Automating Template Creation and Document Benchmarking: Q&A with Kingsley Martin, CEO of KIIAC

Through the contract-automation grapevine I heard about Kingsley Martin, president of KIIAC LLC, a company using proprietary software to help customers analyze contract models. Kingsley was kind enough to speak with me about KIIAC and show me how the software works. It’s sophisticated and intuitive, and I think it meets a real need, so I was happy to turn my conversation with Kingsley into the following Q&A.

Q: What is KIIAC?

A: KIIAC is the name of the company Trace Liggett and I formed in February 2008. It’s also the name of the company’s software. The name stands for Knowledge Information Innovation and Consulting, and it’s pronounced “kayak.” Trace and I have many years of experience in legal practice, legal publishing, technology innovation, and software development.

Q: What does KIIAC software do?

A: KIIAC software creates templates and clause libraries for use in drafting contracts. The technology analyses any group of contracts, or “Reference Set,” and automatically determines what clauses they contain, how those clauses are organized, and the range of standard and non-standard language used in each clause. We call the generated templates and clause libraries a “Reference Standard.” Our technology then allows lawyers to select the clauses, and the language associated with those clauses, that best suit their needs.

Attorneys can use KIIAC in two different ways. They can compare, or “benchmark,” any given contract—for example, a draft submitted by the other side in a transaction—against a Reference Standard and consider possible alternative clauses and language. Or they can use the Reference Standard to identify standard and alternative clauses and use it as the basis for a document-assembly system.

Q: Tell me more about the technology.

A: Our template-management tool is called the Workbench. It consists of two elements, the Indexer and the Template Generator.

Image 1The Indexer reads a given Reference Set, classifies the constituent contracts by type, and breaks them down into their component clauses. The Indexer can analyze a Reference Set consisting of anything from a handful of contracts to hundreds. The contracts don’t have to be selected by hand, as the software efficiently allows you to weed out those contracts that are irrelevant for your purposes. Documents are classified by reference to a configurable taxonomy of document types, such as credit agreements, security agreements and promissory notes. The image to the right shows that part of a taxonomy for contracts related to finance.

KIIAC Image 2The Template Generator then automatically creates a standard document outline, or template. The image to the right shows a template outline for a credit agreement, reflecting the organizational structure found in a Reference Set of credit agreements and indicating how often each kind of provision appears. Icons to the left of each caption heading show in what proportion of the contracts in the Reference Set a given clause appears. For example, the sections entitled “Definitions” and “Representations and Warranties” appear in all contracts in the Reference Set, whereas a section entitled “Guaranty” appears in approximately 25% of the documents. The slider bar at the top of the outline allows you to limit the clause list to those clauses that appear more frequently or less frequently.

KIIAC Image 3

For each provision, such as the section entitled “Loan Documents,” the Template Generator creates a clause library—the list of clauses in the Reference Set documents that correspond to this particular provision. KIIAC will find these clauses no matter where they occur in any document. The image to the right shows a sample clause list. It groups clauses according to the extent to which they conform to the standard language found in the clause set. Icons to the left of each clause caption show how consistent each clause is compared with all instances of that clause in the Reference Set. Lawyers can quickly broaden or narrow their search by using the slider at the top of the clause list. Positioning the slider in the green area limits the group to standard clauses; moving it to the left would include clauses with additional deal-specific language and moving it to the right would include clauses containing less of the standard language.

KIIAC Image 4Lawyers can view the text of each clause in the clause list and see the extent to which it contains standard or non-standard language. The image to the right shows one clause from the list of Loan Documents clauses. It shows the text of this particular clause as compared to all the Loan Document clauses, with standard language displayed in green and deal-specific or divergent language underlined.

KIIAC automatically selects the standard clause—the clause containing the most widely shared language and the least amount of deal-specific language. Users can adjust the standard clause and can also elect to have the Reference Standard display a limited number of alternative clauses, so that those who then use the Reference Standard to create new drafts have limited discretion in selecting alternatives to the standard.

Q: How would KIIAC help a lawyer draft a contract?

