At a Wall Street Journal conference late last year, White House Chief of Staff Rahm Emanuel was quoted as saying, “You never want a serious crisis to go to waste.”

Emanuel was talking about the oil crisis of the 1970s, when, in his view, the U.S. missed an opportunity to begin addressing the nation’s long-term energy needs. It wasn’t a lack of ability, resources or skill that stymied progress on the energy front but a collective reluctance to use crisis as a catalyst for change.

Emanuel could have been talking about today’s Defense Department.

The U.S. Defense Department has a history of responding to crises superbly, with great agility and effectiveness, but too often can be counted on to lose much of the innovation gleaned via crisis — especially those lessons related to materiel and procurement. Some defense officials recognize this history, and they have formed groups, such as the Joint Unmanned Air Systems Center of Excellence, to make sure the innovations of the recent revolution in unmanned air systems are retained.

Innovation evaporation is not a new phenomenon. Take, for example, the Springfield rifle produced for the Union by the military-run Springfield Armory in Massachusetts. The Springfield Model 1861 was the world’s first mass-produced rifled musket, with nearly 800,000 manufactured during the Civil War. Despite its effectiveness, the government allowed the Springfield to fall by the wayside after the war. The rifle had not gone through the Department of War’s traditional procurement process. The armory developed it quickly to meet the needs of a crisis. After the war, the U.S. government reverted to less advanced models delivered under the standard procurement system, and it would be nearly 30 years before the Springfield was fully reintegrated into the U.S. arsenal.

From the Civil War to the 21st century the form factors have changed dramatically, but the challenges of integrating innovation remain strikingly similar.

I realized this last year while participating in a working session on the future of unmanned warfare sponsored by Harvard University and the consulting firm Booz Allen Hamilton, where I work. A group of civilian Defense Department leaders and active-duty and retired flag officers gathered in Cambridge, Mass., to discuss the topic over two days. These officials were used to dealing with the challenges of unmanned warfare, but they rarely had the opportunity to compare experiences. That meeting, and a series of subsequent conversations by the participants, culminated in the publishing of a report, “Unmanned and Robotic Warfare: Issues, Options, and Futures,” describing the chaos that exists around efforts to integrate robotics into a military framework, but suggesting a potential path forward.

Among issues raised during the session was how difficult it is to contemplate a future in which platforms will be dependent on future technological advances. It would be easy to underestimate what might be possible and, with the pressure of an ongoing crisis, equally hard to break out of the “now.” In light of this, policymakers need to refocus their thinking toward required operational capabilities rather than proposed evolutionary changes to platforms. That is one of the issues being tackled by the Joint Unmanned Air Systems Center of Excellence.

Another question raised was this: Why can’t we seem to incorporate the platform, sensor and architecture innovations gleaned in time of war into the baseline of production so the benefits remain with us even after the current crisis is over? This issue strikes me as being systemic in nature, not just applicable to robotics or unmanned air systems. It might be more apparent in those areas because of the inherent rapid pace of innovation, but it’s an issue that cuts across disciplines.

Evolutionary advances in unmanned air systems are perhaps less intriguing than visions of swarming fully-autonomous killer-robot bumble bees as described on the Web site environmentalgraffiti.com. The truth is, as one participant at Harvard noted, “80 percent of what is afloat, in the air, and on the ground today will be there 15 years from now.”

Retired Air Force Gen. John Jumper, a former Air Force chief of staff and a participant in the Harvard working session, told me about his experience advancing the Predator during the Kosovo war. The Air Force and its contractors has previously retrofitted several Predators with laser-targeting devices that enabled them to illuminate ground targets during air-to-ground strike operations. It was a very clear-cut advance in capability, born of necessity and proven in battle. On arriving at his post-Kosovo command, Jumper learned that the innovation had not been adopted more broadly and that there was no intention of including it as part of the Predator program of record. It was being ignored. The story is just too uncomfortably resonant with the story of the Springfield rifle. Why would this happen — and what can be done to facilitate the formal adoption of lessons learned? In a 2003 posting on the StrategyPage Web site, military analyst James Dunnigan observed that the Defense Department is better at identifying lessons than truly learning from them. He noted that “‘Lessons learned’ often become twisted to support pet projects” and that often the same “lesson” is used to support or reinforce widely divergent views.

Although politics might play a role in this arena, it’s counter to my own observations. From experience, the services rally to support troops in battle, especially during times of crisis. Also, I’ve learned never to attribute to conspiracy that which can be explained by bureaucracy. And in this instance, the suspect bureaucracy might, in fact, simply be the regulatory structure of acquisition programs, the incentives and expectations of program managers, and the tension between requirements definition and materiel solutions.

