Day 11 · How to survive a Demo Day

By Karthik Nagapuri · 2026-09-24 · Draper Founder Program #2

A founder's guide to Demo Day and investor Q&A, built from a live panel: last-hour rules from the wings, the questions investors really asked, what each one was testing, how to answer under pressure, and a worksheet for your own four minutes.

Demo Day arrived on Thursday, and the room felt different from every other day of the programme. Same fifth-floor space, same faces, but the chairs had been turned to face a panel, the cables were being checked twice, and every founder was quietly rehearsing a first sentence under their breath.

If you have never been to one, a Demo Day looks like a performance. After sitting through this one, I think it is closer to an exam where you are allowed to see most of the questions in advance, if you know where to look. This chapter is my attempt to show you where to look.

A note on sources. My recording from the day is partial. It caught the backstage advice, the opening remarks and panel introductions, and the first three pitches with their Q&A. I am not going to fill the gaps from memory. So the first half of this chapter is what I actually heard, told as lessons. The second half is clearly labelled reflection: a playbook for preparing, answering and following up, built from those questions. The founders who pitched are kept anonymous, and their numbers are left out. Their stories are theirs to tell.

What a Demo Day is actually for

Each founder had about four minutes to pitch and five minutes for questions. That ratio tells you most of what you need to know. The panel gives you slightly less time to talk than it gives itself to probe. The pitch buys you the questions. It does not close anything on its own.

The panel that day was a good cross-section of who you will meet at any Demo Day: Aaliyah Barker, principal investor and program director at Draper University, who had run our investor night, one-on-ones and Demo Day prep; Rohit, who leads startup banking at IDFC First Bank; Srinivas Kollipara, whom we had first heard at the Day 2 fireside, a long-standing builder of the Indian startup ecosystem; an investment director from a Hyderabad fund; and a community lead from the ecosystem. An investor, a banker, an ecosystem builder, a fund manager and a community person. Five different lenses, five different reasons to be interested, and five different kinds of question.

The session opened with a speaker from the bank's startup programme describing how they support early founders: outreach programmes, connections to investors, joint programmes with venture firms, and help with unglamorous basics like banking and visibility. It was a useful reminder that a Demo Day panel is not only investors deciding yes or no. It is a room full of people who each have a way to help you, if the four minutes make them want to. He closed with a line that got a laugh of recognition from every founder in the room.

Today is difficult, tomorrow is much more difficult, the day after tomorrow is beautiful. But most people die tomorrow. So never give up.

• quoted from Jack Ma at the opening

The organisers had one lovely touch. Every founder was introduced as a superhero: Captain America for the first pitch, Superman ("resilience, purpose and moving forward under pressure") for the second, Aquaman ("navigating complex environments and different worlds") for the third. Cheesy, yes. But after eleven days of being told everything that could kill our companies, it was a kind reminder of something founders forget on stage: the room is mostly on your side. Panels want you to be good. It makes their afternoon better.

The last hour: three rules from the wings

Before the pitches started, there was the usual huddle of founders and mentors near the front. Three pieces of advice I heard there were worth more than most workshops, and they apply to every pitch you will ever give.

Rule one: never stream your demo

A founder was about to play their product video straight from Canva over the venue network. A mentor stopped them: download it and play it locally. It sounds trivial. It is the difference between a demo and four minutes of a spinning wheel in front of investors. Every founder has a story of Wi-Fi betraying them at the worst moment, and the fix costs two minutes.

Generalise it and the rule becomes: remove every dependency you do not control. The venue network, the venue laptop, a login that might expire, a cloud service that might be slow, an HDMI adapter you assumed would be there. If your demo is live software, have a recorded version ready as a fallback. If it is hardware, charge it the night before and carry a spare cable. None of this is glamorous. All of it is cheaper than a failed demo.

Rule two: every claim on a slide is a promise

The second piece was about slides. Anything you put on a slide, especially logos, programmes, partners and associations, is a claim, and investors will check it. The mentor was passing on a point Aaliyah had made to founders earlier, and it was sharp: if one claim turns out not to be true, it does not just cost you that line. It makes the panel question every other claim, including the ones that are completely solid.

That asymmetry is worth sitting with. A true, modest slide earns you a little credibility. A slide with one inflated logo can cost you all of it. So the advice was simple: only put on the slide what you can defend in one sentence, and remove anything you cannot.

