RPA: The Smarter Way to Reconciliations
By Neil Vernon, Chief Technology Officer, Gresham Technologies
Artificial intelligence may be the future. But when it comes to reconciliations, the present belongs to robotic process automation (RPA).
A recent industry found that, while organisations are starting to integrate AI into their businesses, so far the majority are enjoying little to no return on their investments.* Although 9 out of 10 see artificial intelligence as a business opportunity, many initiatives fail, with 70% of companies reporting minimal or no impact from AI.
What will be surprising to many is that the financial sector has one of the lowest AI success rates of the industries surveyed. According to the same study, 28% of financial sector respondents reported a positive business impact from AI in the past three years. In technology and IT, the figure is 48%.
Ultimately, AI has potential, but it is not ready to be the saviour for the unsolved problem of reconciliations – even in today’s ‘new normal’. With financial services firms facing intense margin pressures, the emphasis must be on initiatives that can deliver proven and substantial efficiency benefits. RPA, done right, is that technology.
Streamlining reconciliations is not new, but it is more vital
Fast and scalable reconciliations and exception management resolution has always been important. Now, it is essential. The volume, speed and complexity of trades continues to grow, as do firms’ regulatory reporting obligations. Being data confident at all times has become imperative.
For years, reconciliation was a post-settlement date activity. However, most errors happen on trade date, and typically in the first few seconds. Waiting five days or more before detecting a problem is ruinously expensive. Settlement failures result in reputational risk and significant manual effort to identify and rectify the source of the problem. And the cost of failure will only rise once the Central Securities Depositories Regulation’s (CSDR) mandatory buy-in provisions come into force from 1st February 2021.
Industry participants are increasingly recognising the value of conducting more, and more frequent, reconciliations that start on trade date and extend throughout the trade lifecycle. Carrying out market-facing reconciliations with relevant third parties such as brokers and custodians is one aspect. But the onus has been directed more recently towards inter-system reconciliations within the organisation, as well as reconciling with regulators on trade date to meet transaction reporting requirements.
Trader frauds and misreported transactions are more easily detected once reconciliation processes are stepped up. Failed trades often stem from an undiscovered upstream problem and would be significantly reduced with early and ongoing reconciliations, saving both money and regulatory censure. Regulatory demands for accurate, reliable and timely risk data aggregation and reporting under the Basel Committee reinforce the need for robust, automated reconciliations across an organisation’s various data sources. Now, everything is a reconciliation.**
Tackling exception management with RPA
Investments in more modern enterprise platforms means that real-time reconciliation needs have picked up in response. The big challenge with reconciliations though is how to automate the exception management and resolution processes. Which is where RPA comes to the fore.
Surging reconciliation volumes mean organisations are finding more errors. RPA’s strength lies in fixing those errors at minimal cost.
Success lies in the right approach towards RPA. It is not about integration on the cheap, using robots to connect different platforms together in place of human actions. That approach will only store up bigger problems for the future. Any time one of the platform vendors upgrades or changes its technology, the robots will no longer work, and the systems will stop communicating. Organisations’ technology infrastructures become brittle, freezing them on their existing – and quickly outdated – platform versions.
Instead, where RPA can deliver real value is by automating regular, repeated activities within a platform – which is why it is ideally suited to reconciliations processing.
Partnering for real world success
Collaborating with a provider able to deliver proven technology solutions that offer measurable automation benefits, rather than hyperbolic marketing language, is key.
Experience as a team, with technology solutions that work and a track record of success across different market segments, is what ultimately counts in the reconciliations world. Firms need partners that understand reconciliations as a business and have market experts that can tailor offerings to solve even the most complex of problems. It’s the only way to collaborate. This means partnering with technology providers that prove they can reduce time and costs, freeing up users’ resources to fix the root cause of their systemic problems.
With a combination of broad, cross-customer experience and an understanding of good practice, firms will be ideally positioned to take advantage of the opportunities the new technologies bring and deliver a more optimised reconciliations process.
Focus on what works
Artificial intelligence and machine learning may be the future. In time, they hold the possibility for systems to learn from reconciliation exceptions and intelligently respond to them. But the technology is not there yet to meet firms’ current, real-world needs.
AI and machine learning have become popular industry buzzwords, and the technology providers will continue to monitor and work on developments in the space. Talk about how they can solve the industry’s ongoing reconciliation challenges is, for the moment, just that though – talk. In the meantime, firms need solutions that actually work. And the most effective, proven capabilities offering the biggest automation benefits are to be found with RPA.
* Survey of more than 2,500 executives from 29 industries in 97 countries by MIT Sloan Management Review and Boston Consulting Group
** Banking Supervision’s standard number 239 (BCBS 239)
AI/ML on the buy-side trading desk is a long-term program with short- and mid-term deliverables.
Aiden uses the computational power of deep reinforcement learning to improve trading results and insights.
BondDroid’s AI-generated prices are integrated directly into LTX’s pre-trade analytical tools.
Fintech firm has partnered with Jim Toffey, founder of Tradeweb Markets, to create LTX.
Tradeweb AI tool makes it possible to calculate reference pricing in seconds based on dynamic market data.