2009 March | Business Intelligence Blog

Outlook Series Interview — Part Five: Agent Community Business Intelligence Applications

Friday, March 6th, 2009

Here is part five of Bruce Molloy’s recent interview with host Mike Lippis of the Outlook Series. We have broken the 35 minute podcast interview into five parts for easier reading. This section explores a sampling of business intelligence applications that can deployed using Agent Community.

Mike: Okay, now with regard to business intelligence applications, what is Connotate doing about Data Silo Visibility?

Bruce: Sure, so this is inside the organization, where again, as we said earlier in the call that there are thousands, and in some cases tens of thousands of separate sources, so you can imagine the scenario whereby you have someone, say in research who knows of a particular database that has certain kinds of historical information at the same time there might be another source that has some kind of textural write-ups, or what have you. You could quickly create two Agents to go to those various, those two different areas, bring back the information, mash it up with a standard schema and make it available, whether in Excel or some other format. You could do that easily and it would make available then, this information that normally wouldn’t be seen because the alternative approach is typically going to IT and having a very heavy duty kind of integration program set up and put into place and the reality these days, and I think it’s always been, is that 20% of the applications the most widely used applications get all the attention and resources and the other 80% are never created. And that other 80%, they might, if ever created they might only be used by a small number of people, but none the less, they’re never created. With a solution like ours you can address those smaller applications that certainly are highly valuable to those particular individuals in the corporation.

Mike: Now how can a dynamic, competitive intelligence be leveraged to gain a business advantage?

Bruce: Sure. So many, many ways. Think about having your own collection, your own little personalized army of Agents and with these Agents, you can direct them to do various things, keep track of various things, both inside and outside the firewall. Let’s focus right now on outside the firewall. Out there on the Web, you can have these Agents doing various kinds of surveillance so certainly to keep track of product developments, what’s going on in the product space with a whole list of competitors, 5 competitors, 50 competitors. You can look at pricing, you can look at pricing, not just in general for the products but you can slice it into various areas, geographies. If you have distributors you can look at the difference between the pricing there and so there are multiple dimensions to what can be picked up. Additionally there are things such as regulatory requirements, statewide and on the federal level. You can keep track of what those compliance or those requirements happen to be. In terms of marketing, you can keep track of the kind of campaigns, the kinds of shows, the initiatives, what’s going on there and get it real-time, you know get it when it’s happening, you know, not after the fact when it’s perhaps too late to respond. Additionally you can be tracking overseas developments, financial developments if you have a significant supply chain feeding your particular product. You can keep track of what’s going on in every point in the supply chain. The examples go on and on and on so the way to think of it is if you had a series of helpers who are monitoring this information out in the world, where would you point them? You can probably come up with 7 or 8 important functions that would be valuable for them to do. Now to make it a little more interesting, the Agents can also monitor what’s going on inside the organization. And lastly, you can mash that information together… you might have some historical trends internally and you want to mash them together with the information happening real-time on the Web.

Mike: Is Connotate doing anything to personalize business intelligence?

Bruce: Yes. That’s absolutely what we are focused on. So you can think of what we provide as a kind of on-demand business intelligence solution. And so since users can create on the fly exactly what they need, when they need it, going after the information that might be just appearing, it’s truly personalized. And this is very different than a more “one size fits all” approach for various reasons we often time have to adapt to. So this notion of these highly personalized Agents, this highly personalized on-demand is a key part of how we think about our solution.

Mike: Now, can you quantify how Connotate Technologies is extending the reach of traditional business intelligence?

Bruce: Sure. So, there are various numbers having to do with traditional BI and really the kinds of BI we can make available with these lightweight on-demand Agents, if you will in going after the Web. I mean, I’ve seen some numbers that talk about internal sources represent really 20% of what a corporation needs and there’s another 80% out there on the Web so, you know, you can draw your own conclusions as to just how much a solution of ours can extend the reach but it’s certainly, it’s certainly significant.

Mike: Can you share a use case with the audience in terms of how someone has used these Agents and benefited?