After more than 30 years in this business, I can’t in good conscience simply fall back to the tired refrain that we need acquisition reform — e.g., better, quicker, more streamlined processes that deliver results in a more timely and more effective manner. Of course we need those things. However, it’s a challenge comprised of many moving parts, and, even in best of circumstances, it is a problem that is not going to be solved quickly.

But small changes are possible. In the case of unmanned air systems and robotics, the operational models and uses are evolving at breakneck speed. One small change that might help drive larger benefits is to connect programs of record directly to the end-users and measure, in at least one dimension of performance of the program, how well program officials incorporate user feedback directly into the program.

In a recent conversation with UAV pilots at Creech Air Force Base, Nev., we talked about this problem. I was hoping to gain some insight into the process they used to feed improvements back to the platform programs. The conversation was slow in starting because they really had little connection to the program of record and didn’t feel like they could provide much insight into how proposed changes were evaluated and adopted. What they did convey, through story after story, was an image of a frontier environment where resourcefulness and creativity are key ingredients to satisfying high operational tempo requirements.

They described how new software was developed on the fly to support the needs of specific mission types. These were debugged and integrated in a live environment, used extensively, then set aside in favor of new mission needs. When the need arose again for that first innovation, there was often no one left who remembered how to incorporate or use it, much less what its limitations were. But the choice they faced was this: Develop something new themselves in days, or try to feed a request back to the program of record and have it in years, if ever. How many refinements have been developed, proven in battle and lost?

As in any frontier, another major issue is scarcity of resources. The competition between, say, close-air support missions and the ISR teams is intense and often translated into resistance to any improvements. Why add a close-air support capability to your ISR asset when it only increases the likelihood that you’ll lose control of that asset in favor of some new mission?

The operations centers also are not part of the program of record, despite the fact that improvement to the overall functioning of the unmanned air system often requires enhancements to the platform and the operations center because they are dependent on each other. There is a very real concern that new deliveries from the program of record will lack — and might be incompatible with — homegrown upgrades. And since everything is being used all the time in support of real-world combat operations, there is no traditional cycle of downtime for block upgrades.

While there hasn’t been enough time for a traditional community of users to form in the unmanned realm, there is clear recognition of the competing needs of the various user groups. The air system crew, the exploitation team and the support team — or “the triad of pain,” as they’re known among themselves — compete and cooperate in a high-demand, real-time arena, an arena that is constantly placing more demands on already stressed unmanned air systems. And as usage has expanded, the frontier has fragmented from a loosely coherent squadron, in which a local innovation became the de facto standard, to an array of six separately operating squadrons, or “six little tribes” as they’re known locally. Each has its own tactics, techniques and processes, with more evolving all the time.

All of this evolution takes place in a desolate environment with facilities baking in the 100-degree daily heat. This is a “meat grinder” environment, and although the pilots and users point out that they were in no way equating it to what the “boots on the ground” are experiencing in Iraq and Afghanistan, it’s clear that life in the Nevada desert is a far cry from a Vegas vacation.

Part of the solution might be to build an online unmanned air system user community. This has been done elsewhere with some success. In 2007, Dell launched the IdeaStorm Web site, which it described as “our way of building an online community that brings all of us closer to the creative side of technology by allowing you to share ideas and collaborate with one another.” This Web 2.0-based community quickly came alive with users helping users, just for the “psychic income” of sharing their knowledge. In my own firm, we’ve embraced this approach through a suite of Web 2.0 tools called Hello.bah.com, consisting of wikis, blogs, discussion forums and tag clouds serving user-defined communities.

Marine Corps Gen. James E. Cartwright, vice chairman of the Joint Chiefs of Staff, embraced the value of Web 2.0 in his former role as commander of Strategic Command. He sent a note to his noncommissioned officers, saying: “The metric is what the person has to contribute, not the person’s rank, age, or level of experience. If they have the answer, I want the answer. When I post a question on my blog, I expect the person with the answer to post back. I do not expect the person with the answer to run it through you, your [officer in charge], the branch chief, the exec, the Division Chief and then get the garbled answer back before he or she posts it for me.”

A Web 2.0-based UAS community, appropriately secured, would allow UAS teams to share ideas and best practices while also allowing the programs of record to test new ideas and, especially, to listen to what the community says. Such a community could also provide a “corporate memory” to ensure that mission-specific refinements could be reused and further refined by all when needed. And for policymakers, it would provide insight into the responsiveness of programs of record to real needs.

It’s an exciting time to be working in robotics and unmanned systems. Ideas are free-flowing, and it’s still too early for a morass of “standards” to stifle creativity and innovation. The operational challenges that those “standards” are intended to tame are real and compelling, but there are equally real alternatives to traditional top-down policymaking to solve them — and actually benefit from them.

I believe the leaders in this community will have the vision and courage to embrace these nontraditional approaches to accelerate the continued evolution of this powerful new capability.

Art Fritzson leads the Defense IT business at Booz Allen Hamilton, where he is a vice president.