Don’t Just Click Yes! – AI Start-up Binadox, the IT Compliance Solution.

Artificial Lawyer recently caught up with Michael Kholodenko, co-founder of Seattle, US-based start-up, Binadox, which makes use of Natural Language Processing (NLP) to provide a compliance tool for software licences that can prevent costly legal violations.

At first you may think this does not sound that world-changing, but sometimes it is the very precise uses of AI technology that end up creating a very successful and useful product for lawyers and legal departments.

This would appear to be doubly so, as Bindox points out, because getting your software licences in a mess, or just allowing staff to click ‘Yes – Accept Terms’ can open a Pandora’s box of legal problems. Binadox seeks to meet this problem head on in an automated way.

Fundamentally the software seeks to answer two questions: what happens when an employee downloads new software or subscribes to a cloud service (SaaS) and clicks on the little ‘Yes’ box that agrees to the licence terms, with all the legal issues that suddenly connects to? And, how can a company or law firm, manage this weak point in its digital behaviour?

Now, many people may initially share Artificial Lawyer’s first reaction, which was: ‘It’s just a little tick box. Surely it doesn’t matter that much?’ Well, it looks like it does matter, as Binadox states: ‘If your employees don’t follow licence terms, your organisation could:

– Be sued or face a costly audit by software providers

– Be denied compensation due to limitation of liability

– Permanently lose important data

– Violate privacy, security agreements or regulations compliance in general

To help solve this problem the system shows what software has been downloaded, by whom and what them clicking ‘Yes’ to the licence terms may actually mean.

It does this by reading the licence with NLP and then can compare its legal provisions against the terms that the company or law firm allows. This avoids a company accidentally breaching its own policies and getting into hot water by agreeing to software licence terms that could create business conflicts or financial issues.

An illustration of how Binadox uses NLP to read software licences and compares them to permitted compliance terms.

In short, Binadox is a compliance software platform, operating with NLP to supplement the work of those people in a company, often inhouse lawyers, who manage such issues – if indeed the inhouse lawyers even have the time to look at this problem. It’s possible some companies and law firms simply don’t focus on this issue, but perhaps would if there was an easier solution other than manually checking.

At the same time Binadox says that it helps IT managers to be more efficient with software assets procurement. Its Software Usage Monitoring tracks software installed across a company and identifies used and unused applications, their last-time usage and cost. With that IT managers can discover all the purchased licences that are not used to optimise software costs by reassigning unused licences to other machines, revoking redundant licence agreements, or changing their procurement policy in the future.

There now follows an interview with Binadox founder, Michael Kholodenko.

Michael Kholodenko, Co-Founder, Binadox

When did you launch and who are the founders?

We launched in Q2 2016. Binadox has four original founders: Igor Gutnik, David Gornshtein, Boris Tamarkin, and myself, Michael Kholodenko

What gave you the idea to start the company?

While working at other companies we saw the issues, which arose when users accepted clickwrap agreements without reading them and moved right to installing the software. We also observed the situations when later at the negotiation stage (e.g., at the end of a trial period), these users seemed to be surprised of certain terms which were in the EULAs they had accepted.

Such observations combined with the rise of SaaS encouraged us [founders] to protect end users from the liabilities that could arise from license agreements. This idea pushed us towards developing a solution to automate the process of capturing and reviewing different software licence agreements so that organisations and users could always be aware of the terms they agreed to when they installed software or used SaaS.

How important is it to understand software licence agreements if you are a company or law firm?

It’s very important, because software licence agreements are binding legal documents, which you sign by agreeing to their terms. You wouldn’t sign a document which you haven’t read and don’t understand, the same refers to online contracts. Failing to understand the terms you’ve agreed to can be very costly: from true up costs in case of an audit, to general litigation costs in case of a dispute, or other indirect monetary expenses such as lack of compliance, risk of exposing personal information, etc.

Who in a company normally looks after this type of problem?

It can be a sophisticated IT team with vast legal knowledge, an in-house general counsel or a licensing / SAM (software asset management) expert.

What kind of NLP are you using? Does it spot specific clauses? Or does it mainly spot keywords? How does the compliance check work?

The Binadox NLP engine capabilities are based on machine learning, statistical methods, context and linguistic features which enable the following:

  • Summarisation capabilities, which are implemented in a “Level of details” mechanism, which allows summarizing a document up to different levels, starting from an original non-summarized text and up to a fully summarized text with summarisation of every section
  • License type identification – (paid (commercial) / open-source / freeware / shareware)
  • Detection of common topics- (software license related) and highlighting a part of text, which is related to a specific topic (can be scattered across the document)
  • Clauses detection – highlighting clauses grouped as CAN’T / MUST / MAY permissions.

Additional inter-document functionalities are:

  • Similarity mode– an ability to find contextually similar documents (thus, if there is a document which is highly similar to the currently reviewed agreement, the reviewer can concentrate only on the differences instead of analysing the whole document).
  • Compare mode – an ability to compare two documents. The Compare mode features a section matcher, which is a graphic representation of similar sections – be it due to similar section headers or contexts. A section in one document can relate to more than one section in another document.

screen-shot-2017-02-01-at-17-20-04In general, a compliance review is done by a human expert, Binadox provides the tools mentioned above to reduce time for the review and focus on important issues.

An additional tool to automate the process of compliance review is a rule – one can automate approval and rejection of licences by setting a rule based on previous decisions and a corporate policy.

How much time does this automation of compliance save a company?

It depends. In some cases Binadox can help focus on important parts and ease navigation within the document. It can show only the differences between a new licence and a previously handled slightly different license (in case of updates or minor changes between the vendors), or completely automate the process by applying a rule.

What is your pricing model for clients?

Binadox solution is divided into two parts. The first includes agents which actually capture software installation licenses and SaaS subscriptions’ Terms of Services from end-user hosts in an organization. The second is the Web-based server which presents all the information.

Considering that, we offer:

Licence Analyzer – A free tool with restricted funtionality allowing review of manually uploaded licences.

Starter package – A free version (requires registration) of Binadox suite which supports up to five agents for IT asset management and allows reviewing, sharing and downloading licence agreements.

Enterprise package – A commercial version supporting SaaS-based / on-premises IT and legal solutions which support as many agents as you need.

Do you have any law firm clients/can you name any corporate clients?

At this stage we have only corporate clients, which we are not disclosing at this moment. We don’t have law firms as clients yet.

Do you have future plans to develop/change the product(s)?

Yes, we have a solid road map for many months ahead. A lot of ideas come from customers who give us feedback and share what new features they want to see. Every two months we add new features, functionalities or enhancements.