The Journeyman

Gilman Louie, In-Q-Tel's CEO, is a man in the middle.

If you're lusting for a nostalgic fix of the dot-com days, when venture capitalists' portfolios were as hip as their offices, pay a visit to the Arlington, Va., digs of In-Q-Tel, the six-year-old not-for-profit that was established, according to its mission statement, "to help the CIA and the greater U.S. intelligence community to identify, acquire and deploy cutting-edge technologies." In-Q-Tel finds companies, introduces them to the agencies and then helps them broaden their appeal, both to the agencies and potential corporate customers.

The lobby of In-Q-Tel's office suite is meant to impress. A black Nelson marshmallow chair, a replica of the famous 1956 model, perches in front of a long picture window offering a panoramic view of the Northern Virginia business district, whose streets were awash in new tech money only a few years ago. The walls are relatively spare, save for some mounted press clippings about In-Q-Tel and its energetic chief executive, Gilman Louie.

But the offices lack the brashness of Silicon Valley yore. No foosball tables and free soda machines-at least not in public view. In-Q-Tel, after all, is something of a government hybrid with a foot in two worlds. As you check in at the receptionist station, you log your name into an electronic record. You're issued a visitor's pass, which you must surrender upon leaving. The operation feels rather . . . official.

Which is probably just how Louie wants it. Founded in 1999, In-Q-Tel now has 67 companies in its portfolio. The Sept. 11 attacks as well as an ensuing deluge of blue-ribbon-panel recommendations highlighted the intelligence community's weaknesses in using analytic technology. Today, that's In-Q-Tel's sweet spot.

The company is helping cultivate nearly three dozen young firms specializing in one or more of the following: search, categorization, collaboration and publishing; application integration; visualization; translation; geospatial intelligence; and design, simulation and modeling. Considering that every year the CIA provides In-Q-Tel with a list of the problems it most needs to solve, these categories give a decent picture of where U.S. intelligence managers feel they need to focus.

In a spacious boardroom decked out with wide-screen projectors and stage-quality lighting, In-Q-Tel trots out some of the most promising members of its flock. One, Visual Sciences LLC, headquartered in McLean, Va., received its first In-Q-Tel investment in March 2004. The company's showcase software can take tremendous amounts of data-from a Web site, a database-and turn it into a dizzying array of charts, graphs, diagrams and even maps. Let's say you run Citibank's Web site, and you want to know the pages of the site to which people most often navigate from the home page, and then what account functions they perform there. Push a button and you can "see" this movement, three-dimensionally if you like. Break the information down by day, by hour, by minute, and the program adjusts the visual display accordingly.

What's an intelligence agency's interest in something like that? Well, consider that with a few adjustments, the program also can identify, by Internet Protocol address, who visited the site and what actions they took. An image of the Earth from space appears on screen, and the program zooms down to a map of the United States. Whole swaths of the country light up, each pinpoint of light representing a unique IP address by location. Now, imagine plugging in to the program not Web traffic data but reported data on money transfers into and out of the United States. Imagine seeing it organized-visualized-by amount, sender, recipient, time of year, bank, location or any other of a host of variables the system can account for. Now you get the picture.

Another In-Q-Tel company also nestled in the niche of handling hordes of data. Agent Logic Inc. in Arlington, Va., builds "event management" software. Essentially, it's a meta program that looks across the full gamut of applications running on an organization's system and alerts individuals to important or curious events, some of which might happen so fast they are overlooked. A security analyst, for instance, can set up a "rule" in the program to notify him every time a ship enters a particular port or monitored space of ocean. The ship name can be checked against a watch list or a broader manifest showing which vessels are due in port on a given day. The analyst can set more detailed alerts as well, such as, "Tell me whenever a ship enters the space carrying chemical cargo." There might be a number of systems collecting all this data separately. Agent Logic fuses it together and displays it on screen in a format that looks as basic as the e-mail application you likely use every day.

Policymakers might be less interested in how tools like this work than the broader question of what problems they can solve. At the highest levels, Louie says, intelligence managers and decision-makers want to know whether tools are available that can share information, process large amounts of it, "find a needle in a haystack." For his part, Louie has become something of a provocateur in the hypersecretive intelligence field. Besides being a technology evangelist, he readily challenges the traditional excuse for not sharing information-that it might fall into the wrong hands. Sharing skeptics rarely ask the cost of not sharing, Louie says. He also dismisses the notion that federal intelligence agencies are the best sources of data on terrorism. Cops on the beat, seaport managers, airport security screeners-these are the best collectors, the nodes of an intelligence network, Louie believes. In that vast information pile, there is far more hay than needles. To get through it all requires more than the software In-Q-Tel companies build. Sharing across the entire network and taking calculated risks are what agencies need most, Louie says.

To get there, he is looking to the middle. Lately he's been taken by a theory that the greatest opportunity for change lies not with the green rookies and not with the seasoned veterans but with the journeymen, the em-ployees who've worked long enough to know the ropes but not so long that they're calling the shots (or have closed their minds to new ways of doing things).

It's a fitting fascination, on many levels. By his own admission, Louie is a journeyman in the world of intelligence. Six years in the game and he knows his way around but is no old hand. The companies in his charge are not newborns but aren't yet ready for prime time. Ultimately, the journeyman practitioners in the intelligence agencies, Louie thinks, are most likely to transform those agencies from bureaucratically oriented Cold War fighters to contemporary, technologically savvy intelligence networks. If Louie's ideas, and the In-Q-Tel portfolio, track with the long-term thinking of the intelligence agencies, it may be several years before we see wholesale change. But when it comes, it may resemble little that has gone before.