HISTORY >> MADE FASTER... why automation often fails, but doesn't have to | NTT DATA

Whitepaper - Tue, 12 May 2020 - 10 min read

HISTORY >> MADE FASTER... why automation often fails, but doesn't have to

With business norms swept away by the 2020 COVID-19 crisis, change must go deeper and faster than conventionally thought possible.

Download Whitepaper

Automation for London: Driver of change, or poor performer?

With business norms swept away by the 2020 COVID-19 crisis, change must go deeper and faster than conventionally thought possible. As we emerge from the biggest health crisis the UK has seen, how can London Market players protect revenue, fix inefficiencies and ready their businesses for the rush to recovery?

The answer is a refocusing of technologies, working practices and business models to create an environment that may be starkly different to the current ways of working.

London is among the world’s most expensive insurance markets. For every £1 spent by Lloyd’s customers, around 40p is consumed by costs – either fixed costs or commissions paid to intermediaries involved in the distribution chain. That’s unsustainable in the face of fierce competition from hubs around the world that deliver products faster at lower cost.

However, there is appetite for change. A 2019 London Market survey1 commissioned by NTT DATA reveals that 59% of stakeholders recognise digital as highly important to their business today, rising to 97% for its importance in three years. Furthermore, 70% believe automation can significantly benefit business development.

“New business models such as hiring robots by the hour and avoiding infrastructure and licence costs with a managed service are eliminating many of the risks of business automation that concern insurance executives.”

1 Coleman Parkes Research, survey on behalf of NTT DATA of Lloyds of London focused businesses, September 2019

Automation: great expectations, often unfulfilled

Automation is typically sold on the promise of:

  • Running low value administrative tasks now done by humans
  • Offering costs that are just a third of a human worker
  • Reducing manual processing time from hours to minutes
  • Improving data accuracy and consistency
  • Releasing staff to focus on more valuable work

The reality of automation is often very different. Firstly, organisations often do not understand the complexity and the detail of what’s required to automate processes. The misconception that it’s the technology at fault when a robot fails, instead of the overall change programme, leads to dissatisfaction and loss of confidence.

The traditional approach of buying, hosting and running automation has given it a poor early reputation. The technology is ready and mature. That’s not the problem. The issue is the flawed approach taken by many projects from the beginning.

Automation can also appear to be a complex technology that is only understood by the technical experts. While it’s true that the technology involved is sophisticated, the basic concepts and benefits to be gained are straightforward. It’s important to demystify automation and show that the technology does not inevitably mean the loss of jobs, but will deliver lower costs, greater efficiency and higher profits for improved business performance.

It’s also true that today, these benefits are more achievable thanks to new business models, such as hiring robots by the hour and avoiding infrastructure and licence costs with a managed service that eliminates many of the risks of business automation that concern insurance executives.

Automation relieves insurer’s manual processing burden

For Tokio Marine Kiln (TMK), a key area to address was the manual processing of data. Low value data processing still involved a lot of manual input and there were plenty of systems that were not interconnected, increasing the burden.

As well as the people costs, manual processes caused a time lag between data entry and the availability of that data – limiting the time available for highly-valuable data analysis. It was also prone to errors.

TMK explored Robotic Process Automation (RPA) as a method to minimise manual data input and processing and free up time for analysing data. The concept was appealing, but the fact that TMK handles a relatively small number of high-value policies meant the high, upfront costs involved in a traditional RPA deployment simply wasn’t viable. NTT DATA offered TMK the ideal approach – a completely managed RPA service that could be scaled up and down at will and with a simple, monthly cost.

For Tokio Marine Kiln (TMK), a key area to address was the manual processing of data. Low value data processing still involved a lot of manual input and there were plenty of systems that were not interconnected, increasing the burden.

As well as the people costs, manual processes caused a time lag between data entry and the availability of that data – limiting the time available for highly-valuable data analysis. It was also prone to errors.

Automation has suddenly taken on new urgency

The 2020 global health crisis has transformed the way we work more deeply and more quickly than was thought possible. One of the long-term lessons of the crisis is that new ways of achieving business resiliency need to be built in. We have also learned that it is possible to transform how organisations work at pace and with agility.

Automating processes is one of the most effective ways to build robustness – even if key personnel become suddenly unavailable, you can be sure tasks will be done.

Automated processes can be run on existing infrastructure and simple processes can be developed remotely in only one or two weeks.

Conventionally, the decision on which processes should be automated was largely based on the number of people involved and the frequency at which they ran to achieve a reasonable return on investment. That’s still true, but in addition it’s important to think about the need to automate critical processes no matter how small or how little time they take.

Five London Market insurance processes ripe for automation

Automation can be applied to almost any London Market business function.

Here are five examples of processes that are traditionally highly labour intensive, but which can be automated relatively easily. In each case, NTT DATA has applied automation to substantially increase performance.

