Introduction
Below I explore DealBuilder, QShift, Exari, and
Perfectus IPManager in detail, focusing on the
strengths of each system. I also offer a peek at D3.
Each of these solutions are incredibly sophisticated
tools in their own right. Each tool would result in
significant people multipliers from its use. Each tool
empowers the author of templates to reach numerous
end users and give them the tools to rapidly
create sophisticated legal documents.
DealBuilder's Powerful Relevance Engine
One fundamentally different aspect of DealBuilder
is the "relevance engine" and associated analytic
tools. DealBuilder empowers the author to build
sophisticated rule-based systems with simple
brackets ("[ ..... ]") and braces ("{TestatorName}").
The DealBuilder Author tools analyze the markup,
check for proper syntax and logic, and build a complete
dependency tree that determines the end-user
interview.
This relevance engine enables rapid development
and deployment of complex legal documents. The
relevance determination is not limited to single templates,
but extends to the entire dependency tree of
inserted and assembled documents that form part
of a master template. The relevance engine checks
the dependencies and re-assesses the entire tree of
pages, groups, and variables each time a variable
changes, making the interview entirely dynamic.
Moreover, as the templates change, and the relevance
of variables changes, the interview automatically
reflects the current state of the documents.
If an extraneous variable appears in the interview,
it will be because the variable is used in one of the
documents being assembled. There exist myriad
analytic tools available to the author/developer to
debug interviews, including color-coded templates,
multiple alternate markups, annotated interviews in
developer mode that display the entire relevance
tree for each variable, and the ability to produce
documents in redline mode to show included or
excluded text.
The "authoring" process includes a powerful logic
check for syntax, dependencies, text structure, and
more. Tim Allen, President of Business Integrity, has
an open offer of a bottle of Dom Perignon to any
lawyer who can produce a marked up document
that does not have at least one logic error. Sad to
say, when I built my own DealBuilder templates for
E-Lawyer.com I was unable to collect on that offer
of Dom Perignon. Not to say that the syntax of
DealBuilder is difficult; rather the authoring analytical
engine works so well.
The most recent version of DealBuilder - version
2.7 - included some significant upgrades that built
on the core Relevance Engine that has been central
to DealBuilder's success. The Web-based interview
now includes a tree navigation structure that you
can toggle on and off. This tree is dynamic, like
everything else about DealBuilder, showing only relevant
pages and variables and refreshing automatically
on each page display.
The earlier version included support for what DealBuilder
calls ActiveText, the inclusion of variable
data entered earlier in the interview in Prompts,
Guidance, and Preamble text for variables, as well
as Lookups. With DealBuilder 2.7, you can now
include Computations in ActiveText which opens up
just about any form of reporting into the interview
process.
DealBuilder 2.7 also introduced two features of
particular interest in Corporate Counsel implementations:
(1) transaction server reporting tools and
(2) Lexicon support. DealBuilder has always had a
transaction server where answer data was stored in
a series of related SQL tables. With 2.7, Business
Integrity introduced Web-based reporting tools, an
interface to enable you to data mine the answer data
and produce reports that you can export to Excel.
Similarly, DealBuilder has offered use of Lookup
Tables to create dynamic prompts that could vary
based on state or country. But with 2.7, there is
now multi-language support. This includes localization
for multiple languages in the core prompts
and syntax of the system. It also includes a feature
called Lexicon, which enables you to have multiple language
prompts, with the language of a particular
prompt varying on the language setting of a particular
variable.
Finally, while DealBuilder does not bundle a work-
flow engine, it does contain flags and hooks
that can be used by such an engine, including a
parameter of the document that indicates whether
an interview has been completed fully and is thereby
LEGALLY SAFE to enter the workflow process.
This determination of completeness (and the
criteria thereby) is more than simply a determination
of whether any unanswered variables exist. It
is rather a determination of whether there are any
unanswered relevant questions that the author has
determined are required to be answered.
qShift - Knowledge Management Repository
As a practicing lawyer, I would have loved QShift. I
would build my documents in an outliner, so I could
take advantage of an outline tree to navigate the
different provisions and decide which ones to use in
the final documents. QShift is the ultimate outliner.
It is a desktop client to a Web-based document
model and clause repository, hosted by Ixio.
When you use QShift for a document, you take the
document structure and parse it into paragraph sized
elements. The analytic engine automatically
identifies the outline structure of the agreement,
including the titles of each paragraph and subparagraph.