In practice this means auditing your own deck like a sceptic. "Partnered with" should mean a signed agreement, not a friendly call. "Backed by" should mean money in the bank, not a conversation. "First" and "only" should survive a two-minute search. If you are not sure, soften the word or delete the line. Nobody has ever lost a deal because their slide was too honest.

Rule three: protect your work, and document it

The third was about protecting what you have built. A mentor warned founders not to overshare technical details (the design, how the product actually functions, how you are training your models) with people building in the same category, however friendly the evening gets. Accelerator cohorts are warm, generous places, and two founders in adjacent spaces will naturally end up comparing notes late at night. That is often good. But if a conversation has to go deep into how your product works, get an NDA, or keep it at the level of what, not how.

And separately from that, keep a record.

Document, document. If it's not documented, it never happened.

• a mentor, just before the pitches

He meant it practically: dated photos of prototypes, dated design files, notes of when you built what. Not a full disclosure of your product, just a trail that shows when things existed. He said a habit like that had once protected him against a false claim. The phone in your pocket already timestamps every photo. The discipline is simply to use it, and to keep the files somewhere you will find them in two years.

Read the question behind the question

Now the pitches. I want to walk through the questions rather than the products, because every question a panel asks is testing something specific, and once you can see what is being tested, you can prepare for it.

Case one: a deep-tech hardware pitch

The first founder presented an anti-drone defence system: a fixed-wing VTOL platform with an AI "brain" that decides what kind of threat it is looking at and responds, with a plan to license the AI to hardware partners. The demo video showed test targets and a return-to-base sequence, and the founder added, with a grin:

This is not AI generated.

• a founder, during his demo video

It got a laugh, and it was smart. In a year when every video might be synthetic, he answered the first doubt in everyone's head before anyone could voice it. That is a skill worth copying: pre-empt the obvious doubt inside the pitch itself. He also finished in exactly four minutes, which drew its own small round of appreciation. Time discipline is noticed.

Then the questions, and what each one was really asking.

• "How do you plan to get this into the hands of governments, and which governments first?" This is the go-to-market question, and for anyone selling to governments, large enterprises or hospitals it is the hardest one. The panel wanted to see that the founder knew the door, not just the destination. When the founder said some of his advisors were confidential and he would not name them, the panellist simply said "of course." A lesson hidden in there: you are allowed to protect sensitive relationships, as long as the rest of the answer shows a real plan.
• "Which components are built by you?" For any hardware company this is the credibility question. What is yours, what is borrowed, and where do the chips and motors come from? The founder was candid that early testing used partner hardware and imported parts, because local components are expensive at this stage. Candour worked better than any spin would have.
• "You said you'd license the algorithm. Have you actually started trying to license it?" This one separates an intention from a behaviour. A business model on a slide is a hypothesis until someone has said yes to it. The founder described active conversations with potential hardware partners. Whatever your answer, expect this question for every revenue model you show.
• "What will you do with the money?" The use-of-funds question. The answer was demo hardware, full-time hires, and IP and legal work. Clear buckets, tied to a milestone. That is the shape a panel wants.
• "Are you building for India? Have you looked at who else is doing this here?" Two questions in one: market focus, and whether you respect the competition enough to have studied it. The founder's answer was generous. He said other teams in the same space had invited him to demonstrate alongside them, and that the shared goal was to make the country safer.
• "The math is not matching." One panellist pressed on whether the drone's size and operating altitude really matched the threat it was meant to counter. This is the moment most founders dread: a direct technical challenge, in public. He did not get defensive. He walked through terrain, altitude, endurance and targeting, step by step, until the reasoning was on the table.

What struck me was how few questions were about the technology itself. Go-to-market, sourcing, whether licensing had begun, use of funds, competition. The panel took the demo more or less at face value and spent its time on whether the business around it was real.

His closing line had a nice ring to it, paraphrasing an old saying: strength defeats strength, but they were adding intelligence to strength.

Case two: when your product sounds like science fiction

The second founder pitched glasses that let you control your phone without touching it: unlock it by looking, answer a call, turn on a light when you look at it. The technology builds on sensor research his co-founder had published.

The first question named the elephant directly: the technology seems futuristic, "kind of like Black Mirror", so tell us about the research and how you will scale it into something that works. Here is the general lesson. When your product sounds like science fiction, the burden of proof shifts onto you. The panel will not ask whether people want it. They will ask whether it can exist. Your pitch needs a short, confident lineage for the core technology: where it came from, what has already been demonstrated, and what remains to be proven.