Underwriting – manual risk entry into Policy Administration System (PAS), collation of compliance checking, policy document generation

Taking risk information from a defined format and moving it into an internal system. Pulling risk detail from external systems via a common key (e.g. flood rating via postcode).

Generating policy documentation from details in an internal system.

Claims – First Notification of Loss

RPA paired with chat bots can allow many customers to be serviced effectively and efficiently with minimal human interaction to reduce contact centre strain.

Broker – Creation of the slip and collation of the claims history

Automatically collecting claims history from multiple sources (internal systems, databases and spreadsheets) and collating these into one view for use by the broker. Collecting details of the insurance contract into a pre-defined template, saving manual time to populate fields.

Finance – cash matching, Bordereaux, invoice and PO approval

Collating Bordereaux information into reports and/or dashboards for management. Extracting and recording invoice information (use of OCR for disparate invoice layouts).

Human Resources – new starters and leavers process

Templates for new starters, leavers and changes of accesses can be populated and assessed based on Active Directory which should be the main source of the truth. Moreover, this allows for a self-service model, freeing up line managers’ time, who can be inundated with this kind of work.

Avoid the common pitfalls: five ways to create automation success

1. Choose the right processes to automate

Make automation decisions based on fact, not emotion. Typically, the processes that people dislike the most or are the most complex and time consuming to complete are given priority for automation, rather than the ones that will deliver the most benefits.

In one case, a mid-sized specialist commercial insurance company decided to automate 20 processes. However, an analysis by NTT DATA showed that only five of those processes were appropriate for automation.

2. Robots aren’t always the answer

Assess a process and decide how it could be improved before considering it for automation. Often, a process can be re-engineered to make it run far more efficiently without the need for automation.

In one example, a specialist insurance company wanted to deploy automation to a process to achieve a 34% reduction in time and effort. However, NTT DATA showed that by re-engineering the process, an 86% benefit could be achieved before deploying automation.

3 Start small and build up to an end-to-end process

Attempting to automate an entire, major process at the outset of an automation programme can lead to unnecessary complications and implementation issues. Establishing the initial automation project to encompass the setting up of the environment, such as the IT, integrations and governance, at the same time as the process reengineering and automation can cause unmanageable complexity. Teams become frustrated and automation is deemed too difficult.

Instead, start by automating a simple task to prove that automation works and then progress to the integration stage. A good example is the collection of policy data from a broker, which is then automatically placed into the administration system.

4 Don’t forget to engage the team

The people who will be working with the bots will also feel the most threatened by their implementation. Unless effective change management is put in place with clear articulation of the automation objectives, it is likely that resistance from teams will build as they fear they will lose their jobs.

Communicating objectives is vital with guidance on what staff roles will be post implementation and how automation will help them in their work. Teams will then feel ownership and actively support the changes, as automation brings about new types of roles which can be enabled by providing options for re-skilling staff.

5 Avoid over- committing and paying too high a price

Not every implementation will result in the projected benefits. It is important to be pragmatic and avoid committing to long consultancy engagements, high upfront infrastructure costs and long-term licence agreements.

A common example of over-commitment is organisations paying licence fees for several robots when the automation could be completed by one.

Instead use a pay-as-you-use service to minimise costs and eliminate costly and wasteful commitments. Defining approaches to data is at the core of all systems and vital to ensure individual modules can be introduced, removed, replaced and seamlessly communicate.

How to choose an automation supplier that will succeed

Automation bots are just one solution:

Several other technologies are beneficial and sometimes essential for a successful automation program, such as artificial intelligence and chatbots. Make sure your supplier has a full working knowledge of the entire digital spectrum.

Use someone that knows your industry:

Every industry will have its own processes that are not replicated in a broader context. Use a supplier that understands insurance to shorten the time to obtain benefit.

Don’t pay huge fees for excessive discovery phases:

Processes should be taken one at a time. Ensure your supplier gets the automation working in minimal time and at minimal cost.

Consider a managed service:

It is certain that any automation deployment will change as the needs of the business changes. A managed service deal will run regular reviews of the automation and alert when updates are required.

Use Case: the back office – cutting time-consuming admin

The back office is the backbone of insurance companies, yet far too much back-office effort is wasted on time-consuming admin that distracts people from more value-adding tasks and heaps costs onto the business. Insurance as practiced in the London Market is an administrative behemoth, requiring mountains of paperwork and numerous manual inputs and checks to ensure the smooth running of a company.

Intelligent automation can improve invoicing process and solve a range of associated issues, such as the unacceptable length of time needed to confirm the paperwork and poor data quality owing to errors in manual user validation.

This can be addressed by using combination of Robotic Process Automation (RPA) and Optical Character Recognition (OCR) / Intelligent Character Recognition (ICR).

RPA can interact directly with a user interface and emulate human interactions, rapidly performing repetitive back-office tasks that would otherwise be carried out by people.

It reduces costs and speeds up processes, as well as improving the reliability and quality of the data being processed.

