And why we did not shout about it from the rooftops…
Sumanth Raghavendra, CEO and co-founder of Presentations.AI
You know what’s cool?
One million users.
If you are a productivity startup targeting prosumers, hitting the “one million user” milestone represents a rite of passage - a strong signal that you are on your way to achieving product-market-fit.
The pace at which you achieve this milestone is the second dimension of note. The quicker you get there, the stronger the pulse of your business.
While history has seen the emergence of several fast-growing technology prosumer companies - including household names such as Slack, Zoom, Dropbox and Notion - the mantle of being the “fastest growing productivity app” of all-time belonged to Tome.app - an AI story-telling platform that hit this one-million-user milestone in six months from launch.
But this was in February 2023.
Since then, a small and largely unknown startup has done one better.
That startup was ours.
Presentations.AI is an AI-powered presentation platform that launched on July 7, 2023, and hit the one million user milestone on September 23, 2023. That's 1 to 1 million users in just 84 days - less than three months, nearly twice as fast as the previous record holder.
To put this into perspective:
Our growth journey wasn't just fast; it was hypersonic.
But speed isn't everything, and the real story lies in how we got there and what came next.
Before we get into the details, it might be pertinent to note these facets:
I would have loved to claim that these facts prove that we are marketing savants but that would be disingenuous.
The truth of the matter is that we had to be frugal not out of choice but out of compulsion - we simply didn’t have the resources. We had raised a small funding round many years back but had all but run out of money by the time we were ready to launch our product.
Given this rather grim situation, how did we become the fastest growing productivity app of all time?
The answer can be distilled to a 3-step process - a process that can perhaps best described as “simple but not easy”.
Let's expand on these points.
Paul Graham of YCombinator famously said that the most important thing that your startup can do is “make something people want”. In fact, this remains the YC motto to this day.
It is easy enough to recognize and intuit that this is necessary for startup success.
It is far more difficult to unweave this aphoristic truism to battle the nitty-gritty nuances of the real world.
Take our case as an example.
We are an AI-powered presentation platform.
In a world drowning in presentation tools, did users really need another one? While there are billions of users of Microsoft PowerPoint and Google Slides, how many of them do you believe “want” a new presentation tool?
The startup graveyard is filled with "PowerPoint killers" that ended up more like PowerPoint paper cuts - mildly irritating but ultimately forgettable.
We realized the real unsolved problem wasn't about killing PowerPoint, but about euthanizing the blank slide. People struggle to start from scratch each time they need to create a presentation. Tools like PowerPoint conflate content and design - this makes it difficult for non-designers to conceptualize and convert their thoughts and stories into a visual summary. The irony of PowerPoint is that it is the world’s most widely-used design tool…but its users are primarily non-designers.
Sometimes, being early is the same as being wrong.
Long before AI was cool, we had the idea of using it to help people with two aspects of presentation creation:
We built this AI platform for years in a pre-GPT world. It was like trying to build a spaceship with stone tools and a vague understanding of gravity. This was an arduous challenge as there was no ecosystem to speak of - from ML models to data pipelines, we had to build everything from scratch.
Recognize and capitalize on market shifts.
Then the generative AI tsunami hit.
Suddenly, our headwind became a tailwind.
It was like we'd been pushing a boulder uphill for years, and someone suddenly installed rocket boosters on it.
Not only did the missing pieces in the technology puzzle fall in place, there was another enormous tailwind in terms of consumer interest in a solution like ours which pushed our CAC down to zero.
By the time we launched our public beta in July 2023, there were literally thousands of people knocking on our doors.
Our domain name - www.presentations.ai - transformed overnight from something that people didn’t understand or care about to something that thousands of people “wanted”. From painstakingly having to explain our approach and value proposition, our domain name became a self-sufficient way for potential customers to understand what we offered.
The snapshot below illustrates how much interest there was at the top of the funnel for our solution compared to our peers, almost all of whom were far more heavily funded and higher profile than us:
Picture: Organic Search Leaderboard for our market in early 2023
We witnessed first-hand how market conditions can change rapidly, turning what was your crazy idea for years into the next big thing almost overnight.
Chalk this up to amazing prescience or just plain luck but we ended up making something people want…but that in itself is a condition that is necessary but not sufficient.
What next?
Solving a problem is one thing. Creating an experience people genuinely look forward to is another challenge entirely.
After all, how many people wake up excited to work on their presentations?
“Death by PowerPoint” is a sentiment shared by many. There are chronicles dedicated to “PowerPoint Hell” - folks preferring to endure the pain of dental surgery over the agony of having to create a presentation.
We were determined to change that narrative.
How do we transform the business-critical act of creating presentations from pain to pleasure? Convert a mere tool to a beloved part of our users' jobs and workflows.
