There can't be many industries where the back office is more
traditionally paper-based than in the insurance business. So if any
sector is ripe for workflow automation this is it. Julia Vowler
reports
Could NatWest Life, the life and pensions business of NatWest
bank, justify spending £6m on imaging and workflow technologies?
The answer was yes, but only if the productivity gains achieved
were significantly greater than those usually credited to the
introduction of this type of technology.
That meant setting an extremely high target. Instead of being
content with the standard 20%-30% productivity gain, Project
Cormorant had to enable clerical end-users to do half as much work
again as they used to. In insurance industry terminology, the ratio
of staff to policies had to increase from 1:3,500 to 1:5,262, and
the time spent executing each individual business process had to be
reduced by a third.
"There was no way we'd have done the project if we hadn't aimed
to get a 50% improvement," says Nick Thomas, head of product
services at NatWest Life & Regulated Sales, and project manager
of Project Cormorant.
For all that, it was still a difficult business case to
make.
"The financial case was very marginal, and we set the payback
period for three to four years. If we'd said we wanted payback
within say 12 months we'd not have gone forward," says Thomas.
What complicated the return on investment equation was a factor
that nearly every organisation encounters when making
technology-led investments - the bigger picture. For NatWest Life,
the bigger picture was framed by the fact that the bank's life
assurance business was set up in 1993, but although it was a
greenfield business, it was not an IT greenfield site.
To get to market fast, the company took existing life assurance
IT systems from another company. This meant that it could start
selling policies fast, but the trade-off also meant it had bought
itself an instant legacy system.
"There's been various attempts to migrate us away since then,"
says Thomas, "but to replace all the back office systems would mean
looking at spending over £20m."
Thus, implementing a workflow system to go between the end-users
and the underlying back-office systems was seen as a means of
strategically starting to unhitch the business from its core
processing system, prior to its eventual replacement, when that
makes financial sense.
"The workflow project was seen as an evolutionary step to
migrate our systems forward," says Thomas.
As well as being part of a long-term technology strategy,
Project Cormorant also, he says, had a significant infrastructure
investment content - about £1m-£1.2m of the £6m budget would go on
implementing NT desktops and servers. Justifying infrastructure
spend is always a problem, and the return-on-investment equation is
always uncertain and debatable. NatWest Life used two
justifications, both familiar to those who have to make similar
decisions. One was that an enterprise-wide NT platform represented
prudent future-proofing, by upgrading to a flexible, higher-powered
industry standard. The second was that the cost of the
infrastructure spend was simply to be attached to the business
benefits accrued from the workflow functionality.
The principle is, says Thomas, "the first one on the bus, buys
the bus".
With a very tight business case and a non-trivial infrastructure
burden to carry, the Project Cormorant team were under extremely
visible pressure to do well. Although the products and partners -
Eastman Kodak's Open Workflow, Open Image system, plus Unisys as
systems developer and integrator - had been selected in the second
half of 1996, and the requirements definition completed in the
first quarter of 1997, the "go/no-go" business case review starting
in the spring meant that the project finally kicked off at the end
of June 1997.
Timetable
The design phase, divided into two, business and technical, took
five months until November that year, although build got going in
October and was complete by March 1998. Testing ran from March to
June, and then a phased implementation began, with the first phase
complete by the end of June, and the last complete by August. "Over
six weeks we rolled it out to between 120-130 users, team by team,"
recalls Thomas.
The system was designed to focus on the business processes
within the company's service area (dealing with existing policies
and customers, as opposed to the new business division) which would
yield the most benefit from being automated.
The 60-plus processes were divided into three categories. Those
receiving the highest level of workflow were the high volume ones
that could be most automated. The next level were the medium volume
ones (the bulk of the processes), and the lowest was the collection
of processes that were difficult to define, seldom used, and
usually required human intelligence to make a decision on how to
deal with them.
The top two categories were given full scripts and had lots of
parameters. They had only one way of doing the work. The last was
less automated, leaving room for human decisions.