OCR/ ICR can boost the RPA platform’s ability to automate complex processes involving data supplied in an unstructured and semi-structured way. Invoices are a great example of the kind of variable content that OCR/ICR can tackle successfully, where documents contain broadly the same set of information but in a variety of formats.

Typical benefits can include:

  • Saves time, FTE and money in finance operations
  • Improves data accuracy and quality within invoicing
  • Increases ability to scale and support changing business demand rapidly
  • Improves the audit trail of invoice processing through error logs and SLA tracking

The NTT DATA approach to Intelligent Automation

NTT DATA applies a proven method built on fact, using data and analytics to uncover the reality of a process, not the perception of users. Our aim is to establish business objectives and select the right processes for automation; truly understand and re-engineer the processes involved; to make it more efficient and suitable for automation; and implement a sustainable automation solution.

The method is based on a five-stage approach:


Understand the goal of automation and why it is being set. Work with stakeholders to agree the objectives.


Understand the problem, the environment and collect all the necessary information to make informed decisions about which processes should be re-engineered and automated.


Design the automation and work with users to create the automation model, define governance and describe the expected results.


Implement the agreed actions according to the need and set timescale.


Learn from what was launched and completed in the previous stages to continuously improve, gathering the necessary information.

What is process mining?

NTT DATA applies process mining to understand how a process is being run. This powerful data analytics technique collects operational data in real time from user devices and logs to reveal how a manual process is performed in reality. This helps to create a deep understanding of a process based on fact, not user perception, and opportunities for improvement

NTT DATA has delivered multiple automation projects in the insurance market. We have also developed real-world solutions to demonstrate further possibilities.

The following examples illustrate some of the benefits that can be achieved.

The front office: Bordereaux process

Understand the goal of automation. The pre-renewals Bordereaux process is notorious for its administrative complexity. Reviewing and readying thousands of policies for rollover generates a plethora of time-consuming, low value-adding tasks that distract staff from completing tasks that add greater value to the business. It’s also a process many people dislike because of the mundane and repetitive tasks involved. Time consuming and prone to data quality issues, manually running the Bordereaux process is inefficient and lacks the scalability required to deal with peaks in demand.

NTT DATA helped one insurer improve its front office efficiency by creating an intelligent ‘virtual worker’ to carry out the entire pre-renewals process automatically. The project enhanced and streamlined the existing business process and applied automation to free up valuable personnel.

The project delivered substantial benefits:

  • Helped maintain a steady full time equivalent (FTE) level of staffing during a time of growth, saving the equivalent of more than £57,000 annually
  • Increased flexibility by making it possible to expand the pre-renewal work capacity quickly
  • Improved data quality within the pre-renewal process
Projects in practice - Efficient data transfer improves binding of risks

Direct binding of risks from brokers has cropped up in recent years to handle a greater number of policies in the London Market. The system allows brokers to access coverage for marine cargo risks using an online platform to quote and price the policy. In the conventional set-up, all the documentation had to be manually input into the carrier’s system once the risk was bound, wasting time, resources and effort that could be better directed elsewhere.

NTT DATA created an intelligent automation proof-of-concept solution to automate the process end to end.

First, the system uses RPA to detach the required PDF files from the inbox of the online platform. OCR then reads the data contained within the PDF. The resulting information is entered automatically across 15 screens on the Policy Administration System, so it is clearly available to anyone who needs to review it without disrupting other aspects of staff roles.

The project delivered substantial benefits:

  • Reduced processing times from five minutes to just 20 seconds per policy
  • Increased accuracy of data input to 100%
  • Freed up full-time employees to focus on more value-adding activities
  • Removed the need for complex system integration across multiple platforms
Automation can improve all business areas

Automation can help every player in the London Market – from highly experienced key workers such as underwriters, to back-office administrative operators such as finance staff

The technology, when applied accurately, can alleviate many of the pressures that prevent staff from adding the greatest value to the business and its clients.


Intelligent automation provides an effective way

Market – pressure on costs and pressure to deliver a better customer experience.

Given the challenges of overhauling companies’ legacy IT infrastructure and the cost and competitive pressures looming over the sector, adaptable automation platforms such as RPA and OCR can provide companies with a sharp competitive edge for future growth.

The benefits of automation include: -

Higher flexibility:

Intelligent automation platforms run 24/7 and scale to handle peaks in demand, such as the annual renewals process

Faster processing:

Cuts the time taken for manual processes, such as approving claims, from hours to minutes

Lower cost:

Automation can take on the FTE workload of a typical UK worker for around one third of the cost

Attractive returns:

The typical ROI on an automation project in the London Market is between 30 and 300% per year, depending on the process

Reduced risks:

Automated systems help keep sensitive business and client Wdata and processes in house, lowering the risk of a breach or loss

Rapid implementation:

New systems can be up and running within weeks


Whitepaper - 10 min read


With business norms swept away by the 2020 COVID-19 crisis

Download Whitepaper

Related Insights

How can we help you

Get in touch