These are parsed in to paragraph-sized
(or multi-paragraph-sized) clauses which are then
displayed in a document model.
You can then assign each clause a variable title and
several properties, including an outline level, and
status as optional or required, as well as group them
with several other similar clauses in either a select one
choice or a multi-select. QShift enables you
to see the document model live as you make your
selections as to clauses to include in the particular
model you are building.
QShift shines in the metadata included in the
clause-bank and optionally included in the outputted
document. Each clause, and each document
model can have unlimited annotations. Authors can
add comments on the appropriateness of use. You
can workflow individual clauses until the language is
formally approved and released. The End User can
also add comments. In Author mode you can easily
add comments to the system and notify the original
author of the clause. In the outputted document,
one of the formatting options is a Model Summary
which shows the Title and Text of each clause, its
status, and any comments on the clause.
Clause usage is not limited to a single template. A
clause paragraph from one model can be borrowed
by or shared with another model. Moreover
the system is smart enough to recognize a shared
paragraph and flag the author that changes to a
particular paragraph will affect other models.
QShift proves particularly adroit at handling Word/
RTF paragraph styles. Much of this process is automated,
including the selection of independent Word
style sheets so that the output of a single template
model can be formatted differently depending on
the preference of a particular attorney user.
QShift is not strong on variable logic and does not
currently support repeats and collections. While its
QShift Document Variables wizard is much improved
over an earlier viewing (including the ability to import
and export XML answer sets and field management),
the real strength of QShift lies in the clause
management, melding knowledge management
repositories, and document assembly.
Exari - The Power of XML and Platform Independence
Exari claims it is the only system that can be hosted
on an entirely non-Microsoft based system. Exari
can be hosted on a LAMP server (Linux/Apache/
MySQL/PHP) and run assemblies on an Apple
Macintosh. This is true because Exari adheres to
standards for XML encoding of templates, variables,
interviews, and logic.
Exari (which is SpeedLegal rebranded) is proud of
its XML heritage. The authoring environment is an
XML Editor. Word/RTF documents are uploaded
into a parser which intelligently breaks the Word
document into XML paragraph objects. From there,
you can can group or split the paragraph objects
(as required) into objects that can be separately
addressed. Each object in the system can have
properties and conditions associated with it.
The sooner you can make the conceptual leap from
"text markup" to object properties, the easier you
will find Exari to work with. For the author, the ability
to address an "object" instead of careful placement
of brackets and expressions in a WYSIWYG editor
like Word provides a much greater degree of control
over the formatting of the resulting document.
Along with the XML comes a powerful, Web-based
preview engine. As part of the process, you can see
the document in HTML, but with color coding indicating
the variables and the optional text. You can
hover over these items and see the question that
triggered that particular text, and even resume the
interview at the question that triggered a particular
passage.
Exari offers a unique feature, not present in the
other systems ... the notion of bias. While you can
manually code bias into templates in other systems,
bias in Exari is a property of a field. The bias options
include a particular configuration of defaults that
have a particular slant. In addition, there is an Interview
Style option that offer Express (only mandatory
questions) or Comprehensive (all questions).
Exari ships with a built-in output option to create either
a Word/RTF document or PDF document. You
can manage documents entirely on the Web-server
or through an integrated document management
system. The built-in system uses the user's login to
determine what the particular user can see.
Exari also handles lists of object particularly well. It
includes a List Property that includes punctuation
syntax for list items (a feature also found in Deal-
Builder), enabling you to specify standard punctuation
for list items in a particular list, the penultimate
list item, and the last list item. But Exari takes this
one step further with two unique features. The first
tells the system to "dissolve" a single list item, which
is great when the sublist contains only one numbered
item (e.g. "(A) The testator gives the following
gifts: (1) a gift of ____"). When combined with the
second function to set the lead-in text, this becomes
"(A) The testator gives the following gift: a gift
of ______". I cannot tell you how much advanced
scripting I have done in HotDocs to accomplish this
same result.
Perfectus - .net/XML Object Oriented Programming and Integration
Perfectus is a technological wonder. Where Deal-
Builder excels with its relevance engine that figures
out all the scripting of the interview automatically,
Perfectus gives a power-programmer the tools to
build an EasyInterview that is both comprehensive
and flexible. Taking the best of .NET, XML, and
object-oriented programming, every aspect of the
system is an object with multiple properties that you
can set and script. You can reuse and reference
objects, whether they be variables, clauses, expressions,
outcomes, pages, pictures, or bound strings.
