Capture: “Well you are the experts, you tell us!”

So the Imaging RFP is out and there are no fixed parameters, no volumes, no index requirements and no document types listed. You are on the client / vendor call listening to all the questions being asked when you hear …”well you guys are the experts – you tell us!”

It is difficult to do and I have by no means mastered the art of gently pushing back on the client. But, is it wrong to ask the Client to consider providing defined parameters around their imaging project so when comparing vendors they are comparing apples for apples? After all you want to be their trusted partner and you are already helping them out of the gate create a level playing field for themselves to review each vendors solution and pricing equally. Let me expand a little further……

Without defining a set volume, number of document types, index fields required then the resulting proposals will vary drastically:

– Some will low ball it to get the work and potentially not deliver as they have gone too low

– Some will low ball it to win the business but then have numerous Project Change Requests (PCRs) which will drive up the final cost

– Others will be risk averse and make worst case assumptions thus driving up their price

– Others will be unsure about the work or not want it as the project is so open-ended and with their strong brand reputation go in high

I am sure this applies to many projects in the ECM / BPM space but I think with imaging it may be a little easier to raise the question for open-ended projects. Interested to hear other people’s views and experiences.

 

Fraud is on the UP!

What’s the big deal?

Fraud is not a new problem and impacts practically every market. It occurs when an act to deceive others is committed. In the Insurance industry this deception to obtain benefits or advantages that they are not otherwise entitled to can typically be conducted by clients, employees, vendors and even terrorist organizations.

The following is a brief overview of the Insurance fraud landscape.

Insurance fraud is estimated to be the second largest economic crime in America that is exceeded only by tax evasion. Insurance fraud imposes significant financial and personal costs on individuals, businesses, government and society as a whole so it is widespread but also more worryingly it is growing. Insurance swindles victimize people from virtually every race, income, age, education level and region in the U.S. [8] The economic downturn in recent years has meant an increase in fraud of up to 31% with an estimated total insurance fraud costing more than $96.8 billion per year. A conservative estimate of $20 billion of this is for property and casualty insurance fraud. [9]

These significant financial costs make it easy to see why Fraud is one of the top priorities for ‘C-levels’.

 

Pain Points:

Today, there are a number of pain points for fraud investigators in relation to the claims management process. [7]

Pay and chase process:

In today’s Insurance industry, fraudulent activity is generally found after money is paid out which is not a very successful model.

Data overload:

There is a mix of systems across different vendors, agents and brokers which combine both structured and unstructured data. Insurance companies face a considerable challenge to ingest all these data sources and analyze it in a timely manner.

Better Insight:

Traditional methods of data analysis have long been used to detect fraud and require complex and time consuming investigations. Today, investigators demand a better insight into the claims and relating fraud data due to the sheer mass of data and the need to have up to date information to identify new trends. Many retrospective models are used for investigation today and these need to change to be more proactive.

Collaboration:

Collaboration from knowledge workers with domain experience in financial, economics, business practices and law plays a key part in reviewing cases. These key contributors to the process need a way to quickly and efficiently manage, analyze and review all aspects of an allegation which can be made up of many different types of documents such as incidents, repor not, interviews, vendor, quotes..etc..

Fraud Investigation:

There are four (4) key stages to combating fraud:

Detect:

• Fraud Analytics
• Identification of fraudulent activity
• Best practice is to detect at First Notice Of Loss (FNOL)
• Determine Who
• What are the anomalies
• Learn behavioral patters – although this is getting more difficult as fraudsters keeping changing and trying new techniques
• Apply models

Respond:

• Building the circumstantial case
• Collating the evidence
• Prevent fraud

Discover:

• Big Data – Big Analysis
• Retrospective Analysis
• Analyze the evidence – any source, any data, any time
• What else occurred, what did we miss?
• Determine relationships and patterns
• Obtain direct proof

Investigate:

• Forensic Analysis
• Why send to a Fraud Unit / SIU
• Assign Investigator – physical followup
• Direct examination of the subject

 

Opportunities:

Below are details on opportunities that exist today to assist Insurance companies identify and investigate claims in a more efficient and successful manner:

Point of Sale:

Reducing fraud at the point of sale can save insurance companies millions of dollars each year. Below are some opportunities that could be exploited to insure companies are a step ahead of any potential fraudulent claims:

• Continual add to existing data – adding information helps uncover new trends
• Real time detection – access multiple data sources and flexibility to adjust to changing nature of fraud

Big Data:

The ability to access a variety of both structured and unstructured data sources, remove duplicates and quickly and efficiently analyze large volumes of data for key facts, trends and patterns will provide invaluable for Insurance companies to combat fraudulent claims.

