While a growing group of larger insurers are applying ACORD XMLfor internal and external purposes, some industry insiders believeit'll take an 800-pound gorilla to make standards a standardpractice in insurance IT.

by Robert Regis Hyle

Internal communications can be a good thing or not, depending onwhos doing the talking. For example, the prevailing opinion aboutpeople who talk to themselves is that they have a fewproblemsparticularly if someone answers. But in the world ofinsurance, internal communications have a different meaning. Withcarriers collecting data from multiple sources, internal users haveto be able to communicate with co-workers in a language everyonecan understand. The same holds true for communication with externalcustomers and partners, although insurers have found the answer tothat problem is a tad more complicated.

That is why ACORD XML has become a growing part of the insurancelandscape. Actually, to this point, the ACORD XML standard is theonly one that exists for property/casualty insurance in the UnitedStates, says Rich Maynard, director, personal lines applicationarchitecture services for The Hartford. Standards are evolving,though, and in the slow-moving insurance arena, the evolution isgoing to take some time. It took a decade for the banking industrysIXF data standards to reach maturity, Maynard points out, and theACORD XML standards are less than four years old. Mele Lynn Fuller,interface architect with Safeco, agrees the insurance industry hasto be patient with the direction standards are headed. You have togive [carriers] time and give some of the smaller players time tocatch up, she says. Thats a mixture of money and resources.

Bruce Wilkinson, CIO of Citizens Property Insurance Corp., withmore than $1 billion in premium, says there has been an awful lotof talk about standards, but the number of people discussing themis large vs. the number of insurers actually receiving results fromthem. Were focused here almost fanatically toward the businessresult, says Wilkinson. Technology for the sake of technology isinteresting, but if it doesnt produce something you can see on abalance sheet or on an income statement, then we call it researchand development.

Balance Tipping

Larger insurers are showing an increasing amount of interest inthe ACORD XML standards, according to Matt Josefowicz, analyst forCelent Communications. A recent survey by the research firmdisclosed 65 percent of the top 20 property/casualty insurancecarriers in the U.S. have implemented at least one ACORD XMLproject. The industry moves slowly in general, he says. At thispoint, the question is why not use standards as opposed to why usethem.
When insurers discuss ACORD XML, the dialogue usually relates tohow the standards can be used with external partners, but thestandards also have shown value internally for insurance carriers.Standards have been positioned historically as about communicatingbetween different partners in the value chain, says Josefowicz.When you have an external communication standard, it is hard to seewhy you should adopt it if all your trading partners havent done soyet. Its a classic network problem: The networks not valuable untilyou have lots of participants.

At the same time, however, Josefowicz believes there has been ashift within the last year, particularly at larger companies, tomake the investment in ACORD XML as a means of exchanginginformation among internal systems as well as reconciling thedifferent data standards among the various siloed parts of a singleenterprise. Companies know they will get value from investing inthe standards internally because everyone in the enterprise will beusing them. It can solve a lot of discrepancies between differentdata standards historically used by different pieces of theenterprise, he maintains. In those cases, you dont have to wait foryour data partners to come up to speed. You can start getting thebenefits immediately in terms of uniting data and creating accessto data from within different parts of the enterprise. This alsopaves the way for the carrier to be a step ahead of others whenadoption eventually hits industrywide.
The Hartford is using ACORD XML in its internal messaging suite,Maynard reports. Rather than building it all ourselves, we felt wemight as well leverage an emerging standard thats out there andutilize it for all our internal message structures, he says.

While XML may have been initially developed to communicateexternally, Safeco found it worked particularly well as the carriermoved into the server world from its IBM mainframe environment. XMLwas an excellent way to share data and format data in the PCenvironment, says Fuller. XML-based data is self-describing. I dontthink you can overestimate how much of an advantage there is toself-described data. If you hand somebody an X12 data file or afile we pull off the mainframe, you almost need a description papernext to you to define what the data means. There is no way, shortof having memorized it over years, to understand what this datameans. I can hand anybody who reads English an XML file, and thatparty can read it. It helps in design work. It helps tremendouslyin trouble-shooting any kind of problem. You dont need a programmerto read the file. A brand-new business analyst can read data [inXML].

Better and Better

Fuller was involved with the initial creation of the ACORD XMLstandards and is proud of the work involved. We did a good job onVersion 1.0, but once we got a year and a half into it, we had 600to 700 messages, she says. Volunteers from the ACORD membershipcurrently are improving the architecture to cut that number to aquarter of what they had before, she says.

One of the development teams goals is to make sure the businessdata is organized in such a way users can change technologiesreadily. This is easier in theory than practice, because therealways is a relationship between the technology and how you formyour data in your messages, Fuller says. Were trying to keep it asopen as we can so if schemas arent the right technology in fiveyears, we can roll the data standards into the next technology.

The Hartford worries about the new versions of the standard thatkeep coming, Maynard says. Were wrestling how we are going tocontrolsetting it as an internal standardthe proliferation of [thestandard] across all the applications and make sure we are using astandard implementation of it, as well.

ACORD is working on the release of Version 2.0, and that willbreak with tradition from earlier releases that had backwardcompatibility, according to Fuller. We assumed when we firstdeveloped the ACORD XML standard we probably would have to breakbackward compatibility every one-and-a-half to two years because ofnew requirements and increased knowledge, says Fuller. New datarequirements often are a reason to break compatibility. We hadhoped originally not to make quite so many changes, but we findsome fairly significant architectural changes will greatly improvethe ability to implement the standard, she says.