The second question was, for me, the best exchange of the day.

I think there are maybe two different markets here, and I love that you focused on one.

• a panellist, to the second founder

The panellist praised the focus, then pointed at the other market: people who cannot easily use their hands, for instance stroke survivors with tremors. The founder said they had already built a feature aimed at exactly that group.

That exchange was a small masterclass in how good investors give advice. They did not tell the founder to switch markets. They affirmed his discipline and, at the same time, showed him a market where the pain might be much higher. For a tech enthusiast, touchless phone control is a nice-to-have. For someone whose hands will not cooperate, it could be a must-solve. Same product, completely different position on the pain scale we learned on Day 2 (pain × frequency × consequence).

We usually try to raise the pain score by improving the product. Sometimes the faster way is to ask who already hurts the most, and start there. And if a panellist hands you a second market on stage, the right response is neither "no, we're only doing X" nor "yes, we'll do that too". It is: "yes, we see it, and here is why we are starting where we are."

Case three: network businesses and the trust question

The third founder, a recent graduate building full-time with his co-founder, pitched a hyperlocal coordination network: helping people nearby find each other for shared intents (group buying, sharing a cab, finding a tennis partner), with a trust and commitment layer so you do not have to search, browse and monitor. His insight was that local coordination should be organised by proximity, not by use case, because trust compounds among people near you, and every extra intent you share with the same trusted network lowers the cost of the next one.

Three things he did well are worth copying in any pitch:

• He named his hardest problem himself. He said plainly that this is a network-effect, critical-mass problem, and described moving cluster by cluster, using a WhatsApp agent to talk to a few hundred people at a time. Naming your biggest risk before the panel does shows you have thought about it.
• He named his metrics. Repeat rate, and cross-utility migration (whether someone who came for one intent stays for another). Two metrics, clearly defined, tied directly to his thesis.
• He gave a crisp why-now. Three forces: smartphones make geofencing easy, digital identity makes trust between strangers possible, and AI can extract intent without long forms. A good why-now is a list of things that changed in the world, not things that changed in your company.

The panel went straight at the trust layer, because the trust layer is the product: in a gated community of five hundred households, do you expect everyone to have digital credentials, or is there another mechanism because they are neighbours? Then came two questions that every founder should expect, in this order:

• "Are you generating any revenue?"
• "Is it one-time revenue, or is it repeat?"

The follow-up is the one that matters. Not whether money came in, but whether it comes back. That was Day 2 repeated back to us almost word for word. Did they pay, and are they paying repeatedly?

The five questions every panel asks

From here on, this chapter is my reflection on what I heard, not a record of further events.

Sitting through three rounds of Q&A, I started sorting the questions into buckets. Almost every question was a version of one of five.

1. Is it real? Which components are yours? What is the research behind it? Has licensing actually started? This is observed behaviour versus estimated behaviour, applied to the founder rather than the customer.
2. Who exactly buys? Which governments first? The enthusiast or the high-pain user? A precise answer names a role, a setting and a moment, not a demographic.
3. How do you reach them? Government procurement, a cluster-by-cluster rollout, a WhatsApp agent. Distribution is the part of the market thesis founders most often leave vague, and panels know it.
4. Do they come back? One-time or repeat. Repeat rate. The strongest signal in the room, because retention is the one thing that is hard to fake.
5. Why you, and why now? Your unfair insight or lived experience, and the forces in the world that make this possible today and not five years ago.

If you can answer all five in two sentences each, you are ready for most of what a panel will throw at you. If one of the five makes you uncomfortable, that is the one they will ask.

A Q&A playbook

Reflection: what I would tell any founder preparing for five minutes of questions, drawn from watching these exchanges.

Answer first, explain second

The best answers I heard started with the answer. "Yes, we are in active conversations with hardware partners." "Demo hardware, hires, and IP work." Then, if there was time, the detail. Founders under pressure tend to do the opposite: context, background, story, and the actual answer somewhere near the end, by which time the panellist has stopped listening. Lead with the one sentence you would say if you only had one sentence.

"Not yet" is fine. Vagueness is not.

