Use ONLY the numbers and facts below; do not calculate new figures, round differently or add any fact not provided. If a number you need is missing, list it under "MISSING" and do not guess.

Structure: (1) a summary paragraph of 60-80 words that states the headline MRR figure, the change, and the most important problem; (2) a metrics table copied from my sheet; (3) Wins (2-3 bullets, each with a figure); (4) Problems and risks (each with what we are doing); (5) Asks (specific and answerable); (6) Next month (2-3 commitments).
Voice: plain, direct, first person plural, no hype, no adjectives like "exciting", "strong", "incredible". Do not describe a number as good or bad unless I do. Under 450 words excluding the table.
After the draft, add a section "NUMBERS CHECK" listing every figure used in the prose and the table cell it came from.

THIS MONTH'S SHEET: [paste table]
WINS: [paste]
PROBLEMS: [paste]
ASKS: [paste]
LAST MONTH'S ASKS AND STATUS: [paste]

The "NUMBERS CHECK" section is the most useful line in the prompt. It gives you a list to compare against the sheet in two minutes, and it exposes a figure the AI changed or invented because it cannot point to a source cell.

A filled-in draft, and what to fix

For the sheet above, plus these notes (illustrative): wins were a signed annual contract worth $9,600 a year and a faster onboarding flow that cut set-up from 4 days to 1; problems were a rise in churn after a price change and a delayed integration; asks were an introduction to operations leads at mid-sized logistics firms and feedback on whether to hire a customer success lead now. A typical draft:

Subject: October update: MRR $45.4k (+8.0%), runway 8.2 months

Summary
MRR reached $45,360, up 8.0% from $42,000. Churned MRR also rose, to $1,740 (4.1% of opening MRR), after we raised prices on the starter plan. Net burn was $38,000 and runway is 8.2 months at that rate.

Metrics
[table]

Wins
- Signed one annual contract worth $9,600 a year.
- Cut customer set-up time from 4 days to 1.
- Added 17 new customers; 221 at month end.

Problems and risks
- Churn rose after the starter-plan price change: 6 customers cancelled ($1,740 MRR). We are calling each one to confirm the reason and will report what we hear.
- The accounting integration slipped from October to November. We will confirm a date by the 15th.
- Runway is 8.2 months. We plan to decide on new funding by February.

Asks
- Introductions to operations leads at mid-sized logistics firms.
- Your view on hiring a customer success lead now or after the integration ships.

Next month
- Reduce churned MRR below $1,200.
- Ship the accounting integration, or say why not.
- Report the results of the six cancellation calls.

NUMBERS CHECK
MRR 45,360 and 42,000: sheet, MRR row. Growth 8.0%: sheet, Change column. Churned MRR 1,740: sheet. 4.1%: sheet (calculated). ...

What a founder fixes in a draft like this: first, the sentence "after we raised prices on the starter plan" states a cause for the churn. The notes said churn rose after a price change, which is a sequence and not yet a proven cause, so the honest phrasing is "following", and the cancellation calls are how you find out. Second, "Reduce churned MRR below $1,200" is a commitment the AI invented from the churn figure; keep it only if you will stand behind it. Third, "We plan to decide on new funding by February" is not in your notes at all: the AI inferred it from the runway. An invented plan in an investor update is a serious error, because investors will hold you to it. Delete it, or replace it with the real plan. All three are the usual kind of error: plausible, helpful-sounding and wrong.

Writing the honest version of a bad month

AI drafts lean positive. Left alone, it reaches for words such as "continued momentum" for flat numbers and passes over the line that matters. Counter this in the prompt and again in review.

Suppose the next month goes badly (illustrative): MRR falls from $45,360 to $44,100 after two customers who together paid $2,200 a month left and new sales covered only part of it, cash drops to $270,000 on a net burn of $40,000, so runway is 6.75 months. The instruction that works: "This was a down month. The summary must state the MRR fall and its main cause in its first sentence, using no positive framing. Do not use words such as 'resilient', 'stabilising' or 'momentum'. After the problems, include a short 'What we have decided' paragraph that I will fill in."

A reasonable first sentence from the draft: "MRR fell 2.8% to $44,100 because two customers paying $2,200 a month in total cancelled and new sales did not fully replace them." The sum is (45,360 − 44,100) / 45,360 = 2.8%. It is blunt, and a founder can add context after it, but the first sentence has done its job. Runway of 6.75 months is under seven months and belongs in the headline too, because it affects what investors can do about it, and every month of notice helps. The point is trust: an investor who is told early and plainly can help, and one who finds out later cannot.

