Ex-Google engineer bids to fix core banking

Ex-Google engineer bids to fix core banking

After two years in stealth mode, a British startup run by a gaggle of former Google engineers has come into the light claiming to have solved the greatest challenge in fintech: how to replace core banking systems.

While banks busily build customer-facing technology for the digital era, they are being hampered by underlying software written decades ago which is no longer fit for purpose, says London-based Thought Machine.

The firm cites the Royal Bank of Scotland's infamous IT meltdown of 2012 and the words of regulator Andrew Bailey as he later handed the bank a record £56million fine, warning of a "very poor legacy of IT resilience and inadequate management of IT risks".

Thought Machine's answer is Vault OS, core banking software created for the cloud which uses private blockchain-style technology and has cryptographic ledgers for watertight security.

The company is run by Paul Taylor (pictured) who joined Google in 2010 after it bought Phonetic Arts, the speech synthesis technology firm he co-founded. Taylor pulled in several Google engineers as well as experienced bankers as part of a 50-strong team based in East London for his new venture, which has been in the pipeline for the last two years.

The firm claims that its OS is completely flexible and handles any standard banking functions or business models. "Best of all, it employs smart contracts and machine learning to enable banks to offer a vast array of financial products - or roll out new ones in days. It is 100% future-proof."

Meanwhile, because it runs in the cloud, banks can scale to millions of customers without expensive in-house data-centres and the large operations teams typically employed to keep legacy systems from failing.

When it comes to compliance, every transaction is reported in real time so banks can determine their exact financial position at any moment. These tools are in the core of Vault OS, making implementation of capital-adequacy standards such as Basel III automatic.

Says Taylor: "People have suffered for too long from the archaic software banks are built on and their CIOs stay awake at night worrying if these systems will cope with one more day. It’s time for new ideas, built with today’s technology. Vault OS fixes broken banking and will be the engine for the banks of tomorrow."

Comments: (20)

A Finextra member
A Finextra member 13 July, 2016, 18:231 like 1 like

"It is 100% future-proof."  Wow! Now, that is an OS, being able to predict the future to know it is 100% future proof. Hype aside.  It sounds very interesting.  Good luck to them.

A Finextra member
A Finextra member 13 July, 2016, 23:171 like 1 like

Core Banking is an area few have dared to modernise (with varying degrees of success) - and those that do tend to be burned out by requirements for the replacement system to almost be exactly the same as the incumbent system - often with continually changing requirements as a never ending moving target.  It is nice to hear people are willing to take a risk on presenting a modern platform - it is a gamble.  With a modern Core it will require modern peripheral applications to support Card/Mobile Issuance Management; Transaction Authorisation; and Fraud & Credit Risk management to complement it.

Juergen Rahmel
Juergen Rahmel - IETC Information Engineering Ltd - Hong Kong 14 July, 2016, 02:271 like 1 like

Agree with Matt. Actually, reinventing 'core banking systems' shouldn't be that hard. In its core it's a high availability transactional system. Processing a large amount of transactions should be a commodity, nothing special. Be it payments or securities or similar. And yes, if you do it today, you add in real time, block chain, smart contracts, machine learning and self-driving to fill your buzzword-bingo sheet like everyone else.

However, what banks struggle with is the integration into dozens of applications in the banks' system universe. The interdependencies are massive. Bringing the core into RTTaaS (Real time Tx as a service...) is great, but the whole lot of other systems needs to be exchanged or adapted accordingly. Thus, moving into the right direction requires much much more work and investment than just a better core. It will become a revamp of the whole architecture. This makes it so incredibly expensive and difficult.

And, it is a project that can't be done with new systems and new resources alone. It has to be sold to the work force that was happily supporting the 'old world' over decades and feels threatened by any move into new directions. This is not a plug-and-play exercise, and that's why it is procrastinated all too often. 

Steven Hatton
Steven Hatton - Trusek Ltd - Amersham 14 July, 2016, 06:33Be the first to give this comment the thumbs up 0 likes

If the banks were to adopt an integration platform or bridge to their core platform to manage all the peripheral applications and third party providers meaning just one point of integration, this would take a lot of the strain from the core platform, manage the security issues of having multiple API's and allow for a controlled transition to new systems. Or even allowing the old core platform to continue while new applications and developments continued on a new platform.

