Amid the insurgency in Iraq, American junior officers are sharing the lessons of urban warfare and teaching themselves how to improvise and think creatively via Web sites hosted on the military's secure SIPRnet. Maj. Patrick Michaelis, the founder of one of these Web communities, CAVNET, explains how it works.

While on patrol, the commander of an Army unit learns that Iraqi insurgents are wiring improvised explosive devices behind posters of radical Shiite cleric Moqtada al-Sadr, rigged to explode when U.S. forces tear them down. He posts the information on CAVNET, an interactive community hosted on the military's secure SIPRnet. His information is read by an officer located on the other side of the city, who then briefs his unit, which then uncovers the rigged posters in its sector and safely disposes of them.

This is one example of how CAVNET -- "designed to prepare for the next patrol, not the next war" -- offers junior leaders in the 1st Cavalry a real-time forum for exchanging knowledge and ideas in over 30 mission categories, ranging from "Civil Military Operations" to "Gear." The implementation of CAVNET, which became operational in April 2004, represents a change in how the Army adjusts to events in combat; rather than have incident reports filed within one unit and passed through chains of command, the information is immediately available to commanders from their peers -- both within and outside their units -- who are experiencing similar situations.

Maj. Patrick Michaelis, the founder of CAVNET, is Battle Command Officer and Task Force Chief Knowledge Officer for the 1st Cavalry Division in Iraq. In the following Q&A, conducted via e-mail with FRONTLINE Web Associate Producer Kate Cohen, he describes the site's origins and the intent for CAVNET to "leverag[e] horizontally and vertically what we are learning about ourselves (what is and is not working) and the enemy (what they have adapted to) for advantage." Michaelis also explains how the system can improve the "lag between what is being taught in the classrooms and evaluated at the training centers and what is being executed on the ground in combat."

Can you explain the thinking behind the development of CAVNET?

Essentially CAVNET was created to harness what is being learned internal to the division -- in terms of emerging enemy and friendly TTP [Tactic, Technique, Procedures] at the tactical level -- for competitive advantage against a networked, adaptable, cellular enemy force. The method for implementation is both technological and behavioral.

The technological is the creation of a SIPRnet (classified Internet) knowledge transfer system, accessible to company-level leaders, allowing them to rapidly share and integrate "actionable" (contextually-based) knowledge into their plan, prep, and execute cycle. This leads to an increase in the tempo of operations against an enemy force that is rapidly adapting to our tactics. Leaders are not required to follow what is posted but can "adopt, adapt, or discard" as their mission analysis dictates. As Maj. Gen. Peter Chiarelli, the Task Force Baghdad commander likes to say, "CAVNET is designed to prepare for the next patrol, not the next war."

The behavioral side is really a significant change in the way we look at information, data, and knowledge from a cultural perspective. We, as an organization, are one of the flattest organizations possible when you talk about information and data. Our "systems of systems" approach provides an incredible level of situational awareness. If a unit reports an improvised explosive device at a certain location, or that it has conducted a cordon and search detaining one individual, or that a unit conducted a traffic control point at a certain point, our systems and process do a phenomenal job in making sure everyone connected has that information at their fingertips. Situational awareness is achieved, it shows up as an icon on a computer screen, or a sticky on someone's map board. This is the drive for a "common relevant operational picture" (CROP for short). It is what I consider a manifestation of Col. John Boyd's OODA [Observe, Orient, Decide, Act] Loop -- operate within an opponent's ability to orient and you will create confusion and an inability to understand, giving you the initiative.

But though all of our investment has gone into creating the CROP to provide a competitive advantage, it does not necessarily mean you have common understanding. How a leader interprets the data and information presented to him is a function of his experience. What a division commander will see when looking at the CROP is significantly different from what a new second lieutenant will see. Why? Experience. …

There is another way of creating knowledge that helps create common understanding; it's through transfer of experience. And though we are the flattest organization on earth when it comes to transfer of information and data, we are culturally a rigid hierarchical organization when it comes to transfer of what I call "actionable knowledge." Instead of an improvised explosive device (IED) discovered at a certain grid, how did the unit find the IED? Instead of a cordon and search resulting in one target detained, how did that happen? Did you know that when the unit went to conduct the cordon and search the mark was not there? So the unit left a three-man team inside the dwelling, and the mounted element moved out of sight and sound of the target location, [and] 25 minutes later the mark arrived and was detained. You see emerging tactic, technique, and procedure that can immediately be incorporated into another leader's decision cycle.

A screen shot of CAVNET.

A unit conducts a traffic control point at a certain location. But in this case, the soldier that has to stop traffic has a certain set of rules of engagement once presented with a threat. The soldier can shout, wave, fire a warning shot, then shoot to disable the vehicle. But this evening he pulled out his 9 mm. to execute a warning shot against a vehicle that was driving erratically, and instead of firing the shot first, he used the laser pointer on the weapon and put a dot into the drivers chest, [in his words], "stopping the vehicle on a dime."

