The Turnaround

Zalmai Azmi, the FBI's chief information officer, wants to take things slow and steady.

By late 2004, the writing was on the wall. The FBI's Virtual Case File, a much anticipated program to electronically organize and store mountains of investigative information, was coming unglued. The project was over budget. It was late. And a veritable revolving door of chief information officers and project managers meant that VCF was dangling in the wind with no one to save it.

Zalmai Azmi, who had been named the bureau's latest CIO in spring 2004, recalls FBI Director Robert Mueller asking him, "How did this happen, Zal? How did this happen?" Looking back now, after VCF has been scrapped and the FBI has wasted $170 million and three years of work, Azmi says the answers were obvious.

VCF was too ill-defined, it was too ambitious and it didn't receive appropriate managerial attention, Azmi says, ticking off the reasons as if he'd not only committed them to memory, but believed them with a certainty bordering on faith. To hear him tell it, you'd think the FBI's plans were doomed from the beginning. He remembers looking at the large computer upgrade project Trilogy, of which VCF was one component, and saying to himself, "There's no way we'll pull this off." With a network as big and as classified, not to mention as disconnected and antiquated, as the FBI's, the idea that the bureau would enter the Information Age in a walk was just preposterous to Azmi. After he became CIO, having served in an acting capacity for more than six months, he remembers officials asking, if we get you more money, can you save VCF? "No," he told them.

Azmi is in one of those rare and fortuitous positions for a federal bureaucrat. The projects now under his charge didn't fail on his watch. They suffered from problems encountered months, even years before he joined the bureau. So he can speak candidly about the lack of vision and coherency that led to the FBI's predicament. But Azmi also is straddling a precarious perch, because now he is expected to turn everything around. Under his guidance, the FBI is embarking on a new case management system, called Sentinel, and this time, the agency has laid out a slower, more deliberate approach. Sentinel will be implemented in four phases over more than three years. In the first phase, the Sentinel contractor will build a portal to provide searches of an existing case support system, authentication of users and access control to other FBI systems. It's a baby step.

Not surprisingly, Azmi likes to draw attention to what the FBI has done right since he arrived. And it might be easy to dismiss these achievements were they not such tangible improvements. In July 2004, the bureau established an Office of the Chief Information Officer to centrally manage all technology responsibilities, activities, policies and employees.

There had been CIOs at the bureau, but they controlled almost nothing. FBI divisions managed their technology investments on their own. They had varied processes and procedures. That has changed, and the move was so significant that it prompted the Justice Department inspector general to issue a modest commendation: "Despite the problems in completing Trilogy, it is important to note that the FBI is making progress to improve its IT management."

Azmi can point to other successes, too. The FBI now has an IT enterprise architecture, or blueprint, something it had lacked. Sentinel will be built to that architecture, which is standard practice for most projects. Azmi and his team have developed a methodology for tracking the FBI's technology assets over their lifetimes. He is preparing a "refresh," or replacement, of the bureau's PCs with newer versions. Thirty-three thousand desktops now have implemented a public key infrastructure, or PKI, which allows them to operate more securely over the Internet. And the FBI is beginning the Regional Data Exchange, a program that lets federal law enforcement agencies share information with state and local police forces. The PKI and data exchange projects were completed ahead of schedule and under budget, Azmi says.

The FBI also has launched the Investigative Data Warehouse, a searchable network of 50 databases from law enforcement and intelligence agencies. The agency has loaded the warehouse with almost 70 percent of the data in the bureau's old case systems, a figure that represents about 300 million paper records, Azmi says. The system now has 8,000 users and is available throughout the FBI's field offices, he adds.

Azmi has learned from past mistakes. "We should have developed [VCF] in phases . . . not asked people to swallow it all at once," he says. With new controls over who buys what systems, his staff now asks offices to explain their business case for buying any new technology. If a suitable equivalent to a desired piece of software already is loaded on their desktops, they'll be told to use it instead. And as for Azmi's staff, there's been something of a refresh as well. The most senior member of his management team has been on the job for about 19 months, he says. The newest has been around for less than three.

Students of failed technology projects will notice similar, even eerie parallels between the FBI's woes and those of other agencies. Indeed, the death spiral of VCF was precipitated by many of the same events that plagued the Internal Revenue Service's Business Systems Modernization. That project also suffered from a broad scope, ill-defined requirements and expectations, and an inattentive and frequently changing management staff. The IRS has a new CIO at the helm now as well, and like Azmi, he has focused on tackling the project in manageable pieces with a new staff.

Azmi followed a seemingly incongruous track to his current post. Before coming to the FBI, he was the CIO for the Executive Office for U.S. Attorneys, where he had occasion to meet Mueller. But before that, he was a project manager at the Patent and Trademark Office. In his jump from PTO to Justice, however, Azmi revealed himself as someone who enjoys demanding jobs with a high chance of failure and the risk of professional ruin. He came to the U.S. Attorneys as the office was scrambling for the Year 2000 changeover, which experts predicted might render the nation's computers essentially useless. Failure was not an option. The office made it through the changeover, and while he was there, Azmi helped upgrade systems and developed the agency's first enterprise architecture, among other achievements.

Asked whether he was intimidated by the job, Azmi displays a casualness that either is born of true confidence or is a very good bluff. "Not really," he says. "I wanted something different." He seems to want something different again and has upped the ante with the move to the bureau. Time will tell whether Azmi succeeds, or ends up regretting what he wished for.