Using A.I. to massively scale product research
Arro shoots for a new era of better-informed product development
PreSeed Now brings you an in-depth profile of a different B2B or deep tech startup every Tuesday and Thursday. Subscribe to get it straight to your inbox.
It’s still unproven how many truly successful tech companies can be built on top of the generative A.I. revolution.
Will it be like cloud computing, which powered the SaaS boom? Or will it be more like social media, where only the platform operators truly win?
That’s yet to be seen, but in the meantime we have a lot of fascinating startups proving how many interesting and useful things can be built with the tech.
Today we meet Arro, a startup looking to massively scale the number of customer and user insights product teams can collect.
This is the first time anyone’s written about Arro, but it’s the second time I’ve been first to write about a startup from co-founder Craig Watson.
I wrote about Soundwave, the social music app he sold to Spotify, after I picked up on Dublin VC buzz a decade ago.
I’m currently in Lisbon for the Web Summit. If you’re around, say hello.
Arro is using A.I. to massively scale product research
User research is one of the most important parts of product development.
But it’s one that can involve hours of painstaking work, talking to people and learning about how they use a product and how it could serve them better.
Craig Watson experienced this first-hand when he worked as a product lead at Spotify. Then he saw the challenges product managers elsewhere faced while working as a product discovery consultant.
Time and money, he says, were the main constraints on the number of customers you could engage with.
“An interesting thing in the back of my mind was ‘how could you 10x or 100x the amount of customer interactions you have?’. Because ultimately, it's going to lead to better products,” says Watson.
“The closer you are to the customer, the more you understand their pain points, more likely you're going to actually solve something with technology.”
Then along came the generative A.I. revolution and Watson identified a solution. And thus, Arro was born as a way of automating product user research.
How it works
With Arro, A.I. conducts conversations with users, allowing many more conversations to be had. No longer is research constrained by the number of people you can incentivise to speak to a human at a mutually convenient time.
“You can conduct research while you're sleeping,” as Watson says.
“Instead of having to set up five or 10 customer interviews, which would be standard enough, you could set it as 1,000, or 500, or 10,000 if you want to, and then that works in the background autonomously.”
Arro lets product teams set the questions they want to ask, the context for why the research is being conducted, and the tone of voice the A.I. should have.
When the research is complete, Arro compiles summaries of the conversations, along with aggregated insights to highlight key themes and discoveries.
The video below is a demo Watson has put together to paint a clearer picture of how it works:
The story so far
Arro isn’t Watson’s first startup. He previously co-founded Soundwave, a social music app that was acquired by Spotify. He also co-founded and sold a business selling limited edition prints of classic tech product patents.
Previously based in Ireland, but now in London, Watson has co-founded Arro with Stockholm-based Johannes Herrmann.
The pair had been working on an asynchronous communication platform to help product teams conduct more research. Then, inspired by the arrival of ChatGPT late last year, Watson says they saw an opportunity to scale research conversations in a much more effective way.
They spent most of this year building the product. Watson says it was developed with the assistance of an ‘expert advisory board’ of around 50 people from product teams at well-known tech companies, who acted as “design partners”.
Arro has now launched in private beta.
“Our goal over the next three months is to get the product out into people's hands; the right people,” says Watson.
“What we found so far is it tends to be B2B companies who are scaleups, who maybe don't have dedicated researchers, or who are maybe under-resourced when it comes to being able to conduct product research.
“They’re keen to do more discovery and learn from the customers, and are more willing to work with a startup like ourselves, and try something relatively new to collect those insights at scale.”
As Arro learns from its beta testers, the startup–now a team of four–is working on adding support for images and video to allow for richer experiences.
“We’re building these Lego blocks so that as a product team, you can have standard conversations, you can start bringing in prototypes, you can start bringing in websites, you can start doing screenshots etc, getting feedback from customers that way,” says Watson.
He says there’s also potential for features like displaying lessons learned from other customers, aggregated to show how users tend to think about certain topics.
And eventually, voice capture and voice synthesis could help bring more nuance to conversations, allowing the A.I. and the user to have a more ‘natural’ conversation.
The A.I. platform risk
As OpenAI continues to add features to its products that replicate the work of startups that build on its APIs, is Watson worried about being ‘Sherlocked’ by A.I. model providers?
He says that while some of the tech underpinning Arro leans on foundational models, the startup is building its own proprietary tech wherever it makes sense to do so.
“We’re conscious of how we can strategically build up our own proprietary technology and make sure that if something does change underneath this, we're able to swap it out.
“In terms of defensibility, really it comes down to ‘if your product was turned off overnight would the customer care or not?’ And the way we want to position Arro is that it becomes a research repository…
“Without Arro, you lose access to all the central repository of insights… being able to quick-query that in natural language. It’s the source of record for the company.”
Go deeper on Arro
Discover more about their funding, vision, competition, and challenges:
Keep reading with a 7-day free trial
Subscribe to PreSeed Now to keep reading this post and get 7 days of free access to the full post archives.