This would make the whole process of modernising the banks system less terrrifying.

Ben Hulme
Ben Hulme - BNP Paribas - Singapore 14 July, 2016, 06:482 likes 2 likes

It's a part of the infrastructure that has fallen so far behind, overlooked due to the fact the core is not a direct revenue generator and the mentality of 'if it ain't broke don't fix it' has probably been a key factor.  The reliance on so many other applications mean that the any development is patchy and layered, without the risks of any major work.  When you look at technological advances in Trading AI and and machine learning within the trading world, you wonder why the main core is still in the dark ages.

Kris Hansen
Kris Hansen - SAP - Vancouver 14 July, 2016, 07:142 likes 2 likes Sounds very interesting for sure, and I agree that the sprawl and complexity of legacy core systems needs to be addressed - however, the challenges associated with transforming core banking are more acute on the operating model and change management side. Technology cannot make this easier, it is about re-envisioning how the organization works and breaking down silos while simplifying. Without a compelling business vision even the best software will sit on the shelf!
A Finextra member
A Finextra member 14 July, 2016, 08:27Be the first to give this comment the thumbs up 0 likes

worked with some of the major core banking applications (T24, Finacle &
Flexcube), it will be interesting to know how this technology will be able to
support the business of the bank.

The business case to adopt
cloud based solution to drive the business of the bank needs to be watertight.
Providing quantitative and qualitative benefits to support this move shouldn’t
be that difficult. Some factors to aid technology adoption also need to consider.

1) Connectivity:
Especially in the less developed countries where communication bandwidth is a challenge.
This may lead to poor user experience resulting from limited network throughput.

2) Integration:
Availability of standard APIs is also a factor that should be considered. Implementation
of a service-oriented architecture (SOA) such that one ESB
manages communication with other peripheral applications is also important.

3) Extensibility:
Due to the ever-changing business needs, ease of extending this application by
the banks technical team is very important. The turn-around-time to deliver
business requirements will distinguish a bank and in most cases determine its profitably.

A Finextra member
A Finextra member 14 July, 2016, 08:591 like 1 like

At last!Agreed Ben Hulme!

A Finextra member
A Finextra member 14 July, 2016, 09:061 like 1 like

Problem is "Core Banking" isn't a single platform or set of infrastructure for many Banks. The "If it aint broke don't fix it" adage is fine until it breaks and there's no-one left who knows how it really work, after usual staff attrition combined with major cost cutting by Execs who didn't know what they we're doing to their "Core Banking" systems.
It's not the systems that fail, but the processes and knowledge that understands how to operate those systems.
Likewise i'd be interested to see what VaultOS can achieve

A Finextra member
A Finextra member 14 July, 2016, 09:171 like 1 like My only concern would be: looking at the team involved - no one stood out as having significant domain experience (in Core Banking platforms). It can be the best Tech in the world but if it lacks fundamental functionality it isn't going to fly. "Sadly since filming this programme"...
A Finextra member
A Finextra member 14 July, 2016, 10:35Be the first to give this comment the thumbs up 0 likes

Its always integration that causes issues - and of course the changing world we live in -  This does sound very positive but the core has to be wrapped in applications to handle payments, regulatory compliance, KYC and AML systems etc  it sounds like decision makers thinking about buying or building banks have another choice. until proven in the field this solution is going to be somewhat risky so i would imagine its a start up proposition we will see first to market before this is embraced by any of the incumbents

James Piggot
James Piggot - Finastra - London 14 July, 2016, 10:372 likes 2 likes

The mention of kernel and emphasis on daily builds and commits on the Thought Machine website does make this seem like they are building the Unix for banking. A platform on which banking core systems can be built rather than an out of the box core banking system?

Building a new core banking system for large banks is challenging, Oracle and NAB have been building their NextGen banking platform for the past nine years. It has reportedly consumed hundreds of millions of dollars and is not yet finished.


Stephen Folan
Stephen Folan - Atos Origin - London 14 July, 2016, 13:081 like 1 like

Good luck to VaultOs. New entrants to the banking system will have an easier time getting their infrastructure in place. Their costs will be lower and then the incumbents will have do do something about it rather than 'mend and make do'.