Bruce: Yes, absolutely. So, the one in one in terms of quantitative information out on the Web would be the ability to check various dimensions of a given product’s pricing, how that pricing is comparing with its competitors, how that pricing and the ratio between a given (product price, competitor’s price) how that ratio varies in different geographic areas, how that is related to marketing initiatives. In addition, people are doing some very interesting things to find out more about where a product is going and you can go, for instance, if you look at product pricing or supply and demand and say on an auction site, like an eBay, or something else. Many dimensions that really help with that particular use case. Another is, as we were talking earlier about, connecting up and integrating information inside the organization, there’s something that we call a data silo visibility, essentially making, taking away the silo nature of these information stores and making it possible to rapidly aggregate and integrate this information. So as we take this to a kind of business intelligence application, it means that when a business intelligence analyst has a particular hypothesis or some notion that they want to test, they can quickly set up Agents that can actually remove the silos of information.

Mike: So is there anything else you would like to add to our discussion before we wrap up for today?

Bruce: If this work is interesting, and I hope the audience listening does find it interesting, I would ask you to come to our website www.connotate.com, and there you will see Agent examples. There’s an opportunity to look at various case studies to understand how the system works and I would invite you all to contact us and we’d love to take you through a demo or talk about how this might help you with your business intelligence applications.

Mike: Very good. Well I certainly appreciate you joining us today and wish you the best of luck with your future endeavors.

Bruce: Thank you so much. It’s certainly been my pleasure.

This is the final posting for the Outlook Series Interview. A full audio recording — with slide accompaniment — is avalable on the Connotate website at http://www.connotate.com/outlook2009.aspx

Posted in General | No Comments »

Outlook Interview Part Four: A Scalable, Extensible Framework – What Makes Agent Community Different from the Competition

Thursday, March 5th, 2009

Here is part four of Bruce Molloy’s recent interview with host Mike Lippis of the Outlook Series. We have broken the 35 minute podcast interview into five parts for easier reading. This section explores the key differentiators of Agent Community over other solutions.

PART FOUR: A SCALABLE, EXTENSIBLE FRAMEWORK - WHAT MAKES AGENT COMMUNITY DIFFERENT FROM THE COMPETITION

Mike: How scalable is your software engine?

Bruce: We did a lot of work on the scalability and we are proud to have a number of major publishers as our clients and they are bringing back tremendous amounts of information, terabytes of information from various sources and they’re essentially running server farms of our Agent Community solution and that’s working very, very well. So, the solution is highly scalable, it’s highly reliable.

Mike: Is your framework extensible and can your Agents move between installations and locations easily?

Bruce: Yeah, that was an important design aspect for us, making the Agents portable, making it easy for these to move around. They’re self-contained little nano applications, if you will, that are able to move between systems and this sets up some very interesting scenarios. You could have, for instance within an organization you could have 2 or 3 groups who are creating Agents, creating these kind of on demand applications, if you will. Once created they could be shared between the systems, you could pull a few Agents out of this one, email them to a user on another system and they could be run over there. It also means that you could have larger communities, communities of interest who would share these Agents and you could move them around. It also enables third parties creating these Agents and making them available for various applications. So I’m glad you asked that question because we think it’s a very important part of what we’re able to enable and that is truly these ecosystem of Agents moving between systems and doing their work.

Mike: So how does Connotate Technologies differentiate itself from the competitors out there that it deals with?

Bruce: Sure, so in terms of competition we don’t seem to have much direct competition. Let me categorize that, I mean a lot of the competition seems to be just very hard working people who are bringing information back from a screen, writing it down, cutting and pasting and doing this kind of work. We were doing a demo at one organization and there’s a woman there who thanked us and said “I can’t thank you enough, you’ve given me my Sunday back.” Apparently she had spent many Sundays sort of cutting and pasting information from the Web. Secondly, what people will do is they’ll write some kind of program or perl script to bring information back and again, this has an upper limit. You can’t, you know, you can only do, you know, dozens, maybe hundreds of these. We’re running tens of thousands and even hundreds of thousands of these Agents, so you know, that has limitations. Lastly, there are some tools that are, I guess, I call them programmer’s amplifiers or tools that kind of help you with the programming task, but it’s not point and click. You need to have, in the case of the Web, a knowledge of HTML and these don’t seem to be designed towards scalability, they’re a little more departmental. The other way I’ll differentiate our solution is we talk about this Library, this community, the sharing of Agents, no one else is doing that. In terms of the outputs, the analysis, the time series, the ability to easily personalize the outputs, no one else is doing that.

