15 min read

Free as in Puppy: Google gave away Clueso’s product. Clueso raised $10 million anyway

Google now makes AI videos for free and Synthesia is worth $4 billion. Claude makes studio-level videos within a chat window. Squeezed in between is a 35-person company from Bengaluru which, by its own account, is profitable. Is Clueso about to be flattened, or has everybody misread where the money in AI video really is?

Anyone who has spent time around open-source software has heard that there are two kinds of free.

There is free as in speech, and there is free as in beer.

Old hands will tell you about a third kind, which I find far more useful in real life: free as in puppy.

You know how this goes. Somebody offers you a puppy for nothing, you take it home, and then you spend the next fifteen years paying for food, vets and a new sofa. 

The puppy was the cheap part.

Why am I telling you about puppies?

Because in April, Google began giving them away.

Every Google account now gets ten free videos a month from Veo, Google’s video model. 

Google Vids, its video-making app, picked up AI presenters, AI voiceovers in more languages, and a Chrome extension that records your screen in one click. If you are on Google Workspace’s priciest plan, you can churn out a thousand videos a month. Claude’s recent model Opus 5.5 generates studio-level videos right within your chat window – for free or close to that.

That is the free end of the market. The paid end is not a pretty picnic either.

Synthesia, which makes training videos fronted by AI presenters, raised $200 million in January at a $4 billion valuation, and reportedly turned down a $3 billion offer from Adobe. Guidde, which turns screen recordings into how-to guides, raised $50 million in February. And Atlassian shelled out about $975 million for Loom, the screen recording tool, back in 2023.

So on one side you have one of the world’s biggest companies giving the product away, and on the other you have rivals with war chests in the hundreds of millions.

Not a market any sensible person would walk into with $1.4 million, right?

Meet Clueso.

That $1.4 million is all it had raised until October, when it closed a $10 million Series A. It has 35 employees and more than 1,000 paying customers, and it says it is profitable.

One of those customers, and I promise I am not making this up, is Atlassian (see note above!).

We sent Clueso’s founders a list of questions to find out how any of this is possible. What came back was a description of a business that has rather little to do with making videos, and a great deal to do with looking after them.

It began with a video they couldn’t afford

Akash Anand, Prajwal Prakash and Neel Balar met at IIT Madras. Clueso was their second idea.

Their first was Desklamp, a note-taking app for students and researchers. It had about 40,000 users and thousands of sign-ups every month, but most users tried it once and left.

The founders knew why. Between 5% and 10% of users spent hours in the app every day. Everyone else gave up because they couldn’t work out how it functioned.

They decided to make a good explainer video. They booked a studio, filmed themselves, animated the app’s screens in After Effects and edited every cut by hand. One video took two weeks. An agency quoted them $6,000, which they did not have.

Desklamp got them into Y Combinator in early 2023, where they ran into a harder problem: students and researchers don’t pay for software. Their YC advisers pushed them to look for something else.

For five or six weeks the three tested one problem after another and spoke to hundreds of potential customers. Video kept coming up. Software companies everywhere needed demos, training videos and tutorials, and struggled to make them, as Desklamp had.

Midway through the programme, they dropped Desklamp and started Clueso. 

When did they realise their own pain was a business? “Honestly, later than we should have.”

Why Clueso thinks first drafts won’t sell

We asked the question any investor would ask. Suppose an AI company ships a model tomorrow that can operate real software and produce a product video on its own. Why would anyone still pay Clueso?

The founders said such a model “solves the first video draft,” and that large companies need more than that. “They’re paying for a library of videos that stays accurate, on-brand and up to date across the whole company.”

In puppy terms, Clueso sells the fifteen years of care.

The founders listed five things this involves.

First, the AI has to know the company: its product, help pages, customers and old videos. Clueso gets this by connecting to the tools where that information lives.

Second, brand rules. Clueso stores each company’s colours, voice, music and style in what it calls Clueprints, so that videos made by hundreds of employees look consistent.

Third, control. Teams want to change any part of a video, by hand or by instructing the AI, and get it reviewed and approved before release.

Fourth, updates. When the product changes, someone has to find every video showing the old version and fix it. The founders call this “a system, not a prompt.”

Fifth, the requirements every large company has: security, single sign-on, translation, and publishing to wherever the video needs to go.

Of the five, most outsiders would underrate control most, and the founders agree.

Large companies are particular about how their videos look and very concerned about whether they are correct. 