A: Drafters can use a Reference Standard two ways. By using the computer generated Reference Standard “as-is”, a drafter can benchmark a draft document against the Reference Standard and use the search features of the software to make appropriate adjustments. Or the draft can use an edited version of the Reference Standard to assemble an entire contract.

Using an unedited Reference Standard, the drafter could browse the template outline, search for alternative clauses, and compare a given working document—perhaps a draft prepared by the user or by the other side to a transaction—to the Reference Standard. Such a comparison would show what clauses the draft being compared shares, and doesn’t share, with the Reference Standard and the level of divergence. This would allow the user to quickly identify those provisions that require further review.

The image below shows the outline of a consulting agreement (the middle panel) benchmarked against the Reference Standard (the left-hand panel). The comparison shows the matching clauses in black, shows the missing clauses in blue, and marks in red those provisions using language that differs from analogous language in the Reference Standard.

KIIAC Image 5

In the image above, the software shows that “Ownership of Work Product” uses divergent language. The user can find the standard language in the Reference Standard. The search features allow the lawyer to search by caption (it will find alternative caption headings), by a fragment of text, such as “work for hire,” or by using the entire text of the clause in question. In the image below, the lawyer has used the text of the clause “Ownership of Work Product” as the basis of the search and the software has found the Work Product and Intellectual Property clauses in the Reference Standard. The user can elect to view the standard versions of those clauses and if desired use some or all of it instead of the text in the working document.

KIIAC Image 6

If an organization uses the Workbench to construct a template and alternative clauses, they can be loaded on a web-based document-assembly tool. The document-assembly module can also be configured to automatically select alternative clauses according to document-configuration options specified by the organization. The image below shows an edited end-user license agreement template in document-assembly mode. It shows configuration options that allow the user to automatically select the clauses appropriate for an evaluation license, a commercial license, or an open source license. If the user selects the evaluation license option, the program would automatically select all appropriate clauses, although the lawyer can manually adjust the clauses selected. Once the desired clauses have been checked and the user clicks on “Save,” the program automatically assembles a Word document.

KIIAC Image 7

Q: So far we’ve discussed how your technology would help lawyers sort a mass of data. But does your technology incorporate any features that can help lawyers make qualitative judgments?

A: The Reference Standard outline is linked to an integrated wiki that can be used to collect and share any relevant practice knowledge. The template outline creates the framework for the wiki. Lawyers can create wiki content for the entire template or share their knowledge about individual clauses. The image below shows a merger agreement Reference Standard with related wiki content regarding representations as to organization.

 KIIAC Image 8

Q: So to summarize, describe how you’d help Acme prepare a template software licensing agreement from a set of 32 different software licensing agreements that it has used over the years.

A: Here’s how it would work: Acting as a consultant to Acme, we’d create a Reference Standard by processing the license agreements using the Indexer and Template Generator. The processing would take just a few minutes and would create a template outline and clause library.

If we were dealing with a large set of documents and the aim is to create a Reference Standard as a research resource, Acme would publish the computer-generated Reference Standard to a website for immediate use. But in your example, Acme is seeking to create a standard agreement based on a relatively small set of documents, so we’d use the Workbench tool to edit the template and clause libraries.

First, we’d review the outline and source documents to determine the variation in the Reference Set. If necessary, we’d break the Reference Set into subgroups, with perhaps one for object-code license agreements and another for end-user license agreements, with a separate sub-template for each group. Or we’d tag the clauses found in each sub-type so that they’re included in the document-assembly system.

Then we might refine the outline by rearranging the clauses, changing headings, or combining related clauses.

Then for each provision we’d review the default clause selected by the Template Generator and if necessary override the computer’s selection. We might also select alternative clauses in order to include them in the document-assembly system.

Finally, we’d edit the clause language in the selected provisions to ensure that all language is consistent. All selected language would be reviewed by Acme’s attorneys before the template is loaded into the document-assembly system.

Q: How does your system represent an improvement on the traditional system?