Our next post will highlight Business Intelligence and Competitive Intelligence applications that can be quickly deployed using intelligent software Agents

Posted in General | 2 Comments »

Outlook Interview Part Three: Community Collaboration and Digging Deeper into Agent Capabilities

Wednesday, March 4th, 2009

Here is part three of Bruce Molloy’s recent interview with host Mike Lippis of the Outlook Series. We have broken the 35 minute podcast interview into five parts for easier reading. This section explores the power of community collaboration and digs deeper into the many kinds of Agents available and the capabilities they offer.

PART THREE: COLLABORATION AND DIGGING DEEPER INTO AGENT CAPABILITIES

Mike: So, if collaboration is key to this and I know you had mentioned about sharing Agents between folks. To what extent can the Agents be personalized and shared? Can you drill down on that and share some details with us about that?

Bruce: Yeah, absolutely. So I’ll give you a couple examples, I mean, let’s just stay first of all with the competitive news example. I might set up an Agent and what I’m interested in is perhaps information about corporate development within these competitors. Someone from marketing might use that same agent and that person might be interested in shows or things that are marketing-related. So I could put in some set of keywords that would reflect my interests and our marketing person could put in something having to do with marketing and so the Agents become personalized. That’s a simple example, it actually gets pretty rich because let me take another example. We have Agents that can monitor Amazon, and on the Amazon website, they’re generalized so I might be interested in the top 10 books coming from Amazon, but someone else might be interested in the top ten music CD’s coming from Amazon. So that same Agent could be set up, now you can personalize them in terms of the information that’s coming from them, but you could also personalize how that information is delivered. I might want that information sent to my email, somebody else might want that information sent to a different device, to an Excel spreadsheet or to a database.

Mike: So, you know, we’re kind of focused on this Agent Library and we hit on personalization and sharing. Are there any other key features about that Agent Library that you can share with us that would be important?

Bruce: Sure. It’s been created with what I call “Web 2.0” kind of feel or “Enterprise 2.0” kind of feel and what I mean by that is that it’s just very, very easy to use. It doesn’t require a user manual. Essentially you log into the Library, you search for some Agents, using a search field. There’s also a tag cloud. Also within the Library, you can see samples of what the Agents will do for you if you have these Agents running on a subscription basis once a day going out and getting information, you can see what you’ll actually get, or a sample of what you’re actually going to get before you go ahead and subscribe. We have a Friends and Family version going now and people will send in requests for various Agents and a friend of mine had, was interested in bank rates, interest rates up in New England and that was quickly and easily set up and turned out to be very valuable for this individual who’s in real estate. Once that Agent is created, now anyone can go subscribe to that and find it very easily. You see this constant collection of, constantly growing collection of Agents.

Mike: So, you know, you have this Library and all of these Agents doing this work and they have to be, you know, providing you results. So, I mean how did you go about or what have to done to provide some sort of control over the Agents’ results?

Bruce: Sure. So in terms of controlling the Agents results and being able to view those etc. we have, first of all, a portal. We have our own portal, it’s very easy to use, that will track, let’s just give you another example, it could track job postings for a given company or a given industry and track those in time series over time and make these available within a portal window. Now again, the end user can configure this, set it up. Once these portal views are available they can be configured in various ways, you can have multiple of these on a given tab, you know, compare and contrast, you can mash them up. So it has a lot of powerful features. One of the ways our customers are using this is to see trends and to see correlations, to see what might be predictive on the Web. Once they’ve seen those trends, those higher level patterns they can click on a button within the portal and that information will be downloaded to an Excel spreadsheet. We have an Excel plug-in, so this allows a modelers and quantitative people to do more with the information. Also, within Excel we have plug-ins, the ability to place Connotate buttons within Excel and upon clicking on those buttons an Excel spreadsheet or model can go to the Agent system and pull back information generated by the Agents.

