Rony Seamons, chief operating officer at AMPLYFI, discusses his experience of working on a Clwstwr project.

AMPLYFI is a Cardiff-based company that is revolutionising business intelligence. Its technologists, academics and sector experts develop ways of gathering data through AI, deep search and other advanced technologies, enabling their clients to gain objective insights that can power them forward.

We found out about the Clwstwr funding as part of a happy coincidence

In April 2019, we had a guest speaking spot at a BBC Digital Cities event in Cardiff. Our presenters spoke about how our AI technology is used in businesses, something that really struck a chord with attendees from the creative industries - particularly a few journalists who were there. Some of these journalists thought what we were doing could have huge applications within journalism, and told us that the Clwstwr funding call was coming up. 

We applied for funding because we wanted our skills to help another industry

AMPLYFI focuses on finance, particularly ‘Know Your Customer’ (KYC), ‘Know Your Business’ (KYB), anti-money laundering and connections within the finance sector. We thought there was an opportunity to cross into a very different sector, a chance to penetrate an entirely new market with partners in the journalism world. 

As tech experts, we're always looking to partner with people who understand an industry because that means we can create products that meet particular needs. It’d be easy for us to busy away with our coders creating crazy tech that nobody wants to buy, but the thought of partnering with journalists at the coalface of the industry was a huge driver for us.

We received funding from Clwstwr to do early stage R&D

As soon as we got our funding, we wanted to start getting to know how the journalism industry works. We pulled together an editorial board of key influencers in the local media ecosystem, which included people from JOMEC (Cardiff University’s School of Journalism, Media and Cultural Studies) and people with links to the media. It would be their task to challenge us throughout the project and help us meet journalists, editors and others working in the industry. We also built a user community of around 40 active journalists from across the industry, most of whom were from Cardiff. 

Having active journalists in our user group gave us a great insight

We interacted with our user group through workshops and one-to-one conversations. Our main aims were to get a picture of a day in the life of a journalist and to validate our plans. We uncovered what it's like to be a journalist, what problems they face, which mundane daily tasks take up valuable time, what tools they use and what pain points there are. This was with the view of seeing if we could create a tool that would take some of those time-heavy elements off the journalists so that they could focus on writing.

Our post-workshop analysis helped refine our plans

We wanted to conceive a product that can be commercialised and taken to market. So, we came up with four different user journeys and scenarios for product development and used our data to choose the most viable option for the next phase. We scored each journey on things like how commercial the idea is, how realistic the development roadmap is and how closely aligned it is to what we're already doing. That's how we came up with the one we are going to go for.

Our AI can help speed up the time it takes journalists
to find reliable and relevant facts and sources. 

Our technology could do what takes investigative journalists hours in just one search 

Professor Justin Lewis (the Clwstwr director) described investigative journalism as trying to get to the geneology of a story – you have to peel back layers to get to the source, quickly finding connections between people involved in the story. It’s an arduous manual task with lots of rabbit holes to fall down. An AI system could rinse huge documents, such as those produced in the Panama Papers scandal, through a specialised programme, and, within minutes, come up with all the key names, key players and their connections.

We could build the AI technology so that it functions like a search engine

Whether the breaking news is a person, a topic, a theme or a country, you’d simply type in a search query and watch as it retrieves all of the relevant sources to build a diagram of connections around that topic. It’d be done at lightning speed, trawling court records, land registry and all manner of resources that take ages to go through manually.

One of our USPs is we get to the things that Google doesn't

When you conduct a search on Google, it returns millions and millions of results in less than a second. But look a little closer and scroll to the end of the search pages and it actually only returns anywhere between 200 and 350 articles on average. The big numbers displayed at the top is an estimation of what's available, but it only indexes and displays a small fraction of what's out there. Our technology delves into the deep web and returns many more results.

Also, much of Google's search results are driven by paid adverts and SEO, so what you see on page one might not be the most relevant. Journalists, like most of us that use traditional search engines, therefore need to navigate past these challenges in order to find the best intelligence.


We understand the sector a lot more after this R&D

We have a long way to go, but we can now shape the product into a really useful tool that there is a need and demand for among journalists. We've done all the market research and validation that we need to do, so we’d like to get funding for stage two. We’d use it to create something that’s as close to a market-ready product as is possible through a full engineering build phase. Once the product is finished, we can see it being sold on a subscription basis to big media entities like BBC and CNN.

This round of R&D has felt better than a lot of the R&D we have done with funders

Some funding partners can be very hands off and unwilling to respond to questions, so you're left to your own devices. With Clwstwr, we felt like the team wanted to help every project to succeed. Clwstwr were so willing to give feedback, contribute and help shape things. We’re excited to see what happens next.

Find out more about AMPLYFI's Clwstwr project AI in the Newsroom.