Clueso’s AI records the screens and does most of the editing, but a human always signs off.

Notice what is missing from that list. Hardly anything on it is about generating video.

Founders building on top of AI models would do well to take note. 

The models will make creation cheap for everyone, competitors included.

What a large customer keeps paying for is confidence in the output when hundreds of people are producing it.

The founders say their tool does a different job.

Tools like Loom are for quick recordings. You hit record, talk over your screen and send the result to a colleague or a customer. Nobody expects it to be pretty, and nobody will watch it next month.

A video meant for a large audience, which has to stay up for months, is another matter. It has to follow the brand rules, it usually needs proper editing, and it has to look as though a professional made it.

That is the gap Clueso fills. Customers feed in rough screen recordings, slides, documents and old videos, and get back polished, branded videos with animation, turning “rough, scrappy input into polished, on-brand output.”

So unlike tools that make videos that are forgotten by Friday, Clueso helps users create videos their customers will be watching for a long time.

The feature everyone raves about is a fifth of the business

Clueso’s party trick is that it can retrace the steps in a product and remake a video when the software changes. It is the feature people bring up first, and we assumed it accounted for most of the usage.

It doesn’t. Across Clueso, only 15% to 20% of videos are updates of older ones.

Updates mostly matter for training content, for customers or staff. Marketing videos and demos rarely get updated. Teams make new ones instead.

Among customer training teams, updates make up a third to 40% of all videos.

The founders added that release schedules matter less every year. Most of their customers now ship product changes monthly or faster, and the total number of videos made on Clueso keeps rising.

One still wouldn’t call the feature a sideshow. 

Customer training teams own the largest libraries, have the most videos going out of date, and are usually the first team in a company to buy Clueso.

A feature that wins the first team does more work than its share of usage suggests.

The founders are building their long-term plan on it. 

Clueso would detect a product change, flag which videos need redoing, record and rebuild them, and publish them wherever they appear.

Start with the team nobody courts

Clueso’s way into customers is through people most software salesmen ignore.

At software companies, that is the customer training or customer success team, which maintains large libraries of tutorials and online courses. At large companies outside software, it is the training team or whoever manages internal change.

The contract is signed by that team’s head: a Head of Customer Education or Chief Customer Officer in the first case, a Director of Learning and Development in the second.

Then usage spreads.

A year in, a typical large software customer has four teams on Clueso, usually arriving in this order: customer training, product marketing, sales, then staff training. Product managers and the engineers who demo to buyers often join. A few customers use it company-wide, HR included.

Clueso’s customer count tripled over the past year. The product expanded from how-to videos to marketing videos, animated explainers and videos built from slide decks, and each new format brought another team in at companies already paying.

The marketing got busier too. 

Clueso posted more on social media, worked on appearing in Google results and AI chatbot answers, started some paid advertising, and made it easier to sign up and pay without speaking to a salesperson.

It launched twice on Product Hunt, the site where new tech products get shared and voted on. Both launches were Product of the Week and one was Product of the Month.

The founders say Product Hunt doesn’t bring in paying customers by itself. It brought mentions in newsletters and blogs. The big customers came the old-fashioned way, through word of mouth and conferences.

Why does a profitable company need $10 million?

On to the money, because the round itself is unusual.

Dharana Capital led it. Dharana is a Bengaluru firm that began life in 2022 as part of Vy Capital, raised a $250 million fund in January, and plans to back just eight to ten companies from it. Urban Company, Zomato and Petpooja are among its other bets.

Dharana was already an investor from Clueso’s first round. So were Y Combinator and f7 Ventures, and they came back as well.

Between them, the existing investors put up more than 80% of the new money. Founders Future and Neotribe Ventures were the only newcomers.

To put the size in context, Synthesia raised $200 million in a single round this year, and Guidde has raised roughly $80 million all told. Clueso’s grand total is $12 million.

But hang on. If the company is profitable, why take any money at all?

The founders gave the answers you would expect. 

They want to build faster and try more ways of growing. They are hiring salespeople in the US to go after large companies. They are spending on AI models and on their own AI tooling. And a cushion in the bank never hurt anyone.

Then they gave an answer I didn’t expect, and it is probably the one worth dwelling on.

Many of Clueso’s newer AI features make each video costlier to produce. “We want to stay ahead of the curve and keep offering the best price we can without raising prices for customers,” the founders said.