Actionable knowledge -- what is being learned now that can be incorporated into planning, preparation, and execution. Couple that with our incredible situational awareness capabilities, you have a very powerful set of tools. It is looking at Boyd's OODA Loop from another angle. This is in essence what the CAVNET seeks to achieve -- it is the "how" of operations. It is leveraging horizontally and vertically what we are learning about ourselves (what is and is not working) and the enemy (what they have adapted to) for advantage.

How is this different from the old way of spreading knowledge?

As I stated earlier, our culture is inherently hierarchical and stove-piped when it comes to the validation of "new actionable knowledge." Normally it will stay within the unit based on the way we have created our After-Action Review process. The learning that is achieved will be fed back into the same unit. But wouldn't it be great if that learning could be transferred laterally? …

CAVNET attempts to rectify our approach to actionable knowledge by providing a trusted source for junior leaders to immediately tap into and share the emerging enemy and friendly tactics, techniques and procedures laterally so that like units can arm themselves with the most recent enemy and friendly trends. This implies trust from above and a method to validate without losing the relevance. Maj. Gen. Chiarelli is a big proponent and integral cultural driver in moving from "knowledge is power" to "knowledge shared is power." His implicit trust in subordinates, realization that this is a small-unit war, and chief advocating [of] cuts across the inherent cultural divide created by hierarchical "turf wars" [is] driving this change across the organization.

How did you get involved in this project?

Weird story. I was teaching a course on leading organizational change at the United States Military Academy, and also running Platoonleader.org as a hobby and came to the conclusion that the model [of] the guys who run Companycommand.army.mil and [that] I was running through Platoonleader.org was much more powerful than we all collectively realized. I had an idea for a series of unit-level networks at each of the major Army installations, resourced by high-powered captains who worked directly for the commanding general [and] who would collect, observe, connect, collaborate, and disseminate -- on behalf of the command -- the created knowledge of a unit. The idea was to leverage what is being learned so a new guy coming into the installation could tap into it and come to the organization at a higher level of experience than [what] was currently being accomplished.

Every company-level leader leaves command with a CD full of stuff they have created and collected. Where does it usually go? Nowhere. Into their brain for future use. Why not create a structural requirement that every leader out-processes through this "cell" that can tag it, post it, and provide it for future leaders? Now think of every major installation, under a common taxonomy, that could be tapped into for the purpose of creating new doctrine -- you get the picture.

That was the idea. It was not initially received well by my brothers working Companycommand.army.mil or one of my old commanders who was now the operations officer for the 1st Cavalry Division. "Too hard," "Not relevant," "Won't work," "Not what we're about," were the common responses. I guess the idea kind of festered in the mind of my old commander, Col. Paul Funk, who brought up a variation of the idea to Maj. Gen. Chiarelli, and Lt. Gen. Thomas Metz (III Corps commander) who saw the power of the idea in a different context.

Col. Funk called me up in November 2003, stating he had just talked with Maj. Gen. Chiarelli and Lt. Gen. Metz and that he wanted me to create one of these "things I was talking about" -- in Iraq. That is the genesis that has led me to beat down many walls, create a system, promote the system, interact with the target audience every chance I get to achieve the purpose as stated above.

How long has CAVNET been operational?

I arrived in Baghdad in March of '04 and spent a month harnessing emerging TTPs from units on the ground. The actual site went operational on April 4, 2004. As with any new system there is a learning curve. It really exploded in use around June of '04. …

Today we are getting about 50,000 "message views" and about 100,000 hits to the site a month. Now, relative to the regular Internet this is not a lot. But if you consider my target audience is 220 company commanders, and the percentage of people on the SIPRnet is a fraction of a percent of who is on the regular Internet, that is an incredible number.

How much access do CAVNET's users have to computers or SIPRnet while stationed in Iraq?

I started with three assumptions: One, the enemy we are fighting against is adapting faster than we are at the tactical level; two, we exist in a relatively static environment (we live in forward operating bases) -- therefore company-level leaders have little chance to physically interact; and three, that SIPR access to the battalion level was sufficient.

In the first two cases I was pretty much on the money. The last assumption was wrong. What I was finding is that if access was not relatively available, then the target audience was not going to use it. Think of it like this: Are you going to walk into your boss's office to check e-mail or surf the Web? Probably not.

We spent money connecting the division's company-level leaders to the SIPRnet to ensure they have access and can share what they are learning. In those areas where we have 100 percent company SIPR access, over 85 percent of the companies are using the system on a daily basis (without having to educate them on the system) and two-thirds of the battalions are integrating what is being posted into their daily orders. ... Right now all battalions have access. About 50 percent of the companies have direct access in their command posts. That is a leap from about 2 percent in August. Plans are in place with the 3rd Infantry Division to complete hook-ups in every company level command post.

