Thursday, 4 November 2010

How the law can be used to save millions

For some years I worked with a small but highly innovative software organisation from Australia.  Barristers with a hint of geek about them, they came with the strange notion that law could be used to great effect in supporting the citizen through translation of the law into computer systems.

The company was SoftLaw, formed by pro bono lawyers who spent their days helping people deal with ineffective government departments and with a belief that there had to be a better way.  The better way they envisaged involved translating the archaic language of the law into clear language that could be understood and processed by computers so a client could provide answers to questions posed by the law and receive a judgement based on the answers provided.  The applicable elements of the law were considered those that weren't relevant were excluded to reduce the number of questions to the minimum necessary to get to the right determination.  A simple example, in plain language:

Benefit may be claimed  if
  - the claimant was under the age of 65 at the 1st of January of the current year
  - the claimant has not earned more than 15,000 in the financial year
  - the claimant has no other means of income
  - the claimant is currently unemployed

In the above example there would be a number of questions that could be asked to determine eligibility in an initial data capture, such as...

Name
Date of Birth
Employment status (employed/unemployed)

The translated law would use the provided data against tests to determine whether basic criteria were met before proceeding into a deeper investigation.  If the initial response gave employment status as employed the test would be failed.  The system could then provide a detailed statement of just why this was the case with direct reference to the elements of the legislation that led to the determination.

All seems fairly straightforward.  What was more impressive was that the amount of IT specialist input required to produce systems was minimal but the potential value from developing them was immeasurable.  For example, having a system that could be used to test the effectiveness of the legislation by putting real-world data through draft legislation to determine winners, losers and outliers would be an enormous advantage to the policy makers who want to achieve the best result and cater for extremes in the most appropriate way.

Evidence from Australia at early adopting clients pointed to significant savings from a variety of sources:

  • reduced time effort and cost in reaching determinations of eligibility
  • increased accuracy of determinations
  • reduced number of legal challenges against determinations
  • reduced losses when challenged
  • improved productivity of users - reducing staff costs significantly
  • reduced time to change legislation
  • reduced errors in new legislation
A list of benefits that most software products could only aspire to.  The early adopters were best described as raving fans and with good justification.  However, selling the product was an uphill struggle to say the least and it made little sense in the early days.  Given the advantage of hindsight it is easier to assess the issues:
  • IT people saw threat not opportunity
    • a test production of a system for a UK legal requirement was delivered in one tenth of the time that it had taken the outsourced SI to deliver the same thing using Microsoft technologies (this is fact - I was there)
    • the buyers had just invested in the equivalent of two plastic cups linked by a piece of string, who did these upstarts with the equivalent of a mobile phone think they were?
  • Legal people saw threat not opportunity
    • if the system allowed for demystification of their legal language then perhaps they might no longer seem as special and important as they once were
    • if the software could represent what they did so effectively could none lawyers draft the law in plain English just leaving them with the final stage conversion?
  • Policy people saw threat not opportunity
    • policy is black art, the ability to increase the levels of certainty of result reduces the significance of the problem, policy people in the civil service are traditionally senior - if it isn't so hard to assess and trial alternatives then there is no need for the same levels of seniority
    • if the law was simplified people would know how to deal with it and could turn it to their advantage (citizens aren't supposed to know how to do that apparently)
  • Operational people saw threat not opportunity
    • if a Q&A system can reach a determination in a short time with no human support most of the time then what would operations be about?
    • allowing intelligent self-service would reduce the needs for large armies of advisors, in the civil service having lots of people equates to having lots of power
This all added up to universal resistence other than from a small number of enlightened individuals who could see the potential instead of being consumed by thoughts of doom and gloom.

The Australians worked as well as any I have ever witnessed in my experience, gently and intelligently pressing their case.  Tackling resistance in a sensitive manner requires the patience of a saint and the determination of a mountain climber.  They deserved better results than they achieved.  

Why does this matter?  Because the loser in this situation has not been the software company (they became RuleBurst then took over and adopted the name Haley and eventually sold their product to Oracle where it is now known as Oracle Policy Automation and is buried amongst their other offerings) or the outsourcer or even the client who lacked vision to recognise just what had been placed before them.  No, the loser was the citizen of the UK.  The citizen who just might have seen a better benefit system emerge, or a better and fairer tax system, a cheaper civil service, an engagement with public services in plain English and a Treasury that might have a little more money to spend wisely.

No comments:

Post a Comment