Wins that are numbers, not adjectives

Founders often write wins the way they would write a press release, and AI amplifies the habit. Investors read the update to learn what changed, so each win needs a before and an after. Ask the AI to rewrite any win that lacks both, and to leave a blank where the figure is missing.

Before (as first written): "Great progress on onboarding, customers are loving the new flow." After (illustrative): "Onboarding now takes 1 day, down from 4. Of the 17 customers who joined this month, 14 were live within 2 days." The second version gives the investor something to remember, to quote to someone else and to check next month. If you do not have the number, the honest option is to say what you will measure and by when, not to find a flattering word.

A short prompt for this pass: "Here are my draft wins. For each, say whether it contains a before and after figure. For those that do not, write the question I should answer to supply one. Do not rewrite them with invented figures."

Asks investors can actually act on

An update with no asks wastes the main thing investors can give you: introductions, advice and time. A vague ask ("any introductions welcome") gets nothing. A specific one tends to get a reply, because the investor can picture who to forward it to. Three versions of the same ask show the gap:

• Weak: "Looking for customer introductions."
• Better: "Looking for introductions to operations leads at logistics firms."
• Best: "Looking for introductions to operations leads at logistics firms with 50 to 500 staff who run a shared inbox for delivery queries. I can send a three-line blurb you can forward."

The last one names the person, the size and the pain, and removes the effort from the investor's side by offering a forwardable blurb. AI is good at producing that blurb, and at turning a vague ask into a sharper one: "Rewrite each ask so it names who, what size, and what I will send to make it easy. Keep it under 40 words. Do not add a company name I have not given you." Then check that the person described exists; an ask for a type of contact that nobody has is a wasted line.

Close the loop the following month. List last month's asks with their status, such as "Two introductions received, one meeting held, one prospect trial started". Investors who see their help produced a result are more likely to help again, and those who never hear back stop trying.

Pulling a quarterly view out of the monthly ones

Every third update, or once a quarter, a founder usually wants a longer view: how did the three months compare with the plan? If the sheet keeps the same rows, this is a short job. Paste the three months of data and ask: "Compare these three months. State the MRR change over the quarter and per month, the average monthly net burn, and the trend in churned MRR. Do not forecast. List any month where a number moved by more than 10% and ask me what caused it." The final sentence matters: AI can spot the unusual month, and only you know the reason.

Keep the quarterly note to a page, and add the one thing the monthly update cannot: a plain statement of whether you are ahead of or behind the plan you gave at the last funding round or board meeting, using that plan's own figures. If you changed the plan, say so. An AI draft should never reconcile a missed target by quietly shifting the goalposts; add the instruction "compare against the original targets I provide, not against adjusted ones".

Other kinds of company, same sheet idea

A subscription box business. Replace MRR with monthly subscribers and box margin: subscribers at month end, new, paused, cancelled, and gross margin per box after shipping. The update says which supplier price changes hit the margin this month.

A marketplace. The headline numbers are gross merchandise value, take rate and active buyers and sellers. The risk section watches the ratio between the two sides, since growth on one side with none on the other is a warning.

A hardware startup. The sheet carries units shipped, units in stock, build milestones and cash tied up in components. The asks are often about suppliers and introductions to manufacturers, and the risk section is about lead times.

A services-heavy startup. Revenue is lumpy, so include billed revenue, bookings, utilisation and the pipeline value by stage. The AI prompt adds "never present pipeline as revenue".

The common thread is a sheet with consistent rows and a prompt that forbids invention. For finding which numbers are worth reporting at all, see the AI KPIs tutorial.

Review before sending: a ten-minute routine

This is the control that makes the whole workflow acceptable. Do it every month, in this order.

1. Numbers. Compare every figure in the prose and table with the sheet, using the NUMBERS CHECK list. Recompute growth and runway by hand once.
2. Source of the sheet. Confirm that cash matches the bank balance, and that MRR matches your billing system, not a forecast.
3. Causes and promises. Underline every "because", "will" and "plan to". Each must come from your own notes.
4. Adjectives. Remove any positive word you would not defend if an investor challenged it.
5. Last month's asks. Did you report back on them? Investors who made introductions want to know what happened.
6. Confidentiality. No customer names, personal data or details covered by a contract.
7. Read it aloud. If a sentence sounds like nobody you know, rewrite it in your own words.