[Our next post will highlight the scalability and extensibility of Agent Community GEN2, and highlight key differentiators of the solution ]

Posted in General | 1 Comment »

Outlook Interview Part Two: Intelligent Software Agents — A non-technical approach makes it Easy to Train Agents and Personalize Delivery

Tuesday, March 3rd, 2009

Here is part two of Bruce Molloy’s recent interview with host Mike Lippis of the Outlook Series. We have broken the 35 minute podcast interview into five parts for easier reading. This section delves into the many kinds of intelligent software Agents and how Connotate’s approach makes it easy for non-technical users to train Agents.

Mike: So it sounds like it’s fairly easy to make these Agents. Everyone, including non-technical folks can make these and can they also train them as well?

Bruce: Yeah, we’ve spent a number of cycles to make sure that you could empower the end user and make this very easy. It took a lot of work, frankly, to make it so simple to use. It’s a complex technical problem we’re solving but again the emphasis has been making sure the end user, when they see some information or want to do a particular task, etc. can quickly do that. And I’ll give you 2 or 3 ways that is realized through the simple design of our software. One is Agent creation. If you have someone who’s working in business intelligence or research or an analyst or someone who wants to do price comparisons and that person wants to monitor certain, say, prices or developments or products from a competitors’ site they can very quickly and easily paint the screen, if you will, create an Agent and have that agent then available to monitor over time, every day, or every hour, every minute kind of, what’s going on. Secondly, in terms of the Library and this is, there’s a real multiplier effect here in terms of the Library. As you get people starting to share the Agents, those Agents come to represent really best practices, best ways to get information delivered, to look at it, to compare it, to mash it and as such it’s a repository of expertise that is then shared and multiplied in the organization. And lastly, in terms of output just because it’s so easy to get this output because it’s so well personalized it becomes a solution that individuals, non-technical folks in an organization can use without having to go to IT and get into a long development cycle, if it’s even possible.

Mike: So it sounds like there’s no programming knowledge needed to create these things, how long does it take a given, non-technical person to put one of these together?

Bruce: Great. The Agents can run a gambit from fairly simple to a little more complicated but, many Agents can be created in 5 minutes (watch how easy it is!). I’ll give you one scenario of a hedge fund client of ours. We had been giving a demonstration for them and they had shared with us that one of their programmers had been working 5 ½ days to get a particular script going that would bring back information from the Web. We did a demo on the fly for them, they gave us the website and in 5 minutes it was up and running. Their jaws dropped and they quickly got started with the system.

Mike: Now I know that you had mentioned the delivery capabilities but what about extraction, what kind of, what types of sources can these Agents extract data from?

Bruce: Sure. It’s a great question, again, our mission as we talked about from the top is making it possible to access all this information, all these different types of information, bring it back, normalize it, make it valuable. So some of the sites, or not the sites, but the types of information we can bring back include the internet, intranets, RSS feeds, databases, internal applications, email, PDFs and other kinds of data sources. The Agents are also able to, this is an interesting aspect, and they can automate processes. So for example, if there’s a particular application that you go to and you have to submit your password to get in you can have the Agent securely encrypt your password and do that automatically for you.

Mike: Now, you had mentioned, we had talked about ease of use and, how these can be put together by non technical folks and put together pretty easily and quickly. Can you give the audience a more detailed flavor on how that point and click agent creation works?

Bruce: Sure. Let’s stay with, we were talking about pricing examples; let’s talk about a news, a press release example. Let’s say you want to monitor 10 of your competitors’ sites and you want to look for news releases and when those news releases come out. You could easily set up an Agent in the following way: You would navigate to that press release site, within the Agent Studio, once in the Agent Studio; you would pick a data description, a schema, if you will, of the particular elements you want to map that page. In this case, it might be title, date and story. And what you would then do is simply using point-click you would map those 3 elements onto the particular press release (watch how easy it is to train an Agent!). Now, having done that once, the system would generalize and the machine intelligence would understand when it sees new press releases it would again break them up into those 3 element types and make those available on the platform. That’s easy to do, you can repeat that process on the nine other sites and now what you have is your own personalized feed of information that can be brought back to you through a portal, through email; you can put filters in, Blackberry alerts, etc. to give you a competitive view of what’s going on with your competition.