When rolled out, users got the full functionality of the system
in one go.
"Everything had to be there from day one," he says. "It was big
bang in terms of functionality, but with a slow, phased turn
on."
Right from the start of Project Cormorant there was very
deliberate, highly orchestrated integration between business and
IT. The project sponsor was head of customer services, and Thomas
as project manager was himself a senior business manager.
Several key business users were seconded to the project for its
duration and were actively involved in requirements gathering,
conceptual and detailed design, system and acceptance testing and
eventual training, roll out and post-implementation support.
NatWest Life took care to ensure that 'business involvement' did
not just mean involvement at the level of managers. The system's
end-users, whose way of working would be significantly changed by
implementing workflow, were closely involved too.
"We identified one person in each end-user team of around 25
staff to be the team's representative," says Thomas.
Each team rep was then "on attachment" to Project Cormorant and
was trained up on the new systems, heavily involved in system and
user acceptance testing, and in drawing up the intensive three-day
training the rest of their teams would go through.
Inevitably, this mean a degree of dissociation and diversion
from their day jobs, but this had been allowed for up front in the
project case, both at manager and end-user level.
Having an integrated project team - business and IT - did,
however, throw up some issues, acknowledges Thomas. One was the
familiar one that the very people you want to help on the project,
he says, are the ones that are best at their jobs - the consequent
"tug of love" can be tension inducing. Another, ironically, was
that it was the IT contingent in the team who ran the danger of
feeling dissociated from their own IT environment.
"The IT group attached to us lost their home base and their
roots," says Thomas.
He had to assure them that once the system was running, it would
be business as usual for them back in the IT department.
NatWest Life also ensured that any lack of mental buy-in to the
project by operations managers whose business processes were being
workflowed was avoided by changing their budgets in line with the
project's expected effects on productivity.
"From the moment we said we'd do the project, all the
operational budgets for the areas to be impacted were changed in
line with the benefits," says Thomas.
It was therefore up to the operational managers to ensure that
co-operation with the new system enabled the benefits to be
harvested, or else, says Thomas, "they'd be looking at cost
overruns on their own budgets."
There was also a significant degree of change management
involved. The way that end-users worked would be changed and they
had to be prepared for that, rather than discovering it once the
system had been rolled out.
"There would be a very different way of working - for example,
far more sitting down rather than wandering around manually passing
on files," points out Thomas.
End-user involvement
The outcome of this comprehensive attention to end-user
preparation and involvement was that the company not only did not
see the expected deterioration in productivity during the
implementation phase, but that the productivity improvements
started to come in significantly faster than had been envisaged.
Instead of there being more than 12,000 work items outstanding
within two months of the end of the roll-out, there were only 2,500
- 21% of the projected total.
"We'd expected it to take until six months after implementation
to get back to the level playing field - pre-project work
throughput - but we experienced nothing like that," says Thomas.
"Within about a month of rolling the system out we were level
again, and had achieved our [productivity improvement] targets by
November."
Since then, the workflow system has been continuously improved
to implement corresponding improvements in the design of the
underlying process, in order to make them ever more efficient and
faster to execute. There has also been a follow-on in terms of
change management, in order to defuse fears of deskilling amongst
the end-user workforce now that so much of their job has been
computerised, with the system pulling through the processes
automatically.
"We did anticipate this reaction but we've been surprised at how
quickly it developed," says Thomas.
In response, jobs have been re-skilled.
"We've started to merge different teams to make staff
responsible for more processes," says Thomas. "Their work has
become shallower, but broader."
With the image and workflow system now securely bedded down in
the service section for processing the documentation of existing
customers, "The next logical step," acknowledges Thomas, "would be
to roll it out to the new business area and all the administrative
functions such as sales."
Ironically, the new business area already has workflow, but an
older pre-image implementation, which it is not cost-justifiable to
replace yet. However, staff in this area can access imaged
documents of existing customers where necessary, even though they
cannot use the full workflow facilities.