That's the question we asked ourselves before writing a single line of code.
Our journey began not in a boardroom or a garage, but in countless coffee shops, office spaces, and conference rooms where people struggled with their presentations.
We observed. We listened. We empathized. And what we found was eye-opening:
These weren't just minor inconveniences; they were productivity sinkholes costing businesses billions in lost time and missed opportunities.
So we set out to solve these problems, not by creating a marginally better slideshow tool, but by reimagining the entire presentation creation process.
We asked ourselves:
These questions led us to harness the power of AI, not as a buzzword to slap on our marketing materials, but as the very foundation of our product. Our AI doesn't just assist; it collaborates. It doesn't just format; it thinks alongside you, anticipating needs and making improvements in real-time.
The result?
A tool that doesn't just help you make presentations—it helps you think better, communicate clearer, and present with more confidence than ever before.
From “Death by PowerPoint” to “Love you Forever”.
Creating a beloved product is only half the battle.
The real magic happens when you empower people to achieve things they never thought possible.
We obsessed over every pixel, every interaction, every moment in the user journey. Our mantra was simple: "If it requires a manual, it's too complicated."
The result? Users often tell us they've created the best presentation of their lives within minutes of signing up.
This approach also made our users feel like they were part of something bigger — a movement to revolutionize the way the world communicates ideas.
And they were the pioneers - the first one million converts.
Turns out that making people “want something” is a far more powerful flywheel than merely “making something people want”.
Now in an ideal world, this is where this story should have ended.
Fastest growing productivity app of all time…beloved tool…loyal users…startup success…ride off into the sunset…live happily ever after.
Unfortunately, not quite.
The villain in this tale is what we came to recognize as “AI’s three-body problem”.
This is how it works.
Every AI startup faces three challenges:
While the first two challenges are applicable to all startups, the third one specifically applies to AI startups.
Allow me to illustrate with our specifics.
If you notice the image above, it shows how people were clamoring for a tool like ours in January 2023.
But we didn’t launch our public beta until July 2023.
Why so?
In January 2023, the AI compute costs for generating one presentation was $2. This is just for the LLM to generate text. If you add AI image generation services such as Dall-E, the cost could easily go up to as much as $10 per deck.
So if you wanted to acquire one million users and each of those users created just one deck, the AI costs alone could add up to $10 million.
Which was, of course, exactly $10 million more than we had at that point in time.
So even though we were ready in all other aspects, we simply couldn’t afford to even launch our beta.
Unsurprisingly, it allowed competitors to capitalize on this AI presentation opportunity. For instance, Tome.app - which emerged as the previous record holder for being the fastest growing productivity app. While Tome had originally launched its beta earlier in 2022, it had precisely zero references to AI. However, it had two powerful assets - $37m in funding and the backing of the likes of Reid Hoffman (previously the founder of LinkedIn and subsequently a board member of OpenAI and an investor at Greylock). Tome leveraged these resources to overlay AI on its original product and turbocharge user growth. It then successfully parlayed this buzz and traction to add a further $43 million to its funding kitty.
While it might have been understandable to try and take potshots at a heavily-funded competitor, we appreciate Tome’s skills at fundraising and doff our hats to them - like many other startup aspects, fundraising is a key skill and a craft that a few founders can successfully pull off far better than the rest of us and this has to be appreciated for what it’s worth.
We did try to raise funding at this point in time but it was far tougher for us - partly because we are based in India where it is tougher to find early VC backers and partly because we had already worked on this for years without having much to show for it.
So how does a small startup with no resources compete in an AI market like this?
To start with, we simply waited.
LLMs are the fastest depreciating asset in history - from Jan 2023 to July 2023, the costs of these services fell substantially. Of course, the costs were still non-trivial especially when you had to provision for larger context windows and newer, more powerful models.
But by July 2023, we had figured out how to launch our public beta and grow…without going bankrupt in the process.
To make our unit economics work, we:
Through this combination, we managed to bring down the cost of creating a presentation to just 2 cents per deck.
While a unit cost of 2 cents is obviously far better than $10, it is still substantial when millions of presentations are created each month.
To solve this, we adopted two strategies that were at least somewhat non-obvious.
The easiest way to solve for expenses is to monetize, or at least try to monetize your solution.
The accepted playbook for a prosumer SaaS solution like ours is to have a freemium model - featuring a generous free tier where users can experience the product upfront and then convert a sliver of your user base to a premium paid version over time.
However, given the high AI costs involved, our success depended on finding a balance between demonstrating our value to users and convincing them to upgrade as quickly as possible.
In our market though, this is easier said than done. Incumbent tools such as Microsoft PowerPoint and Google Slides are largely perceived as being free by end users. So there is a high bar to cross if you want customers to take out their credit cards.