[Our next post will pick up the discussion and focus on fosteting collaboration through Agent Community and will further explore Agent capabilities]

Posted in General | 1 Comment »

Outlook Series Interview — Part One: Extend the Reach of Traditional BI with Agent Community GEN2

Monday, March 2nd, 2009

Connotate’s CEO, Bruce Molloy, was recently interviewed on the popular program — Outlook Series — with host Michael Lippis. The interview covered a variety of topics and applications to showcase how Connotate’s Agent Community extends the reach of traditional business intelligence.

You can listen to the entire 35 minute interview here, but our team has also broken the interview into five parts for easier reading. Here is part one — which focuses on the many data sources we need to contend with every day in our quest for valuable BI that will help us meet our business goals.

PART ONE: SO MANY DATA SOURCES, A SINGLE PLATFORM FOR CREATING BI VALUE

Mike: Saving labor on tedious tasks and freeing employees for more thought-intensive work generates a dynamic competitive advantage for your business.

This is being accomplished by creating intelligent tools that supercharge the knowledge management process by enabling a Web 2.0 ecosystem for accessing, sharing and delivering business intelligence.

Moreover, the solution is powered by machine-intelligent software Agents that act as an extension for business staff to collect, aggregate and mash-up near real-time data from reliable Web and enterprise sources. Furthermore, the next generation of BI applications provide on the spot extract transform loading which is created by non-programmers from intranets, email and documents.

Now joining us to discuss how to extend the reach of traditional business intelligence with software Agents is Bruce Molloy. Bruce is the CEO of Connotate Technologies. Bruce Molloy, welcome to the Outlook Series.

Bruce: Thank you. It’s good to be here.

Mike: Hey there Bruce, so can you share some background with our audience on Connotate Technology’s history and general mission?

Bruce: Sure. I’d be happy to, Mike. Let me set up the state of information today so that I can share with you Connotate’s mission. There’s a tremendous amount of information available on the Web inside enterprises. IDC has estimated that digital information is growing as fast as ten-times (10x) during a given 5 year period. This is just a tremendous amount of information. Within enterprises it’s estimated that there are thousands or even tens of thousands of separate data sources, valuable data sources, but unfortunately disconnected. Now here’s where Connotate’s solution comes to bear. Getting at this information, collecting it, collating it, on the Web, getting down to the element level within the enterprise, being able to quickly pull together elements from various systems, it’s very very difficult. And there’s a very surprising number it’s estimated that some professional workers will spend as much as 50% of their time collecting, monitoring, assembling, bringing this information back in order to create some kind of intelligence from it. It’s terribly inefficient. So here’s what Connotate does Connotate has this notion of intelligent software Agents that are able to go out and do a lot of the things we do. They’re able to monitor and aggregate, watch, make decisions, format, and really supercharge various kinds of intelligence processes. And Connotate is unique in its approach in that these Agents can be created by anyone in the organization. You don’t have to be a technologist; you don’t have to be a programmer. And so we see Connotate’s mission to really make available, to make actionable, make valuable, all this information that exists on the Web and in the enterprise.

Mike: Now, can you kind of put some numbers on how many disparate data sources a typical individual corporation has and how challenging is it to make sense of all that data?

Bruce: Sure. So I’ll give you an example, we were over at a pharmaceutical company recently and they had, between the databases, intranet applications and other sources of information about 6,300 separate sources of information. The challenge is how do you get at those? How do you link those together? How do you create value? And using a traditional approach of very heavy duty integration and plumbing, it actually is not economically feasible. So using our approach, you can create these lightweight Agents that look at the various data stores, bring back the information in a normalized fashion and make it available, whether through Alerts or Excel or some other form that it may be consumed in.

Mike: Can you share an overview of the Connotate Platform with the audience?