The answers that landed were not the most polished. They were the ones where the founder stated plainly what was true today, what was not yet true, and what they were doing to find out. The panel seemed comfortable with "not yet". What it probed was vagueness. "We haven't closed a licence yet; we are in conversations with two partners and expect to know by next quarter" is a strong answer. "We're exploring lots of opportunities in that space" is a weak one, even though it sounds more impressive.

Treat a challenge as a request for reasoning

When a panellist says "the math doesn't match", they are rarely trying to humiliate you. They are asking to see your reasoning. The drone founder did the right thing: he stayed calm and laid out the chain of logic, one assumption at a time. If you do not know, say so and say how you would find out. Arguing with a panellist almost never works. Showing your working almost always does.

Protect what must be protected, gracefully

"I won't name them, because those relationships are confidential" is an acceptable answer, as long as it is rare and the rest of your answer is concrete. Panels respect discretion. They do not respect a founder who hides behind "confidential" for every hard question.

Respect the clock

Finishing in exactly four minutes earned a founder a small round of applause. Running over signals that you have not rehearsed, and it eats into the Q&A, which is where the panel actually forms its view. Rehearse with a timer until you finish with fifteen seconds to spare.

Hear the gift in a question

Some questions are not attacks at all. "Have you thought about this other market?" is often the panellist telling you what excites them. Acknowledge it, explain your focus, and note who asked. That panellist may be your warmest follow-up.

Building the four minutes

Reflection: a structure built from what worked in the pitches I watched.

Four minutes is roughly 500 spoken words. That is not enough for everything you want to say, so decide in advance what you will leave out. Every pitch that landed that day had some version of this spine:

1. The problem, in a scene. The third founder did this well: ten people from the same neighbourhood making the same trip every day, unable to see each other. A scene beats a statistic.
2. The insight. The one thing you believe that others do not. "Coordinate by proximity, not by use case" is an insight. "We use AI" is not.
3. The demo. Show, do not describe. And pre-empt the doubt the demo will raise.
4. Who buys, and how you reach them. One customer, one channel, stated precisely.
5. Evidence. What you have observed, not what you expect. Qualitative is fine if it is true.
6. Why now, why you. Two or three forces in the world, and the reason your team sees it first.
7. The ask. What you need and what milestone it gets you to.

When we tried this on our own pitch, the step that exposed us was the fourth. We could describe the problem vividly and the demo did its job, but "who buys and how you reach them" kept expanding into three customer types and four channels. Cutting it to one sentence forced a decision we had been postponing, which is probably the real value of a four-minute limit: it makes you choose.

A few more things I noticed:

• Your first sentence decides whether they listen to the second. Rehearse it more than anything else.
• One idea per slide. If a panellist is reading, they are not listening.
• Your team slide is a why-you slide. Say in one line why your team sees this problem before anyone else does. Lived experience or a hard-won insight, briefly told, is more persuasive than a list of degrees.

Before, during, after: a Demo Day timeline

Reflection: a checklist, not a record of the day.

The week before

• Freeze the deck. Changes in the last 24 hours introduce errors.
• Run the claim audit (every logo, "first", "only", partner) and delete what you cannot defend.
• Write the ten questions you most fear and rehearse answers out loud.
• Record yourself pitching once. Watch it once. It will be uncomfortable, and useful.

The morning of

• Download every video. Copy the deck to a USB stick and your phone.
• Test on the actual screen and the actual clicker, if you can.
• Know who is on the panel and what each person cares about. A banker, a fund manager and a community builder will each lean in at different moments.
• Eat something. Four minutes on an empty stomach feels like forty.

In the room

• Breathe before the first sentence. Look at a friendly face.
• Finish on time.
• In Q&A, answer first, then explain. Write down who asked what.

After

• Follow up with anyone who asked a question that showed interest, while the conversation is still fresh, and refer to their exact question.
• Write down every question you were asked, especially the ones you answered badly. That list is your next deck revision.
• Update your paper trail. Demo Day itself is worth documenting.

Apply it to your startup

You do not need a Demo Day on the calendar to use any of this. Investor meetings, customer pitches, grant panels and even hiring conversations follow the same pattern: a short story that earns a longer interrogation.