• Evaluate fraud in a holistic approach – a combination of alerts across different sources is a warning sign of a policy that is likely to have future losses

Content Analytics:

Intelligent investigation methods in the areas of Knowledge Discovery in Databases (KDD), Data Mining and Machine Learning and Statistics can be used to detect and prevent. Techniques in these methods are generally defined into two categories: statistical techniques and artificial intelligence. [4] There are opportunities to:

• Use advanced statistical analytics – get away from guesswork
• Not solely rely on prior coverage information – repeat offenders provide inaccurate information continually
• Not just focus on high-premium states – repeat offenders take advantage of lower-premium
• Screen all vendors used in a claims process
• Utilize visual relationship graphs to help determine fraud partnerships with individuals, employees, vendors ..etc..

Case Management:

An effective case management tool can enable fraud management teams to quickly and easily manage cases across the lifecycle of the investigation this enabling the members to make quicker and better informed decisions.  A key capability that doesn’t exist today in internal case management is the ability to integrate and collaborate with 3rd party Anti-Fraud industries such as the Coalition Against Insurance Fraud and National Insurance Crime Bureau (NICB) on specific cases and trends.

Finally, the final contents of the case should feedback into the overall Fraud Detection process to allow the Insurance provider to reduce the risk of similar fraudulent acts in the future. This feedback capability should also assist Underwriters to enhance existing underwriting rules to tighten Insurance policies going forward.

 

References:

[1] Cambridge Intelligence: Business Intelligence

http://www.slideshare.net/neo4j/cambridge-intelligenceandkeylines-neofraud

[2] CNSI Whitepaper: Emerging trends in Claims Management and how they can be used for fraud detection

http://www.cns-inc.com/doclibrary/busdevdocs/Emerging%20Technologies_new3.pdf

[3] TrendWatch: New developments about fraud in America

http://www.insurancefraud.org/article.htm?RecID=3269#.UeyTUI3FWew

[4] Wikipedia: Data Analysis in Fraud.

http://en.wikipedia.org/wiki/Data_analysis_techniques_for_fraud_detection

[5] Gartner: Magic Quadrant for eDiscovery Software

http://kcura.com/relativity/Portals/0/Documents/Gartner%202011%20Magic%20Quadrant%20for%20E-Discovery%20Software%20Report.pdf

[6] Pitney Bowes: Portrait Software

http://www.portraitsoftware.com/company

[7] IBM: DeveloperWorks IBM ECM Senior Product Manager interview

http://www.youtube.com/watch?v=c9e3_T4SDy4

[8] Insurance Fraud:

http://www.insurancefraud.org/IFNS-detail.htm?key=16792#.Ue08pI3FWew

[9] Erie Insurance Facts:

http://www.erieinsurance.com/about/insurance-fraud-cost.aspx

[10] National Insurance Crime Bureau:

https://www.nicb.org

[11] Business Insurance

http://www.businessinsurance.org/10-most-common-types-of-insurance-fraud/

[12] MasterCard Worldwide Whitepaper:

http://www.mastercard.com/us/company/en/docs/Modeling_white_paper.pdf

[13] Property Casualty 360 Reports:

http://www.propertycasualty360.com/2011/04/04/fraud-fighting-over-the-next-5-years-the-future-of?t=investigative-forensics&page=2

[14] G.K. Palshikar, The Hidden Truth – Frauds and Their Control: A Critical Application for Business Intelligence, Intelligent Enterprise, vol. 5, no. 9, 28 May 2002, pp. 46–51

The ECM Platform versus Department Needs pitch

So you are all set with your talk track to your next potential client, you have your ECM slides ready to go, some product roadmap details, email archiving configuration details, maybe even some data growth slides to scare them into needing to do something sooner rather than later.

