Categories

How to Create the Treasury Tech You Always Wanted

Low-code/no-code (LCNC) programming offers non-IT professionals a chance to self-create the tools they want and need. TMI talks to Nicholas Franck, a long-haul corporate treasury and banking professional and LCNC convert, about the technology that could be a treasury game-changer.

No, that headline is not clickbait. Those treasury functions that have discovered LCNC software already know it is entirely possible to create and maintain the applications they need, when they need them.

Admittedly, LCNC users in treasury are currently few and far between. But every movement has to start somewhere. And if that ‘somewhere’ requires little or no prior programming knowledge – just the enthusiasm to build what vendors or IT departments won’t – then surely that is an attractive proposition that will gain momentum quickly.

An expanding market

Indeed, LCNC technology will account for more than two-thirds of application development by 2024, Gartner predicts. A study by market data provider Statista has also found that LCNC development tool spending will grow from just under $13bn in 2020 to $65bn by 2027. So, with the technology barely a decade old and the market growing apace, it is a good time for treasurers to get on board, start building the tools of their choice, and join the band of so-called ‘citizen developers’.

Nicholas Franck
Treasury Executive & Low-code/No-code Evangelist

One of the key attractions of the LCNC approach to software building is that it is relatively easy and can be self-taught, often using free online tutorials. What’s more, it can be explored using free platforms, at least for the basics. Developing more complex and scalable applications is catered for by professional or enterprise platform variants, which do have a cost, although, for corporate treasuries, even this is likely to be low compared with internal IT billings.

Another persuasive argument for adopting the LCNC approach is that treasury employees creating and maintaining applications according to their own needs and timelines will not typically need to wait on (or pay for) often scarce, overworked and non-treasury expert IT resources.

IT is often against other departments building applications, at least initially, as resources are stretched, and tech team members fear they will have to clear up any messes created. In this case, though, using LCNC helps them deliver more efficiently. Treasurers rarely have the time or systems knowledge to draw up precise system specifications to be able to develop valuable and user-friendly tools. On the other hand, IT doesn’t know how important the application will be compared with other departments’ projects and, consequently, doesn’t know what priority to apply or which resources to allocate to a development.

In the past, compromise and disappointment have often been the outcome, on both sides. But with LCNC, treasury can develop its own applications and user-friendly front-ends. If a more robust version is required, or the application is going to be used long term, IT already has a fully working application on which to base its development – something much easier to learn from compared with documents, spreadsheets and the hand-written meeting notes used in the past. A working prototype saves both sides significant time and effort.

Shallow learning curve

LCNC platforms are easy to learn because they incorporate an intuitive ‘drag-and-drop’ method of creating applications. By dragging and dropping blocks, the whole NC build process becomes a visual experience – a person in treasury discovers a need and uses the NC platform, which typically offers hundreds of pre-built blocks, to build an application that satisfies the requirements.

The code that makes things happen in the blocks is unseen for NC systems. Not only is this far easier for the beginner to grasp, but the block coding format used in all NC platforms is also an excellent way to begin to learn computational logic. Once thinking about applications in these terms, new possibilities open up. For example, an initial project might involve setting up a database for treasury KYC documents, automatically renaming the documents so that they are easy to find. A subsequent extension could be to automate sending out emails with requests for records to be updated and filing documents that are sent back, also by email. Later, a further enhancement could be to send information from the database to whoever or wherever the data is needed, whenever.

After starting with an automated NC application, using it and refining, it takes only one additional short step to create more complex and scalable applications by adding LC tools.

In an LC system, the code is visible but purposefully limited in complexity (hence ‘low-code’). The more professional an application is in appearance and functionality, the more likely it is LC rather than NC. Again, LC starts with drag-and-drop simplicity but now adding extra functionality and sophistication requires some rudimentary coding knowledge. Like NC, LC can also be self-taught, with many online tutorials available (simply search ‘I have x problem, how do I resolve it…’). The learning curve for most is not steep: platform developer Omnia found that 70% of new LC users gained the necessary skills in under a month, and 30% did so within two weeks.