But a valuable lesson that we learnt very quickly in our journey was that it is pointless to try to charge users a flat fee of say $10 to $20 per month for your tool. For one thing, it undermines the value of your solution and for another, it is almost impossible to build a big meaningful business by charging a trivial sum like this.
Today, there are enough experts who have opined that the right monetization strategy for AI startups is to charge for outcomes rather than tools. We figured this out early on our own in an interesting way. While our solution was very powerful on a number of fronts beyond the AI aspects - features such as collaboration and graphic-heavy templates with rich media - the outcome that our customers wanted and valued most was far more prosaic - a way to export the decks to PowerPoint!
We started with charging $25 for exporting one deck to PowerPoint. We saw substantial uptake. We then increased the price to $49 per export. There was no discernable drop in order count. We doubled it again to $99 per PowerPoint export and yet again, there was no drop in conversion.
We realized that people might be reluctant to pay even $10 for a tool that competes with PowerPoint but they are more than happy to pay $99 for a single well-designed PowerPoint that saves them hours of effort and pain. It seems obvious in retrospect but it was an epiphany for us at that time.
We then introduced a Pro plan that allowed users to create and export as much as 100 PowerPoint files in a year - we priced it at $198 per year - just 2X the price of a single PowerPoint export. This glide path and staging of value resonated with customers and we saw tens of thousands of purchases.
Another interesting learning was that while we had powerful sharing capabilities within our app, it was not used anywhere as often as we had originally estimated. Turns out that our users didn’t want their peers to know about our existence - it was their “secret weapon” in a manner of speaking and paying $99 to export it to a normal PowerPoint with zero traces of our AI capabilities was a small price to pay to continue enjoying this advantage. Whoever thought that “anti-virality” could be a viable business model!
Traditional startup wisdom says that the market rewards growth - startups that grow fast are recognized and rewarded by large fundings and higher valuations. However, what goes unsaid is that this rule is applicable only to the top 1-2% of startups - the ones founded by “pedigree” VC-backable founders with the right credentials.
The rest of us have to play a different game.
I still remember one VC encounter that was particularly poignant. In the midst of our dizzying growth, I pitched to a VC, a Harvard MBA no less. We were on the cusp of becoming the fastest growing productivity app of all time yet we were told that it is still too early and we needed to “show more traction” plus the mandatory “please let me know how I can be helpful” throwaway!
Given that the VC-funded growth model was all but ruled out, we had to make some tough choices to make the business model arithmetic work.
One of the big decisions we took was to clamp down on growth. This might seem incongruous, all the more so in a post that talks about fast growth. But there was some method behind this madness.
One of our key learnings as we grew our user base was that the AI hype is both a boon and a bane. While it cuts down your customer acquisition costs, it also brings in a raft of “AI tourists” who have no intention of ever buying your product. Given that there is a non-trivial cost to service these free users, they represent a blackhole in your balance sheet.
So we implemented a set of measures to weed out such casual users. We developed a rubric that routed low-intent users to cheaper models. We put limits on certain geographies and personas. We even removed all social and incentivisation features that rewarded users for inviting their friends to our platform.
So while our competitors like Tome.app and Gamma.app grew their respective user bases from 1 million to 20-30 million over the course of a year, we deliberately grew more slowly and thoughtfully.
But we are much happier for it.
Why so?
1 million users is cool.
You know what’s cooler?
$1 million in revenue.
That was our monetization target. It took us only a few months to get to one million users but it took us one year to get to the $1m ARR mark. Getting to this figure was not a vanity pursuit - it was the point where we broke even and turned profitable on a cash flow basis.
Since getting to this figure in June 2024, we have grown by leaps and bounds in the period thereafter. While we were aware of the power of compounding, it is quite surreal to see it first-hand.
But that is a story for another day.
While one million is a good milestone, it is just the beginning of our journey. As we look ahead, we're more excited than ever about the potential of AI in productivity tools. We're working on new features that will make creating presentations even more intuitive and powerful, including:
We're just getting started on our mission to transform how the world creates and delivers presentations. The rapid growth we've experienced is a testament to the need for innovation in this space, and we're committed to continuing to push the boundaries of what's possible.
None of this would have been possible without our amazing users, dedicated team and patient investors. To everyone who has been part of this journey - whether you've been with us from day one or just joined yesterday - thank you. Your trust, feedback, and enthusiasm drive us forward.
As we look to the future, we promise to share more tales and learnings and invite you to continue this journey with us. Together, we're not just changing how presentations are made; we're redefining what's possible in productivity software where AI empowers everyone to communicate their ideas with impact and ease.
It's time to scale!