Clueso charges for a mix of user accounts and usage, with usage counted in AI credits or minutes of finished video, and custom plans for the big accounts. Right now its priority is getting customers to use the new AI features. They are doing just that, and Clueso’s costs have gone up along with the usage.

Old-school SaaS never had this problem. You shipped a feature and it cost you almost nothing to run. In an AI product, a better feature can make every single video more expensive to make.

So part of this $10 million is there to let Clueso keep improving the product without passing the bill on to customers, at least for now. Looking after other people’s puppies, it turns out, comes with a food bill of its own.

There is a second thing I like about this round.

Our Portkey piece worried that India cannot fund its AI winners before a foreign buyer scoops them up. Clueso points to one way out. If you are profitable with 35 people, you don’t need much outside money, and when you do raise, the investors who have watched your numbers from day one will happily take most of the round.

Clueso’s designers ship their own code

Three days after the funding news, Akash put out a post saying Clueso was looking for the best designers in Bangalore.

A freshly funded startup going after designers and not engineers? People asked him about it.

His reply was that AI tools now let designers write the code for the screens users see and ship it themselves, and that jobs which mostly consist of turning designs into code are on their way out.

That sounds like the sort of thing founders say on social media. At Clueso, it happens to be how the place runs.

Both of its product designers write code that goes live. They carry a project from idea to design to code to release.

It started when Clueso redesigned its product. The designers already cared intensely about how each screen should feel, and once they began building the screens themselves, quality improved and the gap between design and release shrank sharply. These days the team often skips Figma and goes straight from an idea to working software.

Before every founder reading this fires their front-end team, a few caveats.

Engineers at Clueso still decide what goes live. They review the designers’ code, and they own how the system is built, how fast and reliable it is, and how secure. The company is deliberately keeping engineering small while design grows.

It also helps that Clueso makes a creative tool, and that its lead designers are seasoned video editors who know in their bones what a good editing tool should feel like.

So I’d take a narrower lesson than the hiring post implied. If the feel of your product decides whether customers stay, let the people with the best taste build the screens, and let your engineers spend their time on review and on the plumbing underneath.

Most of the team is in Bangalore, but Clueso is now setting up in the San Francisco Bay Area, beginning with salespeople for large customers and marketing staff. The product gets built in India and sold next door to the American companies that sign the big contracts.

Don’t fight the models. Get hired by them

None of this means Clueso is safe.

Google already bundles screen recording, voiceovers and AI presenters into Workspace. Synthesia is pouring its new money into interactive AI tools for training. Guidde says its guides can serve AI assistants as well as humans.

So what is Clueso’s answer to the big AI companies?

Join them.

It has built a connector, using a standard called MCP, that lets assistants such as Claude and ChatGPT create and edit videos inside Clueso. “When someone asks Claude or ChatGPT for a product video, we want Clueso to be the video layer their agent calls,” the founders said. “Every time the models get better, our agents get better too.”

Its newest product, Clueso Agents, runs on the same thinking. Customers no longer have to record their own screens. Clueso’s AI moves through the product and records it, with the company’s Clueprint steering the edit.

A team writes a script or a set of instructions, Clueso makes the video, and anyone can jump in and change it. When something goes stale, Clueso can spot it and produce fresh versions of both the video and the written guide.

Other formats are coming, though the founders are in no hurry. Clueso already produces some written content, such as step-by-step guides and help articles, and some customers export videos as presentations with one scene per slide. Interactive product demos are the likely next stop. But they want to be the best at video first.

Ask them where this ends up and they describe Clueso as the tool that keeps a company’s product information current, whatever the format.

Google can keep the puppy

Let’s go back to where the founders began. One product video, $6,000, two weeks.

AI has pushed the cost of that first version close to zero, and Google is handing a good chunk of it out for free. If Clueso were in the business of first versions, this article would be an obituary.

It is very much alive because it sells what comes afterwards: getting the video right, on-brand and approved, publishing it, and fixing it every time the product changes. It sells that to one unglamorous team, and then to the next team down the corridor. And it has raised money partly because doing this well keeps getting more expensive.

If you are an Indian founder building on AI models, I’d assume the worst. Assume the model companies will soon be able to create whatever your product creates, and ask yourself what a large customer will still need from you on the day that happens.

Google can keep giving away puppies.

Clueso has just bet $10 million that somebody will pay for the next fifteen years.

Stay close to the frontier.

New programs, founder stories, and signals from India's AI layer — in your inbox when there's something worth reading.

No noise. Unsubscribe any time.