From little acorns…

Among the first platform providers was Zapier, a cloud-based drag-and-drop tool launched in 2011, primarily for process automation. Fast-forward a few years and, in Zapier’s own words, the technology enables users “to automate the most tedious parts of a day-to-day job”.

Now, with many more platform providers, users can do far more than automate simple process flows. LCNC applications can bring together information from multiple data sources, including OCR, manipulate the data, perform calculations, make applications run and show results. Even AI analysis of the data is possible. At the higher end of the platform market (away from the free platforms), it’s possible to self-build professional-looking and enterprise-scale applications that external and internal partners can use.

… to a big forest

The arrival of LCNC for treasury can be traced back to the shift from installed technology to cloud-based SaaS delivery. This development happened in parallel with IT using infrastructure as a service (IaaS), with offerings from Amazon’s AWS and Microsoft’s MS Azure, and platform-as-a-service (PaaS), which sits on top of IaaS and manages the operating system and middleware IT systems. This outsourcing trend has grown into ‘x-as a-service’, a generic term covering an increasing range of cloud-based services previously provided only locally (banking as a service, BaaS, for example).

PaaS, in particular, provided the cloud-based development and deployment environment necessary for LCNC with its templates, pre-built connections and code blocks, enabling users to create their application ideas from scratch.

There are many LCNC providers, and, naturally, Microsoft is competing hard to be a leader in the space. Power Automate, essentially an RPA tool for business process flows, offers NC drag-and-drop capabilities with hundreds of pre-built connectors. While Power Automate obviously works well with all MS applications, it can, in addition, launch and integrate with non-MS software, including competitors’ and, although a simple tool, also provides AI functionality.

An example any treasurer can relate to is cash forecasting. Excel-based forecasts are prone to error and/or are difficult to maintain but modules available from external providers are often never quite perfect for a company’s individual needs. Using LCNC, a powerful cash forecasting application tailored to the company’s specific requirements can be built starting small and becoming more sophisticated over time. In this context, users can automatically download data from multiple sources into a single pool, ready for analysis, report-building and presentation. In addition, using the same data acquisition and aggregation processes, flows of policy documents, audit and compliance reports can be automated enterprise-wide. The list of possibilities is endless.

Sticking with Microsoft, another LCNC-enabled platform of interest is Power BI, a well-known data visualisation tool. It uses API connectivity to import information from data sources such as the web, ERPs and TMSs. With Power BI, users can create interactive data analytic dashboards that display both high-level and detailed views and offer scenario-planning functionality.

Some LCNC platforms also offer advanced enterprise-level scalability. For example, applications such as Inventify enable more sophisticated (but still non-IT) users to handle more data items and deliver more robust security and performance. This performance-at-scale can be ideal in a situation where, for example, reconciling forecast versus actual cash involves trawling through hundreds of thousands or even millions of invoices and cash flows.

It’s worth pointing out that at this stage of LCNC’s development, treasury can use these tools, but they are not treasury-specific. But this is an advantage. Because LCNC platforms can facilitate automation and add new workflows from multiple sources, they don’t need to be a treasury product per se. The tools’ function-agnostic capability means that different departments – and IT – can share costs and integrate applications across departments, which external systems often struggle to do.

Catch-22?

All that being said, LCNC has a few downsides to consider. Perhaps most pressing for treasurers is finding the time to use it. Small and shrinking treasury functions may feel they cannot add to their daily workload by becoming citizen developers, their departmental resources already being stretched to the limit.

They may also feel they do not have the time to start and maintain the documentation of LCNC development work. Documentation is not obligatory, but it is essential if a system is to have a life beyond its developer. As with any software, if the knowledge of how the application is built and works is unavailable due to a lack of documentation, the tool is already on borrowed time.