The Perfectus IPManager brings all the objects together
into a powerful development tool for designing
the pages, variables, and variable layout objects
that makeup the Easy Interview. The system also
includes a visual representation of the data flow and
constraints between each of the page objects of the
Easy Interview.
Object orientation also extends to template objects
and clauses. These objects and clauses can exist
inside the IPManager Package or you can pull from
an external source, such as templates hosted in a
DMS like Hummingbird. The Query editor applies
equally to variables as well as data binding to external
sources.
Like DealBuilder and Exari, Perfectus includes configurable CSS (Cascading Style Sheets) that enable
the developer to brand and style the interview to
fit into a larger Web site. Perfectus includes layout
options of such flexibility that they are typically only
found in a desktop-document assembly software
environment like GhostFill. You can control and con-
figure each object on the page, with the final page
representing the sum result of the properties of the
constituent objects.
Perfectus takes the next step to include a built-in
process manager options in its Package Explorer.
The process manager includes control over templates
that are part of a package, including rules
over those templates, choice of conversion to RTF,
PDF, HTML and XML, and choice of distributors
including posting to a DMS system or e-mail. The
Hummingbird integration features that ability to pass
variables into profiles as well incorporate tags to
associate a collection of outputted documents with
a single EasyInterview and each other, including
the ability to resume an interrupted interview or to
restart as a fresh new interview.
The Easy Interview is particularly easy on the eye.
It includes variable colored typeface as well as type
size. The page editor enables you to drop HTML
text, including text with variables and conditional
outcomes right on the page. Prompts are displayed
in bold black, help text in small gray font, and
example text is in small light blue font. The Easy
Interview page editor even includes support for a
Picture Object in the interview.
D3 - Macro Suite and Clause Management
MicroSystems, makers of D3 declined to respond to
my request for a demo and access to their system
for this article. From a brief demo I was able to have
at New York LegalTech, I found D3 lacking in many
of the tools of a full-power document assembly
system. D3 focused more on incremental improvements
in efficiencies, rather than on the People
Multipliers evident in DealBuilder, Perfectus, and Exari.
While D3 may be easy to use, its lack of power
tools does not make it viable as a true document
assembly engine that can be used with "Untrusted
End Users." It works best as a "better drafting tool"
for lawyers and paralegals, offering major improvements
over standard Word-processing and macro suites.
Like QShift, D3 focuses on paragraphs and clauses.
Its use of XML encoding is interesting, including the
ability to bind data from various sources to the XML
tagged fields. However, it entirely lacks any ability to
generate dialogs. Questions are presented seriatim,
one at a time, solely in the context of the document.
As a knowledge management tool, it also lacks
many of the advanced features found in QShift for
collaboration on clauses, tracking of commentary,
and versioning.
Conclusion
I expect that I may have lost some of you in the
preceding discussion. It is hard to distill the essence
of each of these systems into a few hundred words
and still do justice to them. These systems have
hundreds of unique and interesting features, and I
cannot say you would be wrong in choosing any of
these systems.
The people multipliers are there if the system is
properly implemented. None of these systems
are "purchase and forget." They all involve some
degree of work. The more you put in, the more you
will get out.
Roy Lasris, developer of Pathagoras believes that
template creation should be a painless, almost
thoughtless process. QShift, DealBuilder, Exari, and
Perfectus include rapid authoring tools that can be
used by attorneys and paralegals with minimum
training to simplify and speed the path to automated
templates. But they also give you the tools to take
the next step to build systems that produce final,
PDF-ready documents, rather than "better first
drafts." The People Multiplier comes only when the
quality of the outputted document is truly ready to
use.
In assessing the work required to build a Perfectus
IPManager package, a QShift document model,
an Exari XML Template, or a DealBuilder template,
please bear this in mind: How much would you
spend to bring on a new attorney for a year? And
how much time (valuable attorney time) would you
extend to train that attorney in your processes and
methods until he or she becomes a useful member
of your firm? Now factor this time into the equation
in choosing whether and which online document
assembly engine to purchase as your development
platform.
Click here to download the PDF of Other Document Assembly Systems - Part 2 of 2