• If you are a deep-tech or hardware founder, prepare the "what is yours, what is borrowed" answer before anyone asks. Honesty about sourcing builds more trust than claiming everything is in-house.
• If you are a consumer or network-effect founder, name the critical-mass problem yourself, show the metric that proves density is working, and have a precise answer on how you build trust.
• If you are a B2B or government-facing founder, expect the go-to-market question first. Know the buyer, the budget line, and the step after "they liked the demo".
• If you are a student or first-time founder, remember the panel is not expecting a finished company. It is testing whether you think clearly about what is true, what is not yet true, and how you will find out.
• Whatever you are building, ask what your version of "this is not AI generated" is: the single doubt everyone will have in the first ten seconds. Answer it inside the pitch.

Resources from the day

Demo Day had no workshop slides of its own, so these are the people, frameworks and tools that shaped the session, plus the earlier-program ideas it kept testing.

Who was in the room

• Aaliyah Barker, principal investor and program director, Draper University, on the panel after leading the cohort's Demo Day prep.
• Rohit, startup banking, IDFC First Bank, on the panel, with an opening talk on how a bank's startup programme supports early founders (outreach programmes, investor connections, joint programmes with venture firms, banking and visibility).
• Srinivas Kollipara, long-standing builder of the Indian startup ecosystem, on the panel.
• An investment director from a Hyderabad fund and a community lead from the ecosystem completed the panel, alongside the Draper University India team who hosted.

Frameworks and ideas

• The four-plus-five format: about four minutes of pitch and five of questions, which means the Q&A is where the view is formed.
• The five question buckets (my own sorting of the day's Q&A): is it real, who exactly buys, how do you reach them, do they come back, why you and why now.
• Answer, prove, admit, commit: the shape of a strong answer under pressure.
• The claim audit: every logo, partner, "first" and "only" on a slide is a promise you must be able to defend.
• Document, document: dated photos, dated design files and a build log as a simple paper trail.
• "One-time or repeat?": the revenue question that tests retention, not just a first sale.

Tools mentioned

• Canva, with the rule to download your demo video and play it locally instead of streaming it.
• Your phone's camera roll as a timestamped record of prototypes and designs.
• NDAs for any conversation that has to go deep into how your product works.
• DigiLocker-style digital identity, named by a founder as one of the forces behind his why-now.

Quoted in the room

• Jack Ma: "Today is difficult, tomorrow is much more difficult, the day after tomorrow is beautiful."

From earlier in the program

• PSMBT (Problem, Solution, Market, Business, Team) as a conviction checklist, from Day 1.
• Pain × frequency × consequence, and "did they pay, and are they paying repeatedly?", from Day 2.

What to take away

• The pitch buys you the questions. Four minutes earns the right to be tested for five. Prepare for the five.
• Panels probe the business, not the tech. Go-to-market, sourcing, use of funds, repeat revenue. The demo is the entry ticket.
• Five buckets cover most questions: is it real, who buys, how you reach them, do they come back, why you and why now.
• Answer first, explain second. Lead with the one-sentence answer.
• "Not yet" is acceptable. Vagueness is not. State what is true today and how you will learn the rest.
• Focus is praised, and a second market is a gift. Acknowledge it and explain why you started where you did.
• Protect the demo. Download it, test it on the venue screen, carry a backup.
• Every claim is checked. One claim you cannot defend damages every claim you can.
• Document everything, with dates. It costs nothing until the day it saves you.
• Repeat beats revenue. "One-time or repeat?" was the sharpest question of the session.

Try this

1. The five-bucket drill. Write the five questions (Is it real? Who exactly buys? How do you reach them? Do they come back? Why you, why now?) at the top of five cards. Under each, write your answer in no more than two sentences. Ask a friend to circle any sentence that is vague, and rewrite it.
2. The fear list. Write the ten questions you most hope nobody asks. Answer every one out loud, recorded on your phone. Replay and cut every answer to under 30 seconds, with the answer in the first sentence.
3. The claim audit. Go through your deck line by line and mark every factual claim (logos, partners, numbers, "first", "only"). For each, write the evidence you would show if challenged. Delete or soften anything without evidence.
4. The second-market test. Ask three people outside your industry who else your product might matter to. If a stronger pain shows up, write one sentence on why you are starting where you are, and keep it ready.
5. The failure rehearsal. Run your full demo with the Wi-Fi off, the video downloaded, and on a laptop you did not prepare it on. Fix whatever breaks. Then do it again the morning of the pitch.
6. The paper trail. Today, create one folder with dated photos of your prototype, your design files and a short log of what you built when. Add to it every week.