Teams of two or more can split the work: one person owns the sheet, another reads the draft against it. A fresh pair of eyes catches the confident invention that the author of the notes reads straight past.

Handling the data safely

Revenue, cash and runway are sensitive. Choose a plan whose data settings you understand: business plans of the main assistants do not train on your content by default, and individual plans have a setting to switch training off, so check the setting is off before pasting figures. Strip names of customers, employees and counterparties from anything you paste: "a logistics customer" is enough for the AI to write a sentence. If an investor has asked for confidentiality, or a term sheet restricts disclosure, keep those details out of the tool altogether. The team plan comparison for startups covers the plan choice, and tool cost is worth watching while you are managing runway: see which startup credits are really claimable.

Making the routine stick

Pick a fixed date, such as the fifth working day of the month, once the books have closed. Put three blocks in the diary: day 1, update the sheet and check it against the bank and billing system (30 minutes); day 2, run the prompt and review the draft (30 minutes); day 3, send. Keep every sent update in one folder so next month's draft can start from this month's. Paste the previous update into the prompt as an example of tone and structure, and say "match its layout; do not repeat its sentences". Over a few months, add two or three things to the prompt as you notice them, for instance a line such as "always show churn as a percentage of opening MRR, not of closing".

The test of whether it works is simple. Are investors replying with specific help, such as introductions and advice, more often than before? Are you spending less than an hour a month on the update, and sending it on the same date each time? Has an investor ever caught a wrong number? If the last answer is yes even once, tighten the review step before you tighten anything else. Consistency is the thing investors reward, and the sheet, the prompt and the review are what give it to you without making the update a monthly dread. If you want to go further on the numbers side, automating monthly management reports shows how to have the sheet fill itself from your accounting and billing tools.

Questions founders ask about AI-drafted investor updates

Should I tell investors I used AI to draft the update?

There is no rule that you must, but the content is yours and you are answerable for every number. If an investor asks, say so plainly. What matters more is that the figures are right and the tone is honest, since investors read these updates to judge whether they can trust you with bad news as well as good.

How long should a monthly investor update be?

Short enough to read in three minutes: roughly 300 to 500 words plus a small metrics table. A long update is not a better one. Put the headline numbers and the asks near the top and detail beneath, so someone who reads only the first screen still knows where you stand.

Is it safe to paste revenue and cash figures into an AI tool?

Use a business plan or an account where training on your content is off, and paste only what the update needs. Leave out customer names, personal data and anything under a confidentiality agreement. If investors share your confidentiality worries, tell them which tool you use.

What if the month was bad?

Say so in the first paragraph, with the cause and what you are doing. Investors who learn about a miss from a buried line lose confidence faster than those who hear it directly. AI can help you word a difficult update calmly, but should never soften the numbers.

Further reads

• AI Startup Credits and Discounts a Small Firm Can Actually Claim — Cut the cost of AI tools while you are managing runway.
• AI HR Software for Startups: Which Tool Fits Which Job — Plan the hire an update might ask investors about.
• Startup AI Tools for Growth: 8 That Earn Their Subscription — Which AI subscriptions earn their cost at a startup.
• Apollo.io for Startups: Set Up AI Prospecting Step by Step — Build the pipeline your next update will report on.
• Virtual CFO Pricing: Build a Monthly Package With AI-Prepared Reports — See how monthly finance packages are structured.
• AI Accounting for Startups: Monthly Close, Burn and Runway — A founder's monthly close routine that connects checked books to cash burn, runway and accountant review.
• AI Agents for Startup Fundraising: Investor Lists, Outreach, Questions — A founder-controlled investor research and outreach workflow with a source log, approval gates and data-room checks.
• Can You Automate a B2B SaaS Business With AI? What to Hand Over First — A risk-ranked order for handing small B2B SaaS jobs to AI: support drafts first, onboarding next, billing chases last.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: ChatGPT Business pricing page and OpenAI Data controls help article, Claude pricing page (training defaults, read November 2026); standard SaaS metric definitions (monthly recurring revenue, churn, burn, runway) as used in common investor reporting.