Who uses it?

... Although I originally designed it for just the commanders to share among themselves, others contributed that I had not foreseen that provided the same purpose as I stated above. The intelligence and analysis folks found this [as] a way to mainline their products to the lowest level, allowing them to get immediate effects across the battle space...

I look at the site usage as three concentric circles. The first circle is my target audience, the 220 company-level leaders and the 1st Cavalry Division preparing for the next patrol.

The second circle are follow-on forces. [Their] having the ability to tap into what it is the division is sharing in terms of emerging enemy and friendly tactics, techniques, and procedures can help shape their training and preparation. Right now our follow-on force is on the ground conducting their relief in place. There is a convergence between the first and second circle in helping prepare Third Infantry Division for the fight.

The last circle is the Army as a whole. There is increasing desire from the schoolhouses and doctrine writers to tap into what is being shared among company commanders. I call it closing the gap between the institutional army and the operational army. There is a lag between what is being taught in the classrooms and evaluated at the training centers and what is being executed on the ground in combat. Somehow there has to be a way to close that gap.

My solution is two-fold. The first is to allow complete access to the CAVNET to any who have access to the SIPRnet -- I'm not interested in hoarding what is being shared -- again, increasing the tempo of operations to reflect reality on the ground. The problem is that the SIPR pipes back to the rear are slow. It takes a long time to load the site, so there are interested parties who are attempting to create a mirror of the site for data mining from the States.

The other method is through inviting current doctrine writers, those on the platform at the school houses, and those creating scenarios and evaluating at the training centers to come work as an embed for a 30 to 45 day cycle. This achieves a dual purpose. The first is that I get an embed who interviews, connects, populates, and demonstrates utility (moving leaders from "lurkers" to "interactors") of the CAVNET to the brigades I place them in. They will go on roughly 25 to 40 patrols with units to understand context. The second purpose (and it is debatable which is more important) is that they mainline what they are learning back to their organizations, closing the proverbial "gap." I only keep them here for a maximum of 45 days so they do not lose relevancy to their parent unit, but can take back and integrate into what they are doing what they have observed and learned. At present we are attempting to solidify this program with TRADOC [Training and Doctrine Command].

Can you give examples of how CAVNET has been used in Iraq?

Yes. Two off the top of my head. A leader posts a report that his unit experienced an IED that was cloaked by a poster of Moqtada al-Sadr. On the other side of the city, a commander taps into the CAVNET and reads the post. Though he is in another part of the city, he has been involved in operations that require removing posters posted on IIG [Iraqi interim government] projects. He briefs up his leaders before they execute a normal combat patrol. One sees a poster that mirrors the description given by the original post. Instead of ripping it down, he calls EOD [Explosive Ordinance Disposal], who discovers that it is rigged as an IED.

In another instance, a scout platoon leader -- in this case the scout platoon leader from 1-8 Cav. -- was given the mission to conduct sniper operations. He had never really executed a mission like it before. He looked on the CAVNET, where a commander from 1-9 Cav, in another part of the city, had posted notes and TTPs from employment of snipers over the past months. The Scout Platoon Leader from 1-8 was able to integrate what he had read from the CAVNET into his planning, preparation, and execution cycle.

How quickly can an officer expect a response after posting?

Immediately after posting the post, it will show up on the site. … I usually do not see a classic hoard of responses to a post. The knowledge is just posted. I know there is acknowledgement by the number of visits to the posts. If someone has something to add, it can be added in minutes, days, or weeks later -- creating a running record of different TTPs reference the same topic.

This was a fundamental change for me after I had run Platoonleader.org, which is designed for discussion. I could expect numerous posts because the community is so large and distributed. In the case of CAVNET, one of my baseline evaluation criteria is that it had to compliment a commander's battle rhythm rather than complicate it. Too many things we design take away from a commander's ability to command. It was my belief that if a commander could not get on, get information, post information, and get off within about 10 minutes, it would be useless. The software is designed around very simple messaging board software that pushes new information to the user if logged on by visual cues of new posts, or through e-mailing the post to a user if interested in a particular forum.

Does the Army monitor CAVNET? If so, how?

There are efforts throughout the Army to both replicate and harvest what is being posted on the CAVNET for the benefit of the entire Army. It's that final circle I was talking about.

... The staff judge advocate, concerned with the rules of engagement uses the CAVNET to get the message out to leaders of variations and vignettes associated with ROE. [S]imultaneously they provided overwatch on all the posting ensuring they were within the rules of engagement in terms of content -- lesson learned from operating a traditional "community of practice" [versus] a "knowledge transfer system" in combat. As the Platoonleader.org guy, I could allow off-mark posts knowing the community would police itself. In this environment, in combat, I am very cognizant of the postings because of the potential outcome of someone adopting another's tactics that may not be within the commander's intent or rules of engagement.