A Finextra member
A Finextra member 14 July, 2016, 13:46Be the first to give this comment the thumbs up 0 likes

I don't think Vault OS is intending to replace existing core banking system as such projects are more about project management and system integration than anything (as most comments already state). I see challenger banks and new banking "arms" of brick-and-morter banks as the addressable market. Built new instead of try to change existing plattforms. The later will miserably fail as we have seen in other disrupted markets already.

Ketharaman Swaminathan
Ketharaman Swaminathan - GTM360 Marketing Solutions - Pune 14 July, 2016, 16:401 like 1 like

A quick glance at ThoughtMachine's website should be enough to persuade new banks to seriously consider VaultOS as their core banking system. But I'm not sure whether it offers a compelling enough value proposition to prompt existing banks - using mainframe or open CBS - to migrate to VaultOS. Ergo, the company's mission  "replace core banking systems" - is somewhat hazy.

João Bohner
João Bohner - Independent Consultant - Carapicuiba 15 July, 2016, 01:422 likes 2 likes


"the company's mission  "replace core banking systems" - is somewhat hazy."


In my opinion is considerable hazy:


"Single source of truth for all transactions"

 This is possible, for any type of product, only, and only thanks of a STR - Standard Transaction Record  Architecture for any kind of 'product'. It does not seem that the VaultOS, support this architecture.

"SkyScraper allows applications to safely and securely connect to an existing bank account."

 So, this assumes that there is already a bank account, and therefore an old 'core system' is in place ...

"Once connected, you can read transactions and make payments."

 Once connected... So you're going to connect with the 'old' system!

Where is the replacement of the old 'Core System'?


You can read the transactions from an existing bank account... in which format? Standard?

You can make payments... How about other transactions such as: collections, depreciation, appropriations, monetary restatement, fines, royalties, default interest, dividends, contingency among many others?

"...categorising transactions into 60 separate types."

 Why 60? Why not 50 or 400? Or a flexible number of categories, according to the context?

Maybe Paul Taylor will show to us the blue skies of VaultOS, briefly...


Ketharaman Swaminathan
Ketharaman Swaminathan - GTM360 Marketing Solutions - Pune 15 July, 2016, 10:48Be the first to give this comment the thumbs up 0 likes

@JoãoBohner: All good catches! I'm coming around to your view that ThoughtMachine's mission is considerably hazy!!

A Finextra member
A Finextra member 22 July, 2016, 11:302 likes 2 likes

From personal experience, everyone's comment on here is extremely valid.  The political, financial and reputational issues associated with core banking platform replacement have sunk many ships in the 30 years I've been involved in this area.

This remains a holy grail that has long remained just out of reach and at best is solved by metadata levels, operational data stores and the 80/20 rule.  It's not so much the technological challenges, more the internal vested interests and empires that have grown up to support banks' current core banking solutions.

My personal view is that the "me-too" investments by banks in Fintech are largely a smokscreen for 'skunkworks' developments around Blockchain to discover whether it really is the panacea for the woes affecting banking.

I'll watch this space carefully, but my bet is that this company is acquired by one of the big banking application vendors befiore it ships its first product.

A Finextra member
A Finextra member 27 July, 2016, 06:391 like 1 like

As a concept i like the thought process of 'core banking as an OS' and perhaps it might solve some of the problems which trypical core banking applications face.

But more than anything else, i feel what most banks failed in the past is where they percieved and used core banking just as 'system of records' and 'automation of tasks' to some extent.  

The key to the success of any bank with their core systems is when they have thought of it in terms of 'Architecture'. IT Architecture that is completely cohesive to 'Business architecture' and business model of the bank. This is what is always broken.

Businesses evolved faster and IT Architectures remained frozen in time. The holi grail is there. Perhaps techologies also didn't help as well (because it was never built for agility).

So my take on Vault OS is that unless and until they think this as a 'constantly moving ship', and build an Operating Environment with technology that can morph itself with time, their innovation would be perhaps criticized in 5 years time.

I am sure they would have thought about it. I wish there is much more to it than what we know just on the basis of this article.

Best wishes to them, and good luck to our industry for many more such innovations.

João Bohner
João Bohner - Independent Consultant - Carapicuiba 27 July, 2016, 13:12Be the first to give this comment the thumbs up 0 likes

@Niraj Vaidya,

Fully agree with your thoughts.

The key is 'an ARCHITECTURE for the constantly moving ship: Technology AND Business.'