Good to go! Or so you think…but who is your audience?

Great if your audience is the IT department but as budgets become more constrained in companies it is the business who are approving spending. So how do we communicate the right message to a business audience. Many times in the past I have had to present to both an IT and a Business audience which may be made up of Procurement, Finance, Legal and other departments. It is difficult to keep all departments interested during a pitch and certainly our previous slide deck won’t win anyone over other than some technical questions from IT.

Below is a typical slide (unsure of source) showing how we the various components of an ECM platform can come together to address the various business units needs. Again something IT may get excited about but to Joe our Finance Manager and Brendan in Procurement this doesn’t exactly answer their questions or help them with their goals.

So back to understanding out audience. What are the pain points and challenges that various department heads face today or even better what are the ‘hot buttons’ for the C-Levels as ECM platforms aren’t cheap and signing off $100,000’s isn’t a business decision that is taken lightly. Let’s think about who may be at our meeting and the challenges their department faces:

Procurement:

– Contract negotiation is headache with multiple versions of contracts being emailed and red-lined

– Wouldn’t it be nice to know the existing contracts that are up for renewal in the next 60 days so negotiations can start early and avoid the last minute rush

 Engineering:

– Manual retrieval of drawings takes too long

– Can’t find the latest product specifications

 Legal/Audit:

– Legal holds are killing us – more resource and money spent trying to find and produce the right data – if we can find it!

– All this data needs to be kept – where to store it? Keeping backups after backups

Membership Services:

– Escalating overtime costs as of our application backlog

– Enrollment process cycle times are too long

Customer Service:

– Wait times are too long and we are losing customers

– We are unable to find the latest update information – our members get frustrated as we need them to resubmit information

Finance Manager:

– AP department may be missing discounts on early payments

– Even worse maybe the company is late on their payments hence incurring penalties

…and so on, and so on.

To summarize, know your audience and keep it simple and relevant to them. Split the time during a presentation equally between the various departments, representatives that attend. Think more about the department or the vertical being addressed and how that sits on top of the ECM platforms as opposed selling ECM.

 

 

BPMN Tutorials

Check out the following slideshare presentation from IBM on utilizing the Business Process Management Notation (BPMN)

Really useful for Business Analysts, Process Engineers, Workflow Designers and Architects to use this standard especially if there is a large corporate project with many processes involved. The great thing from an IBM and other leading workflow vendor products is that if using the BPMN template for MS Vision you can import your process design directly into the system eg, Process Designer. I have varying success with this in the past, mainly with more basic workflow.

bpmn tutorial from uhuru1973

Conference Call

I am sure a lot of people can relate to this!

Uncie James

On 27th September 2013 I became an Uncle. My little sis, Leigh and her husband Paddy had a beautiful little baby girl named Ruby Rose Keating.

Very excited to eventually see my niece at Christmas when back in Ireland. She obviously, gets her good looks from her Uncie 😉

Pricing the Unknown

• How much will it cost to migrate from our legacy system to a cloud solution?
• How much will it cost to outsource all our current capture operations?
• How much will it cost to transition our capture services to your imaging center and utilize OCR?
• How much will it cost to consolidate our various content repositories to one ECM platform?
• How much will it cost to seamlessly integrate your workflow into our ERP?
• How much will it cost to upgrade our BPM system?

I always sigh when posed with these types of questions and unfortunately it is something in the consultancy world that you are frequently presented with. I like and dislike these questions. In some respects it means our pitch, our services or our solution interests the client and there is opportunity to further the sales process. At the same time it is so difficult to put an accurate figure on such high level questions especially when you don’t know the client’s systems, technology, operations and integration points. Generally, these questions come from CFOs, CIOs, Strategic Sourcing, Procurement or VPs of Operations who are all seeking to make cost savings. The reality being that there isn’t one person in the company who has the complete understanding of the current state and all the costs associated so how can an outsider provide a figure?

Generally, there is a three (3) step approach to tackling this type of high-level pricing queries.