Bruce: Yes. So first of all, just to talk a little bit more about the Agents. What the Agents are — are scripts or little programs, if you will, that know how to do something extremely well. And taken together as a group they can do some very interesting kinds of processing you can’t do any other way. The Agents can be created (and are created) through a simple point-and-click interface. The Agents are then shared with groups of people and the Agents can output to an array of devices. So let me drill down a little bit more on that platform that the Agents reside on.

First of all, our platform is called Agent Community and it’s a community for Agents to come together to be shared, to be used in various ways, and it’s a community for users for various knowledge workers within the organization to get those particular Agents they may need and would want to share. Three aspects to the platform, one is the Agent Studio. In the Agent Studio you can bring up a web page, you can define elements, schema, so you could come up with, for example a pricing schema that might include a product name, a description and perhaps a price and availability. And you could quickly paint, the screen, if you will, with this point and click interface to tell it that this is a price, a description, an SKU, availability, etc. And once you’ve done that, once you’ve trained the Agent, the Agent will continue to go to that page and bring back those elements. It can look at other similar pages on the same website and bring back literally thousands or even millions of pieces of information; it can do it over time. The same approach can be used with databases, Excel spreadsheets, email, and a whole array of other kinds of formatting of electronic information. So this is very powerful and the technology behind it is a patented machine learning technology that really no one else has. It’s very unique and it supports 2 things: one, ease of use, and secondly, reliability and scalability. The second part of the platform is the Agent Library. So once these Agents are created, they’re given names, they are typically run so that a data set is made available so you can view in the Library what that Agent will do. And here’s the neat part, once the Agent has been created others can come along and personalize those in various ways. I’ll give you some examples; it could be an Agent that goes to an electronics site and is looking at prices of various kinds of electronics products. Now, I might set it up so it’s looking at cell phones. You could go to the Library and you could configure that same Agent so it would be looking for LCD’s. And in various ways you can put in filter terms and various kinds of date ranges and other information to truly personalize those Agents. Also you can take the Agents and you can mash them up so if you have Agents that might be running, say something is running outside the organization and you have something that’s running inside the organization you can combine those Agents and put together some sort of correlation between those two different lines of data. Finally in terms of delivery, the Agents can deliver to, and I’ll say, two different types of destinations. One would be databases and sort of machine destinations, databases, web services, etc. But then also human destinations, so you can send that information to Excel spreadsheets, to your Blackberry, to email, to any number of alerting devices, in fact you can also send it to a mobile device. You could set up a little Agent to run on an iPhone and bring back various kinds of information.

So that is the Connotate platform, it’s called Agent Community GEN2 and it’s a means, it’s a whole solution for creating these Agents, sharing them and delivering some real value.

[Our next post will pick up the discussion and focus on Agents and how this non-technical approach makes it easy to train Agents and personalize delivery options]

Posted in General | 4 Comments »

Pages

Archives

Categories

Archive for March, 2009

Outlook Series Interview — Part Five: Agent Community Business Intelligence Applications

Outlook Interview Part Four: A Scalable, Extensible Framework – What Makes Agent Community Different from the Competition

Outlook Interview Part Three: Community Collaboration and Digging Deeper into Agent Capabilities

Outlook Interview Part Two: Intelligent Software Agents — A non-technical approach makes it Easy to Train Agents and Personalize Delivery

Outlook Series Interview — Part One: Extend the Reach of Traditional BI with Agent Community GEN2

Pages

Archives

Categories

Archive for March, 2009

Outlook Series Interview — Part Five: Agent Community Business Intelligence Applications

Share This Post!

Outlook Interview Part Four: A Scalable, Extensible Framework – What Makes Agent Community Different from the Competition

Share This Post!

Outlook Interview Part Three: Community Collaboration and Digging Deeper into Agent Capabilities

Share This Post!

Outlook Interview Part Two: Intelligent Software Agents — A non-technical approach makes it Easy to Train Agents and Personalize Delivery

Share This Post!

Outlook Series Interview — Part One: Extend the Reach of Traditional BI with Agent Community GEN2

Share This Post!