starekase - Fotolia
Fintech Fluidly combines Google’s cloud and AI tech to ease SME cashflow management woes
London-based fintech Fluidly explains how its recent all-in move to the Google Cloud is supporting its continued growth, and delivers new features to customers faster
Helping small and medium-sized enterprise (SME) owners get a good night’s sleep is how Caroline Plumb, OBE describes the overarching aim of the cashflow management-focused fintech, Fluidly, that she founded in 2016.
“We know that more than two-thirds of business owners say they can’t sleep at night because they’re so worried about their money and their cashflow,” Plumb tells Computer Weekly. “And so what we do is help businesses see into that financial future, and provide them with control, certainty and confidence around it.”
This is achieved by offering SME business owners (and their accountants) access to an artificial intelligence (AI)-powered cashflow forecasting platform that draws on past transaction data to help small businesses make informed, real-time decisions about how best to spend their money.
So, for example, how will hiring a new member of staff, paying an upcoming bill or investing in new office equipment affect the company’s cashflow in the future?
Since Fluidly’s inception, the number of SMEs that rely on its cloud-based software has grown considerably, prompting the company to undertake a revamp of its hosting arrangements earlier this year, resulting in a wholesale migration of its services to the Google Cloud Platform (GCP).
“We started on a provisional platform with a couple of hundred SMEs, and then [we had] a few thousand SMEs connected,” says Plumb. “And, today, we have 20,000 to 25,000. The old platform would have fallen over, so the migration was really about achieving scalability and robustness.”
Time to make a move
The migration began in January 2019, and was completed six weeks ago, coinciding with a rebuild of the entire Fluidly platform, says the firm’s CTO, Mike Hancock.
“As we rebuilt things, we migrated them over and onto Google,” he says. “It’s a high- risk move, but there’s been a lot of planning going into it, and there hasn’t been any downtime. There have been some bugs, but nothing major. It’s all been remarkably smooth.”
Fluidly’s offering was previously underpinned by Heroku’s platform-as-a-service, but the AI component of its services has always been based on Google technologies, says Hancock.
“When we started building the AI platform, we evaluated other public clouds, and Google came out top in terms of its support, and it still is at the cutting edge for machine learning,” he adds.
“The website was originally hosted in Heroku, which is great for just doing simple things, but there was never any support for AI.”
As mentioned above, the organisation had scalability reasons for wanting to switch up its infrastructure arrangements, but the decision to expand its use of Google cloud also has a performance improvement consideration to it, says Hancock.
“We’re obviously moving large amounts of data around and having it all in one datacentre with Google’s fast network connections makes a real difference,” he says.
For example, the company says the move means it can now generate cashflow forecasts for its customers far more quickly than before, but it is also in a position to roll out changes and add new functionality to its platform at a faster pace.
To achieve this, the firm is making use of the Google Cloud Platform’s AI platform, which is a data science development environment that is supporting the creation of machine learning-based features for the Fluidly platform.
Adopting agile ways of working
At the same time as the migration and rebuild of the Fluidly platform was going on, the company also undertook a revamp of its software development processes by embracing agile working and the concept of continuous deployment and delivery.
“The reason for doing it was to deliver value to users much faster,” he says. “So we have moved to a full, continuous deployment model where we are shipping [new code] to production on average between 30 to 50 times a day.
“It means we can develop and ship features on the same day, and see visible progress in the app on a day-to-day basis.”
The technology changes, coupled with Fluidly’s newly adopted software development approach, has created a demonstrable difference to how the company works but also the atmosphere too, says CEO Plumb.
“These technologies have changed the culture, behaviour and cadence overall of the company,” she adds. “The pace feels different, and you can feel the momentum.”
Read more about cloud migrations within financial services
- The financial services community has gone from being one of the least likely sectors to adopt cloud to becoming one of its keenest users, as regulator attitudes to using the technology have become more accommodating.
- The extent to which financial services institutions across the UK and US have embraced cloud has been laid bare in a piece of research from Bristol-based software house YellowDog.
On that point, the pace at which Fluidly’s 12-strong software engineering team now operates is markedly quicker.
“Engineers typically now raise three pull requests a day, and get them reviewed and merged within the same day,” says Hancock. “In the past, you would work on something for two days, raise a big pull request, and it wouldn’t get reviewed for another three days, at which point they’ve moved on to the next thing.
“Now it’s just this constant conveyor belt of small changes that people are constantly reviewing and pushing, and pushing out changes.”
With its new infrastructure setup in place, its technology teams embracing DevOps-like software development techniques and new functionality being added to the Fluidly platform all the time, Plumb and her team are turning their attention to ensuring even more small business owners can get a good night’s sleep.
“As I said, our purpose as a business is to help all those sleepless business owners worrying about their money,” she says. “We’re hoping to bring this to millions of businesses, and we’re in the tens of thousands today, so there’s a long way to go.”