However, whilst waiting for it to be worthwhile to replace the
older workflow system, the company has gained further benefit from
the new system by selling the intellectual property rights to
Unisys, which has implemented an adapted version at another
division of NatWest bank. This has given NatWest Life additional
return on the considerable amount of bespoke code written jointly
by its in-house developers and those at Unisys.
The final benefit is perhaps less tangible, but no less
valuable. Part of the integration contract with Unisys, says
Thomas, was that NatWest Life wanted a good representation of its
own IT people on the Unisys team. The result is that NatWest Life
now has its own experienced experts at designing, building,
integrating and implementing an image and workflow system.
"We wanted the knowledge," says Thomas.
At a glance
The company: NatWest Life & Regulated Sales
The challenge: How does an insurance company introduce a
new system that will deliver a step change in productivity (target
50% improvement) in less than two months, without alienating its
staff?
The solution: By bringing in an image and workflow system
and integrating it with existing systems with the close
co-operation of both technology partners and key business users,
working to a detailed change and benefit management plan
The Computer Weekly/Buy IT case studies offer an in-depth
analysis of a successful IT project, with expert comment from a
panel. BuyIT was launched in 1995 by the DTI and an alliance of top
industry bodies. BuyIT has selected best practice examples on a
range of projects. Each case study is scrutinised by the BuyIT team
of experts who make their recommendations and comments. The BuyIT
Computer Weekly Best Practice Series is endorsed by Fit for the
Future, a CBI-led, government-backed campaign to get business
learning from business.
Key benefits of the project
- Clerical productivity is up by 50%.
- The infrastructure has been upgraded to NT
- Insulation from legacy core systems
- A ready-made image and workflow system ready to roll out to
other areas of the business
- Revenue from selling the rights to the bespoke software to its
system implementor
- A skilled and experienced image and workflow in-house team with
knowledge transferred from the systems implementor
The technology
Hardware: Dell desktops and Compaq servers plus Eastman
scanners
Software: Microsoft NT client and server, running SQL
Server 6.5; Eastman's Open Image, Open Workflow; bespoke Visual
Basic front ends to provide the infrastructure development and
customised business change aspects, such as "robots" to pass work
along queues
Imaging hardware and workflow software: Eastman Kodak
Systems integrator: Unisys
Project management: NatWest Life
The workflow system
The new system takes in documents from several sources,
including paper letters from customers, faxes and printed reports,
which are scanned in batches, indexed and prioritised. The system
then routes work to the appropriate team, each member of which
processes the item, before the system passes it on, and pushes the
next work item to the clerk, on a date basis unless there is an
alternative priority - complaints are treated more urgently, for
example. The system is also linked to the customer call centre,
where operators can generate a telephone note which is passed to
the workflow system for actioning as a work item.
Critical success factors
Why did this project succeed? Peter Duschinsky, secretary of the
BuyIT Best Practice Group, lists his top 10 critical success
factors:
- A well-documented business case with a clear focus on delivery
within budget and clear and unambiguous goals set to define and
measure benefits
- A very integrated, goal-oriented team focus between the
in-house staff, Unisys and Eastman with successful skills transfer
from Unisys
- Strong project sponsorship by the executive steering group and
good business buy-in
- Good communication throughout - early user involvement and key
stakeholders identified and well-briefed in advance of
meetings
- Operations managers financially committed to realising the
productivity benefits from the new system
- A well-managed roll-out. End-users knew what they were going to
get before they got it. There were few surprises when the system
arrived, and each user team had its own staff expert
- Good post-implementation support provision to resolve any
problems
- Change management - issues were addressed early on and included
local plans to control the impact of implementation upon the
ongoing business operations - a degree of disruption to the day job
was envisaged and allowed for
- Training was intensive and comprehensive
- Strict adherence to formal approval and review processes
(including contract management reviews with the supplier)
throughout the project
Coming soon
An important part of the rationale for doing business over the
Internet is to cut costs and over-heads and pass these savings on
to your customers. When you have over a quarter of a million
customers these savings could be considerable. But how to achieve
them? The solution: Sutton & East Surrey Water, with 260,000
(mainly domestic) customers, is looking to the Internet as the way
to allow its customers to access their accounts and meter reading
information, change payment options and pay their water bills,
potentially saving an estimated £1 per bill in postage and
stationery costs.