Depending on the application’s functionality and level of integration, IT may require proper documentation from the outset, not least for security purposes. LCNC platforms typically use encryption, but this might not be enough. Where APIs are used, IT will undoubtedly want to assess their security before approval to operate. Clearly, there is a balance to strike between freedom of creativity and control.

Another conundrum is that while the time-poor nature of treasury is partly solved by process automation, this is often achieved by using plug-and-play proprietary systems. Of course, LCNC also brings easy automation to the table, but for some, its adoption may be a catch-22: you can’t save time with it unless it is built, but you can’t build it if you don’t have the time. Deciding which to use will be a question for the individual treasurer, with the blessing of the C-suite. The availability and waiting time for IT resources will be a critical consideration.

One challenge for LCNC platform use is a potential issue for the whole business: data quality. Process automation, in treasury or any other function, significantly increases dependency on data integrity. This applies whether using LCNC, a bought-in system or an internal IT-built tool. A business needs good-quality data. It can be de-duped and cleaned, yet as soon as more than one version of the truth exists (lack of version control in spreadsheets is common) or there is a change in a data source, the integrity of data will rapidly degrade. After all, it’s easy to understand that automating rubbish creates only more rubbish.

While LCNC adoption in treasury is unlikely to spark a data management revolution, as automation projects in general rise up the corporate agenda, so too will the realisation that data quality is paramount. Some larger firms are already taking this seriously, with the appointment of a Chief Data Officer with the backing of the CEO and board to make necessary changes.

Time to engage

As a result of the time constraints mentioned above, successful LCNC applications are those where the creator start smalls, get under the hood and make iterative additions, adjustments and fixes, and develops over time. Bought-in systems tend to be ‘black boxes’. When complications or new needs arise, external specialist intervention is needed. This may add to the cost of ownership, but, more importantly, if help is not immediately forthcoming, it may delay the execution of that business process. If sufficient people in a treasury department have knowledge of the businesses’ LCNC applications and management support to automate, solutions can be created much faster.

In a world where reaction speed is a business imperative, it’s worth noting research and advisory company Forrester’s assertion that getting results from LC development platforms is potentially 10-times faster than traditional development techniques. This means that, yes, LCNC is a valuable tool now, but also that, given the growing trend among corporates for digitalisation, it could become essential for treasurers to understand them, if only to stay in line with the rest of the business.

Training and enabling one or a few tech-savvy employees is a simple way to start using LCNC in treasury. Given the many possibilities, though, companies may begin favouring the recruitment of people with pre-existing skills.

The best way to start is to build a small and usable application, work on it until it does what is required, and then keep going. The capacity of the more professional enterprise-level LCNC platforms to enable applications to grow with the business ensures considerable opportunity for those willing to engage.
With LCNC experience, treasurers and other corporate professionals can support their organisations’ digital transformation with the exact tools they need, not the tools someone else thinks they need.

Here’s a summary of what to expect.

LCNC at a Glance

Upsides

  • It requires little or no prior programming knowledge
  • LCNC can mainly be self-taught using free online tutorials
  • Basic platforms are often free or free add-ons to existing purchases
  • It uses an intuitive ‘drag-and-drop’ block-based method to build applications to which simplified code can be added if needed
  • Platforms offer hundreds of pre-built connectors to load data, manipulate it, manage workflows and integrate with many providers’ systems for whatever the need is
  • Applications can range from simple, short-term usage, to complex, scalable, enterprise-wide tools
  • It enables rapid application development and maintenance; reliance on IT or vendor resources is reduced or removed

Downsides

  • LCNC requires time and effort to get the best from it
  • For robust performance, good-practice application-building processes should be implemented. If there will be many LCNC applications, training and treasury- or company-wide standards may eventually be needed
  • Applications need to be documented as fully as possible to ensure service longevity
  • Encryption is a feature of most platforms, but IT approval may be required, especially where third-party access is enabled
  • As with every automation project, data quality is essential