Gimmal Blog

Read the latest thought leadership and industry news from the experts at Gimmal!

All Posts

AI, IG and what it all means: Our Interview with Reid Smith

Recently, Gimmal interviewed Reid Smith, co-founder and CEO at i2k Connect. The mission of i2k Connect is to revolutionize information discovery using its novel artificial intelligence (AI) technology, informed by industry knowledge, to transform unstructured documents into structured data.

He spoke at length about artificial intelligence, information governance, GDPR, and a host of other topics. We've condensed some of the highlights below.

Gimmal: How do you see the growth of machine learning and artificial intelligence affecting the way that information governance will evolve in the next five years?

Reid Smith: The first thing is that manual information governance is not sustainable. I see AI and machine learning as the way forward, and it's not just machine learning. It's all the advances, including natural language processing, in addition to what's going on in machine learning.

I see that organizations will be able to use AI to automatically read, tag, summarize, find and analyze documents, and not have to rely on people to do all of that by hand. Also, to track new documents as they are written inside the organization or as they appear on the internet. A way to look at it is: imagine that your search engine could bring to bear industry and company knowledge, to find your documents.

Then, going farther, imagine that you could analyze the information in your documents with as much ease as you can for the structured data in your databases and applications. I see that being enabled by all of the AI technologies, including natural language and machine learning.

G: What can organizations be doing to prepare themselves and their legacy content to take advantage of this change?

RS: It seems to me that it is always important to answer these questions: What are your processes? What is your governance model? What are the policies that you want to adhere to?

Then, assume that you're not going to have people do all of the nuts and bolts, low‑level work to adhere to those policies. You will be able to use AI technology to do much of that work.

There's a quote that I wanted to come to from Mike Alsup. [from this article:] "Transparent systems aggregate a multifaceted understanding of who you are, and what you are doing, in order to cater to you and to your organization's requirements and the government's policies."

I think that's right on. If you will allow me, I would lengthen the sentence slightly, to say this – “... understanding of who you are, your industry, your company, your job, and what you are doing ...” All of those kinds of knowledge – of your industry, your company, your job, as well as what you are doing – are going to get built into the technology over time.

Another point to keep in mind when thinking about what technology to employ: It may soon be important to be able to explain decisions taken by AI-based systems. Here I am thinking about GDPR (General Data Protection Regulation). Today, this is not possible for systems based on statistical learning (aka deep learning).

G: Very well. Then one final question, Reid. This is a more pragmatic question. There's always that person that says, "You know, I just wanna put it some place and know that that's where I can find it, and whether I go back 10 years or 20 years from now, I want my place."

I would think that the only way you get around that is to either use a strategy where you allow some of that, or in combination with this type of approach, or you're just better and so you become better than they are.

RS: There are two things I have taken away from my experience. As I said at the outset, manual information governance is not sustainable, because people do not have the time, training or 24x7 consistency to do this work ... in addition to everything else they are expected to do.

The other is that you've got to minimize the amount of change you are asking people to deal with. It's hard enough to ask people to adopt a new technology, and the farther away it is from the workflows that they're familiar with, the harder it's going to be.

As a result, our premise is, “We don’t care where you put your stuff, as long as our technology can get to it. Put it where you want to put it. If you want to navigate it the way you always did, we'll support that. If you want to also navigate in a different way, and use search more than you did before or a combination of search and navigation, great. We're fine with that, too."

G: At Gimmal, we think that's the best perspective as well. We want to thank Reid for his time and this informative interview. You can learn more about i2k Connect by visiting them at

New Call-to-action

Related Posts

3 Tips to Ensure KORA Compliance

There has been a spotlight on the Kansas Open Records Act (KORA) in the media lately, largely due to recent violations. Under KORA, any individual can request public records from government bodies. If all requested records are not provided within in a specific timeframe, these organizations are subject to significant repercussions. This is merely one example of a ‘sunshine law’. The purpose of sunshine laws is to provide transparency into government agencies by giving the public access to local government proceedings.

Creating a Retention Schedule that Works

Creating a usable, automated, and simple file plan is an important part of ensuring records are managed in a consistent manner and that you are protected from legal risks, such as failure to disclose information during a discovery proceeding or the unauthorized leakage of information. The first step in the process is creating a retention schedule, which outlines how long records are kept in accordance with the organization’s obligations and the law.

How to Manage Your Sprawling Content

Sprawling content, the spread of content across multiple repositories, has been a thorn in the side of records managers since the dawn of document management. Consolidation of repositories, which began in the early 2000s, at first looked to be the solution. However, it ended up highlighting the problems of content sprawl due to the high costs of consolidation as well as need for records managers to manage multiple file plans. Federated records management offers a solution to these problems but doesn’t offer the same locked-down approach with regards to regulation that consolidation can. Consolidation of repositories and federated records management both have pros and cons and, depending on your organization’s content management processes and repositories, one can be more beneficial than the other in the long term.