Eli5 build custom software. The beauty of this is we can create almost any type of software — we have our own in-house Low-Code platform full of building blocks for IBM AI, machine learning algorithms, Blockchain technology, secure API’s, IoT connections, and integration technology.
This past year, we have been working on a big Regulatory Technology Platform (RegTech). So we thought we would do a run-down of the key 8 learnings:
This is likely to spike in 2021 due to Covid-19 and an increasing push in the EU to be more sustainable and environmentally friendly.
The figures speak for themselves. For banks, there are more than 800 individual regulators worldwide. The number of regulatory changes they have to deal with on a daily basis has increased from an average of 10 in 2004 to 250 in 2020. It is a similar story within taxation and other heavily regulated industries. Every year companies spend more than $100bn on their compliance functions. HSBC alone employs more than 6,000 full-time staff in compliance, at a cost of more than $3bn a year.
Covid-19 is behind most of the recent budget and staff cuts. However, this is nothing new. Most markets have been trying to do more with less in some way or other since the economic downturn in 2008.
The main thing that has been driving change and improving efficiency is technology. Think of collaborative work environments like Google Docs and Notion, or team communication tools like Slack and Hangouts. These make a massive difference to our own workflows at Eli5. And on the custom side, the bespoke RegTech solution we have been building has made it possible for a compliance team to complete their normal tasks in 25% of the time.
It is now possible to digitise data with ease — from PDF’s and scanned documents to video files. Once you have consistent data, most repetitive, manual tasks can be improved or fully automated. In particular, Natural Language Processing can help to sift through, prioritise, and even respond to queries and tasks at a vast scale — and on top of this machine learning algorithms can work to improve systems and processes over time. With regard to regulations, that means the software can read documents, summarise them, automate responses and even recognise where to link to other source materials.
Interfacing is another key issue. This means how the user interacts with the software, what they see on the screen and their journey through tasks. Making this simple and user friendly is the most important part, but critically there needs to be version control for tracking tasks and auditing, as well as shared spaces for remote collaboration in real-time. I have worked for many companies where either they can’t find off-the-shelf software that suits their needs, or even when they do, that their teams do not adopt the new way of working. This is why many more companies, especially larger ones, are seeking custom solutions that are moulded to suit their methods of working.
A stakeholder can greenlight a project, then an independent consultant can help you to unearth inefficiencies, and finally, a software company can build a solution. But it is essential to work with real end-users throughout. We often say what makes our development company successful is that we don’t always build exactly what a senior decision-maker wants. We only build what the users and their feedback are telling us needs to be done.
The good news is that, when it comes to software, we have frameworks to make this process much faster. The original Agile Manifesto was devised by software developers. Our Sprints mean designing a working prototype takes under 3 months and for each new feature after that we work in 2-week Sprints.
When it comes to Process Automation, our integrated solutions and collaborative workspaces that connect data, processes, and people. That means a lot less time on unnecessary tasks like emails and meetings, and more time in a space where every team member can oversee all tasks and activities.
We get a lot of people asking if No-Code and Low-Code platforms are going to make our job as custom developers obsolete. Many people we speak to expect us to hate them, when in fact the opposite is true. They are great tools for making simple products or prototypes for bigger ones — especially when you go one step further than the mainstream products and develop the foundational building blocks of code yourself.
No-Code platforms are things like Squarespace or Wix, that allow you to drag-and-drop elements into a workspace and build simple things like a website. Low-Code is a platform that allows developers to reuse common snippets of code. So for us, having our own repositories of code ready to go helps us save loads of time when prototyping for clients — which means they get a quality product faster and cheaper (although I should note here you can also build for full Intellectual Property with custom — not possible with No/Low-Code).
The RegTech we have built is an online platform. This means that an entire compliance team could transition to remote working with absolute ease during the Covid-19 pandemic. There was nothing new to learn and they didn’t need to download or install anything. There is no way you should be building any new tool without considering remote working options.
What is different about this project is the complexity of the product (e.g. Natural Language Processing, AI, Cloud and Blockchain technologies) and the depth of involvement we have with the end-users. The traditional software development trajectory involves a brief from a client, then a team would go and build a product. You get a delivery in 1–2 years, maybe more, and then the developers are only involved again when issues and bugs are found — or the entire system needs another upgrade.
This process is truly game-changing as we are in constant feedback and improvement loops. During the average 2 years it takes for a big consultancy to develop something like this, much can change. By using these short development loops where we consistently build, test, and iterate side-by-side with the end-users, you get a much better product much faster, with almost zero technical debt.
The economic downturn and the rapid speed of technological advancements already meant almost every market was competing to be the most efficient, innovative and provide the best for their employees, clients and customers. We have only seen this intensify with the Covid-19 pandemic. So if you want to understand how digital products can improve your processes, or you ever have a general techy question you’d like to ask, just shoot me a message :)
There is nothing wrong with solely building products for clients. In fact, it's one of the best ways to gain a lot of experience and attain specific knowledge that can be leveraged to build your own products. However, the agency model is not a great business model and unless you truly enjoy every part of it, it's wise to avoid scaling an agency. There are benefits to remaining small by design. At a smaller scale, it's easier to make a profit, you are free to pick projects that bring specific knowledge and new opportunities, and handling your operations can still be done without it requiring too much of your attention.