Memoirs of a Citibank technology warrior - How the bank built the CORE (Part 1 of 4)

• Ajit Kanagasundram shares that Citibank’s first consumer banking system had to be abandoned after spending $50 million 
• The CORE project succeeded in rolling out an international platform with superior functionality for over 18 businesses worldwide within four years
• The project had a robust base platform, good technical direction, business buy-in and no micromanagement from New York

This is the first of four parts. Click here for part two: The global cards project;  Part three: The “Systematics” project; and part four: Lessons learnt from the “Global Go to Common” project. In this part, Kanagasundram shares how Citibank built its retail banking system and the lessons learnt in the process.

Ajit Kanagasundram, formerly managing
director and head of automation, Citibank

There has always been a conflict between those who want totally centralised processing and those who prefer a degree of decentralisation. An example is the Global Go to Common, where over $1 billion was spent without achieving the required results and which was cancelled last year. This shows the folly of the former model – nearly 40 years after it was first tried in Citibank with the Consumer Banking System (CBS) project but on a much larger scale.

The 70s and CBS

John Reed when he was the head of consumer banking was a true visionary and a man of immense drive. He realised first and more than any other senior bank executive the potential of technology to transform consumer banking. He will be remembered for 3 major projects – one an unqualified success and the other a partial success or failure, depending on your point of view. The third was his sponsorship of the global telecommunications network (GTN) which I will expand on in the section on the Asia Pacific processing centre (data centre), and which was not only way ahead of its time but a total success.

The unqualified success was his pioneering of the automated teller machine (ATM), which is commonplace today, but which when he launched it was thought to be impractical and prone to failure and vandalism. He proved the naysayers wrong by acquiring a lab and manufacturing facility in California called Transaction Technology Inc (TTI), which developed and manufactured Citi’s own ATM machines with a revolutionary touch screen.

The other effort was to build an integrated, all-purpose banking platform, named the consumer banking system (CBS) –and the best brains in the bank, TTI and outside were recruited for it. They did all the right things – built a data model with an integrated Customer Information File (CIF) linked to all product processors. This was the first time that such an ambitious project had been attempted but despite the best efforts of all concerned it was abandoned after a few years after $50 million had been expended – a huge sum at that time. It failed because it was a bridge too far – the technology infrastructure was not mature and they attempted too much in one go – a mistake that was to be repeated on a much larger scale 35 years later with the Global Go to Common project.

But the CBS project had two positive legacies – it created a cadre of trained bank technologists who would have a significant impact later, and the customer data model created for it would later re-appear as the global relationship banking customer relationship model that was implemented in South America and Asia. It was therefore not a total waste of time and effort as it laid the foundation for later successes like the CORE project and systematics.

The CORE project

The next attempt to create an international platform had modest beginnings. Jack Berger, a senior technologist in New York, was looking for a banking system that could be rolled out internationally to serve the growing consumer businesses in Latam, Europe and Asia in mid-1985. He found it in a small software house in Florida called Newtrend – they had a banking package that ran on a mid-range IBM computer.

There were two unique aspects to this banking system - first it was based on the System 38 machine which had unique futuristic technical features – an integrated database and communications interface, a capability where the operating system handled file placements and other unique features all based on a 48-bit addressing system. All way ahead of its time and also much cheaper to run as IBM did not charge separately for the operating system software as they did for mainframes. It also required much fewer technical staff for maintenance. The second feature was that the Newtrend banking application itself was an integrated software package with DDA, loan, time deposit and general ledger processors all linked and integrated with a customer information system, all utilising the unique advanced features of the IBM System/38.

But the drawbacks were also daunting – it was designed for small banks in the South East of the US with none of the features that were required for an international bank, chief of which was a multi-currency capability. But it had the structural integrity on which to build the advanced features required for international banking, so it was chosen.

The system was renamed as CORE – Consumer Oriented Retail System. The team dispersed to the respective regions and collected requirements and passed them on to the technical team in New York to develop a version of CORE that was suitable for Citibank’s international business. By mid-1987 CORE release 1.0 was ready and implementation started for Japan in Asia and in Europe and Latam as well. There were subsequently six monthly new releases of the base with added functionality. Within three to four years, all the international businesses were done with the exception of the large business. The individual businesses were responsible to fund their own projects and hire their own technical staff. New York provided the base software and generous technical assistance where required.

The CORE project succeeded in rolling out an international platform with superior functionality for over 18 businesses worldwide within four years with a very modest central outlay – this was the world's first for any bank. It succeeded because it did not overreach itself with too ambitious goals, the base platform was robust, the technical direction at all levels were superb, there was no micromanagement from New York and above all they got the buy-in from the businesses. It allowed for a standard platform globally, while resisting the temptation for total standardisation, which still allowed for the exchange of functionality between businesses. It also fitted in with the prevailing management ethos of “Paradisation” – a clumsy term describing the strategy of pushing technology management to individual units rather than centralisation.

Click here for part two: The global cards project;  Part three: The “Systematics” project; and part four: Lessons learnt from the “Global Go to Common” project. 

Watch the TABLive interview with Ajit Kanagasundram.

Ajit Kanagasundram is formerly managing director, technology, Asia Pacific at Citibank. The views expressed herein are strictly of the author.