What the Buy IT experts say
Alistair Fulton
Chairman, BuyIT Best Practice Group
We selected this project from a number of workflow
implementation case studies because we particularly liked the
excellent way in which it was managed.
It is quite an achievement to be able to introduce and bed down
new systems and processes to over 120 users in less than two
months. And look at the results - within six months NatWest Life
achieved step changes in productivity, major gains in process
consistency and much lower running costs.
The Success Factors panel indicates some of the reasons for this
success. NatWest Life's own post-implementation report shows up
others. Like all projects this one had its problems, clearly
identified in the post-implementation report, such as:
- Sign-off of the business case took longer than
planned
- Incompatible standards resulted in some confusion
- Detailed business requirements were (inevitably) too vague and
led to a lack of clarity in later phases of the project.
But Nick Thomas was able to resolve these problems and learn
from them. He should be congratulated on a job well done.
Colin Thompson
Deputy chief executive, British Computer Society
NatWest Life won a well-deserved place as a finalist in the 1999
British Computer Society Management Awards for this project. Our
assessors found it an exemplary implementation of workflow and
imaging technology and gave high marks for each of the three
aspects of the assessment criteria - business impact, relationship
with users and management of the development.
This was without doubt a challenging project. It was the largest
in the company's history and it employed technology new to the
business. Yet this would serve as a model for almost any aspect of
project management and change control.
The fact that the project was so clearly business-led was a
major factor in its success. The early integration of business and
IS teams, the focus on the preparation for change and the rigour
and care with which a difficult business case was established and
monitored, all gave confidence to business managers and users. In
addition, the detailed analysis of business process provided a very
clear picture of priorities, both for the initial development and
for future enhancements.
One other important ingredient was noted by our assessors - a
total dedication to excellence at all stages, coupled with an
impressive honesty in identifying, and learning from, any
shortfalls.
Daniel Dresner
Head of standards, NCC
Workflow applications are very much the systems that pump the
business blood around the organisational veins and arteries. It is
critical that suppliers are wholly in tune with the organisation's
body rhythms, otherwise the new organ that they are supplying will
be rejected with catastrophic results. This case study is exemplary
in the way that NatWest Life:
- Ensured that difficult targets were made achievable by give and
take across the operation
- Allowed for dips in productivity while the new system was
grafted onto existing practices. This meant that it was able to
come up to speed and accelerate in a planned and synchronised way
to achieve business benefits
- Integrated their approach to technology and operational
procedures
- Realistically managed expectations.
Recording problems for resolution and successes for repetition
is a key practice that will engender a partnership between
suppliers and their customers. Such teamwork is to be
commended.
NCC is currently involved with a number of these aspects within
the standards arena - see the National Computer Centre's Web site
at www.ncc.co.uk/standards/
Steve Goodwin
Managing director of The Houndscroft Partnership and
vice-chairman of IEE's infomatics division
Any project which has a relatively short timescale - especially
a major one like this - will need very careful planning for all of
its stages.
There are many examples of good practice in this case, and Peter
Duschinsky has already listed his top 10 on p23 - so let me bring
out just a few others that, in my experience can make the
difference between success and failure:
- Vital to most projects is a close level of co-operation between
business and IT and, in this example, it appears to have worked
very successfully with both sides working well together at all
levels
- A second important concern in projects of this size is the
effects they have on people, both during and after implementation.
In this case, there has been involvement not just at management
levels but across all the staff - a necessity if you want to get
everyone on board and supporting the project
- Similarly, the concerns of those assigned to the project about
what would happen to them afterwards were considered, as were the
concerns of end-user staff about de-skilling
- Finally, the post-implementation phase has been reviewed well,
so that lessons can be learnt for next time.