Finding a Reason

The new version might spur usage, Fuller asserts. Carrierslooking to do business internationally, she adds, will have aninterest as they seek ways to communicate with partners theyvenever communicated with before. As we streamline the standard andmake it more logical, we open ourselves up to being much moreattractive to implementation by new partners, she says.

Many companies will stay with one of the earlier visions,though. Its going to take an 800-pound gorilla to move many of usto Version 2.0, says Fuller. And who could that 800-pound gorillabe? A major business partner, she says, citing Marsh as an examplefor carriers or St. Paul Travelers as an example for agencies andvendors.

When a major business partner implements a standard, we aremotivated to make a concurrent move. It makes good business sense,says Fuller. It doesnt make any sense to move from Version 1 toVersion 2 until you have to. Carriers will continue to operate onVersion 1 even if they implement Version 2 because not all partnerswill be compatible with the newer version at the same time. Younever turn off the old standard, says Fuller. I can see in thefuture where an insurance carrier will be transmitting certainlines of business and certain transactions with Version 1 and theothers with Version 2.

Easy Does It

The overriding aspect of dealings between carriers and agents isease of doing business. Carriers that utilize independent agentsknow they cannot demand much from agencies unless the carriersthemselves are willing to pay for upgrades. Fuller points out suchan event would be a rarity. Even as big a carrier as The Hartfordknows it cant force anything down an agencys throat. Were at themercy of our partners, says Maynard. Most of The Hartfords agencypartners utilize one of the two major agency management systemvendorsApplied or AMS. I dont believe its any secret, [Applied andAMS] would prefer we use their standard as opposed to the ACORDstandard, he says. Applieds connection with IVANS andTransformation Station produces a form of the ACORD XML messagecarriers can use, but the fact the two sides have been unable toagree on a specific standard doesnt eliminate the need the agenciesand the carriers have for each other.

One of the things we need to do is make sure we are as easy todo business with as possible, says Maynard. If that means theApplieds and the AMSes are not going to be sending us ACORD XML,but rather their own flavor, and we standardize on ACORD XMLinside, then we need to figure out a way to translate [data]seamlessly from what [agency management systems] are sending us toour version of the ACORD XML standard.

Maynard knows smaller carriers and smaller independent agenciesare unlikely to send data to the other side in any type of standardformat. With smaller independents, were just happy when they sendus a fax, jokes Maynard. Some smaller agency management systemsproduce a flavor of XML, according to Maynard, but The Hartfordstill has to translate what is sent into a flavor the carrier candigest. That can be frustrating, but carriers such as The Hartfordunderstand it all will take time. The adoption of the standardacross the industry is not at a level where I think we would likeit to be, but with any standard thats only a couple of years old,what do you expect? asks Maynard.

He says he often is asked why insurers cant do acquisitions likebanks do. One of the reasons is the banks have the IFX standard andeveryone communicates that standard the same way, so its a loteasier for them to perform transactions, he says. In the case ofproperty/casualty insurance, the ACORD XML standard is emerging,but there are lots of interpretations left up to the individualinsurance company. So when we put our proprietary twist on it, youbegin to dilute the capability of the companies to have a fluidtransaction set going among them all. That [IFX] standard may be 10or 11 years mature. ACORD XML is only a couple of years mature.When were sitting here eight years from now, we may say ACORDstandard is the best thing since sliced bread and allowed us tostandardize the property/casualty industry, but I think were atleast five years away from that.

Wilkinson says his company doesnt tip its scales toward onevendor simply because it offers data in an XML format. Most of thetime, the selected vendor we choose is delivering us the bestbusiness result, he says. If we had three vendors delivering us theexact same business result and one of those used an XML interfaceand the others used some kind of proprietary approach, that wouldgive the [XML] vendor the advantage, but we just havent found that[scenario] to be the case yet. We look at the business result firstand then what its going to take to support it.

Survival or Strategy

Kimberly Harris, research director for Gartner, believesinsurers have to classify their technology strategies between whatthey have to do to survive and the tools that are considered morestrategic and are just nice to have. ACORD XML has been consideredmore strategic, she asserts, but it is becoming more mainstream. Alot of companies that werent looking at it before are using it oradopting it in different forms, she says. Its almost reachingmainstream mass adoption.

Josefowicz adds the trend in insurance is to move toward usingservice-oriented architecture and internal Web services, and thesedepend on XML. So as long as youre moving to XML that way, youmight as well use standards-based XML, he says, as opposed toinvesting in and creating your own data standards, which arentgoing to be leverageable when you talk about moving thosecommunications beyond your firewall.
There is value in using ACORD XML internally, he says, and it iseasier for carriers to justify the investment internally. You canget the return without depending on your trading partners to beenhancing the standards, he points out.

Smaller carriers likely will continue with a wait-and-seeposture toward ACORD XML, Josefowicz notes. With fewer large-scaleintegration issues, smaller insurers are less likely to invest inservice-oriented architecture. [Smaller insurers] dont have quitethe systems integration problems, typically, because they dont havelarge or diverse architecture in most cases, he says.
Harris believes ACORD still has a long way to go in rounding outits standards. Clearly the end-to-end insurance process is notfulfilled yet, she says. There are gaps and certain areas that aremore complete in covering every functionality. More vendors need tobecome certified, not just say they are ACORD compliant, which isthe buzzword now but really doesnt mean a lot. The certification tobe able to support the transaction is really whats important.Companies that are using the standards need to start using them onwide-scale projects so they can use them for both internal andexternal projects.

Wilkinson still is not sold the industry will come around to anysingle standard. In this industry, you have to realize time kills alot of technology, he says. By the time XML is inside the insuranceindustry and reaches a peak of acceptance so that it becomes acommodity, something else may come along that blows that [XML]technology away.