A: Those of us involved in document automation have long realized that one obstacle to widespread adoption of document-assembly systems is the cost of creating and maintaining templates. In my experience, it can take months to create a contract template. The template developer must review numerous documents, determine how they’re structured, and find and review all the alternative clauses. It’s an extremely time-consuming task, requiring both legal experience and skill at organizing large amounts of information. And templates created the traditional way have to be periodically reviewed and adjusted to reflect changes in the law and industry practice.

We find that too few people consider the costs of maintaining the templates. Often lawyers using templates edit the templates on their computer and those changes are not captured. As a result, over time the final documents increasingly diverge from the templates, and the templates have to be redone every few years. Our system can feed the final documents back into the system to audit compliance and capture improvements to the templates and clause libraries.

By automating both template creation and maintenance, KIIAC can significantly costs and thereby make document assembly more of a viable proposition. On our website, under the Financial Benefits link, you can get a better indication of the extent to which using KIIAC would reduce your costs.

Q: Do you have any competitors? If so, what distinguishes KIIAC?

A: Our principal competitors are Wordsensa and BaselineNDA, both of which have been featured on your blog [here and here]. Wordsensa has a bearing on contract drafting and automation, whereas BaselineNDA automates contract review. KIIAC addresses both those tasks.

KIIAC differs from Wordsensa in that KIIAC analyzes both the content and structure of transactional documents and presents document templates in an outline view familiar to practitioners. Both Wordsensa and KIIAC can tell you where similar language appears in a set of documents. But KIIAC goes further because it understands the structure and organization of contracts. Our software can distinguish between clauses in the buyer and seller’s representations, even though the language may be very similar. And KIIAC can distinguish between language regarding adequate insurance coverage that’s phrased as an obligation and similar language phrased as a representation. Furthermore, by capturing the structure and organization of a document, KIIAC can not only find matching clauses, it can tell you what you may be missing. For example, KIIAC can tell you what representations are typically found in a credit agreement and which representations might be missing from your draft.

KIIAC differs from BaselineNDA in that KIIAC is based on a statistical-analysis engine, meaning that it can develop reference standards and benchmarking for any contract type without having to incorporate in the software the logic and understanding of a subject-matter expert. And KIIAC can be used with any kind of contract.

Q: What’s involved in implementing KIIAC?

A: KIIAC can host your documents on a secure web server or we can install our software on your network. The Workbench is a Windows application and would typically used by a consultant or by lawyers charged with creating the templates. The web applications used for drafting and document review are available to all authorized users, without your needing to install specialized software.

Q: Your system has a bit of a garbage in, garbage out problem, in that the language you copy from the Reference Standard is unlikely to be particularly rigorous. Do you have any thoughts on that?

A: Yes—we provide the tools to organize the document template, create clause libraries, and group the clauses by language conformity, but technology doesn’t dispense with the need for a strong editor to retool the selected language so that it makes use of consistent, pre-approved usages. KIIAC and A Manual of Style for Contract Drafting would make a perfect team!

Q: What’s your pricing?

A: Initially, we won’t be charging for the software. Instead, we’ll be charging a consulting fee for analyzing and building templates.

Q: How can readers find out more about KIIAC?

A: By visiting or emailing me at You can test drive KIIAC by registering for a password to our website. The site contains a number of Reference Standards created from contracts filed on the SEC’s EDGAR system. Users can see how each Reference Standard is organized, see what clauses they contain and how frequently those clauses occur, search for alternative clause language, and upload one of their documents and benchmark it against one of the Reference Standards. But users should note that the Reference Standards on the website are in their computer-generated “as is” form, without any editorial adjustments. We wanted to show how the software is able to automatically create template outlines and clause libraries without the need for that sort of editorial input.

Categories Q&A

About the author

Ken Adams is the leading authority on how to say clearly whatever you want to say in a contract. He’s author of A Manual of Style for Contract Drafting, and he offers online and in-person training around the world. He’s also chief content officer of LegalSifter, Inc., a company that combines artificial intelligence and expertise to assist with review of contracts.

1 thought on “Automating Template Creation and Document Benchmarking: Q&A with Kingsley Martin, CEO of KIIAC”

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.