1.) Discovery Services
I was like to lead with suggesting a deep dive analysis which can be a number of days or even weeks depending on the company size, number of sites, complexity and system integration. Sometimes client’s understand this and willing to pay several thousand dollars to perform such an analysis as they understand they don’t know everything so how can we ask someone else to price it. They also realize that they get a more accurate figure as opposed to a high-level guestimate.
However, it is often the case that the client doesn’t want to make that initial investment as they are just seeking to understand if it something they want or can afford to do. They may never progress the work so why pay to investigate it. At this stage I get some what wary but often try to encourage the discovery services based on providing an accurate project plan, tasks, schedule and pricing. It also helps to ‘sweeten’ this idea by suggesting to discount the discovery services cost from the implementation cost if they decide to progress.

2.) ROI Model
What happens if the Discovery Services costs aren’t favorable to the client? As a next step I turn to an ROI model. If the client can supply what it costs to run the current services and we have an idea on how much it will cost to run the future services then an ROI model can be developed.
At this point there is some onus on the client to provide figures and if they are really keen to investigate and potential engage your services then they will take the time to compile the necessary figures. Therefore if we have the current and we know what the potential future figure may look like then we can estimate the potential monthly savings. We still don’t know the total transition cost but because we have before and after costs we can make assumptions on the transition costs and run various iterations to determine the payback timings.
Eg,

• Current = $30K / month
• Future = $20K / month
• Savings = 10K / month
• Transition scenarios = $5K , $10K, $20K

ROI Models:

• $5K Transition scenario

Month 1 = $25K Savings $5K (Immediate payback)

• $10K Transition scenario

Month 1 = $30K Savings $0K
Month 2 = $20K Savings $10K (2 month payback)

• $20K Transition scenario

Month 1 = $40K Savings -$10K
Month 2 = $20K Savings $0K
Month 3 = 20K Savings $10K (3 month payback)

A simple example but you get the concept.

3.) Pricing Estimate
Inevitably, there are clients that insist on a pricing estimate for the transition, migration, implementation or upgrade that won’t see the value in the discovery services and won’t have time to collaborate on an ROI model.
Therefore if the opportunity has been qualified and is worth it we begin a high-level estimate exercise making a lot of high-level assumptions. This in itself can be an art and I believe those with a project management experience in the similar engagements can assist in breaking down all the high-level task into a low-level and begin to put an estimated number of man days beside each task.
Bucketing the overall project into phases is the first step. Eg, Implementation/setup platform
Then the transition work should be broke down into those low hanging fruits followed by rollout in manageable chunks. Each of these buckets should then have the typical deployment life cycle applied to them: Data Gather, Solutioning, Implementation, Testing, Go-Live.
This framework gives you the ability to define the lower level tasks and place estimates beside them. This should be a group effort including the relevant SMEs for implementation, migrations, imaging, testing ..etc.. This will help get a more accurate final figure. Once all is said and done and we have our estimate I usual safeguard this with a +/- 30% variance.

Summary:
We have covered an overall approach to pricing the unknown and then if forced into a figure how this maybe broken down by man days into the various phases and steps to get an estimate.
People or Sales become very focused on the figure which isn’t a bad thing as they should feed into discussion despite being low level they know the client better than anyone and hence should help guide the figure from their coaching or from the expectations shared by the client. The key thing for me is that at these early stages we are just trying to further the sales process so the figure we provide is only an estimate (big disclaimer required). In many instances the client is just seeking to understand if this work is $2 million or $20 million and when they will get their return. By socializing the Budgetary Planning Estimate (BPE) with the client further and by reducing a day rate or increasing the estimated variance we can generally achieve a figure that allows the sales process to progress to the next step and hopefully a successful contract execution.

To code or not to code….that is the question!

In many Enterprise projects there is often an immediate need to start developing or coding a solution to a problem to meet deadlines.  This approach can be spurred by AGILE development methods and working nimbly and dynamically to execute small focused packages of work. Many developers further fuel this situation by wanting to learn the latest coding techniques or work with the latest API libraries and jump straight in to explore new coding capabilities.

From an Enterprise Architecture (EA) perspective there can be 3 main challenges with the foregoing:

1. EA principles are geared towards being able to facilitate inevitable changes with minimum disruption and effort and creating new code bases isn’t always the path to take
2. AGILE development has obviously a great role to play in development based projects and has had many proven successes. However, is it suited to all types of development? In many large scale projects involving the integration of many systems does the AGILE methodology lend itself to managing the complex business rules, integration layers and schemas/contracts between systems? Or does the traditional, well thought out and planned Software Development Life Cycle (SDLC) lend itself better to these Enterprise backend workstreams and leave AGILE for front end user applications?
3. Shouldn’t some form of due diligence be performed in the organization to determine what is already in use today or what market solutions can potentially solve our problem?

The Reuse, Buy, Build concept is an oldie but it is sometimes good to be reminded of such approaches before tackling a new project and jumping in with the latest a greatest development tools and techniques.

1. Resuse – obviously this can take place at different levels. Maybe it is a complete and existing imaging solution that simply needs extended to another department or potentially just a subcomponent such as the OCR functionality is required to achieve a desired classification. A great tool for this is the Application Portfolio Management (APM) which takes time and effort to create but if maintained and kept uptodate then no EA division should be without it to help make strategic decisions
2. Buy – What does the market offer today and also tomorrow? Has the due-diligence been performed to ensure a software package, a solution or service isn’t available today? By being able to implement something off the shelf, great time and cost savings can be achieved. Even if only 80% of the functionality or requirements are met AND can be added to via additional coding or configuration then this must be better than starting from scratch? Price obviously has a part to play and ofcourse if there isn’t a clear winner then a cost/benefits analysis should be performed to support the final decision made
3. Build – So we can’t resuse anything and there is nothing feasible on the market that meets our specific requirements then we must resort to developing a solution to our problem. Whether this be utilizing our internal resources or a third party agencies we want to ensure that the proper platform is selected, future development or expansion of services are considered. To take our Reuse, Buy and Build approach further we may even consider Re-Building something that exists today before starting with a blank canvas or at the very least listing those existing components, libraries and routines that can be utilized in our new development project

A common assessment criteria which many will have come across in deciding on whether to Reuse, Buy or Build is based around the following categories:

More can be learnt from a great article on the Architecture Journal: http://msdn.microsoft.com/en-us/architecture/aa699448.aspx

As companies and projects demand resources to do more with less then hopefully the foregoing acts as a short reminder of some of the basic EA concepts before reinventing the wheel with your next development project. One final thought if you do end up developing your next solution then understand it is your time to create something innovative and potentially path the way for future development projects. So think disruptive – something new, something innovative, make use of those new coding standards, tools and techniques to create some revolutionary that forms the strategy for your company’s or department’s solutions and services.

GoPro: Vegas and Grand Canyon by helicopter!

My first outing with my new GoPro Hero 3 was Whale watching in Alaska with my girlfriend. Unfortunately, things didn’t go to plan and the best footage of the whales breaching right beside our boat didn’t get recorded! Ooops – that didn’t go down well with my girlfriend. Lesson learnt: don’t use your GoPro for the first time when on a once in a lifetime excursion!

Anyway, second test run with the GoPro was a helicopter tour over the Grand Canyon and Vegas. What an amazing trip and this time I managed to capture it. Well done me. Some lessons learnt this time are:

– Don’t point the camera directly into the sunlight when moving at speed as you get large gray blinds occurring on your video

– Keep the camera as steady as possible – using both hands is recommended

– Any great landscapes or views you want to capture then take you time to pan round them slowly and even hold at certain points

I also used the GoPro Editing Suite for the first time which is pretty intuitive even for me. Adding title slides, music, trimming sections of film, merging, fading in and out as well as applying filters are all straight forward tasks. Maybe with time I can find out how to do some of the fancier slow motion, freeze panes, timelapse and perspective editing and blog about it. Make sure you have Quicktime installed as it is a prerequisite for certain functionality in the editing suite.

Anyway, I would really recommend  the helicopter ride. We managed to get a deal with Sundance helicopters and did the touch down tour in the Grand Canyon which started out late afternoon and returned to Vegas just as it was getting dark and Vegas was lighting up. Awesome. The trip was expensive but if you are thinking about it look out for Groupon deals before you head out as a bargain can be had!

We offer 3 kinds of service…

I liked this image when I saw it. It makes sense when working with demanding clients that want to haggle on price and enforce super tight deadlines!