- The Washington Times - Sunday, August 19, 2012

Months before the Army’s ill-fated 5th Stryker Brigade was to leave Washington state in the summer of 2009 for the war in Afghanistan, its commander became convinced that he needed a particular type of equipment to counter cunning bomb-makers.

The buzz was spreading among combat commanders that an analytical software platform named Palantir could soak up and analyze all sorts of battlefield data on the enemy, and then provide clues as to where roadside bombs — the No. 1 killer of Americans in Afghanistan — and their makers might be located.

Soldiers liked its ease of use, while some complained that the Army’s official intelligence-analysis system, the Distributed Common Ground System, was cumbersome.

Col. Harry Tunnell IV, who headed the 5th Stryker Brigade Combat Team, arrived in Kandahar, Afghanistan, in early July 2009 without the fully programmed network of Palantir he wanted.

Months later, after heavy casualties, some of his soldiers told headquarters that their mission would have gone better if Palantir had been with them from the start.

This is the untold story of how the Army, according to those close to Col. Tunnell, resisted his requests and relented only when he was in actual combat. The Palantir shipment did not arrive until February 2010. By the time it showed up and soldiers became acclimated to it, the brigade was in its last months of deployment.

The Washington Times obtained a January 2009 memo addressed to Army headquarters and carrying Col. Tunnell’s name specifically requesting Palantir six months before the brigade was dispatched to Afghanistan.

The memo says that Army-issued computer gear, including the common-ground system, exhibited an “intelligence gap” in sharing data company to company.

“The Palantir system allows the [company intelligence support teams] the ability to collect and mine data from various sources to combine into an actionable report for the commander,” said the memo, which was drafted at the Stryker Brigade’s headquarters at Fort Lewis, Wash.

Army headquarters at the Pentagon told The Times a different story. A spokesman said the Army never received an official request from Col. Tunnell until July 23, 2009, when the brigade had settled into Kandahar. The spokesman said the brigade obtained Palantir before it left for Afghanistan.

The Army said in a statement: “The memorandum you’ve provided regarding an Operational Needs Statement for Palantir Analytical Software in support of the 5th Stryker Brigade Combat Team [SBCT], 2nd Infantry Division, is inaccurate and unsigned. 5/2 SBCT deployed to Southern Afghanistan with the Army’s most advanced Network-Centric capabilities at that time, and Palantir.”

A different story

Col. Tunnell, and some of his men, tell a different version.

Col. Tunnell was interviewed last week by a staffer for Rep. Duncan Hunter, California Republican. Mr. Hunter, a Marine Corps veteran of the Afghanistan and Iraq wars, charges that the Army resisted allowing units to buy Palantir because it sees it as a threat to the institution’s own system, the common ground system.

Staffer Joe Kasper, who conducted the interview, quoted Col. Tunnell as saying that he made formal requests to the Army’s intelligence headquarters, known as G-2, but was told to use standard-issue networks as opposed to the off-the-shelf Palantir, made by Palantir Technologies.

“He asserts that formal requests for Palantir were definitely made, working through the chain of command before they arrived in Afghanistan,” Mr. Kasper said.

Col. Tunnell said the 2009 memo was generated at his headquarters, but he could not recall whether that is one he signed.

As to the Army contention that Col. Tunnell had Palantir from the start, Mr. Kasper said the unit took demonstrator models that Palantir Technologies provided for pre-deployment training, but they were not configured for threats in Afghanistan. New Palantir components did not arrive until February 2010.

Col. Tunnell wanted Palantir as early as 2008 after seeing it demonstrated on base.

Col. Tunnell’s written training guidance for the first three months of 2009 states: “The Brigade S2 [intelligence officer] will continue to coordinate for Palantir and integrate into the IWfF once received.”

The “IWfF” is the Infantry Brigade Combat Team Warfighters’ Forum, in which brigade commanders can talk over issues with senior Army leadership.

Col. Tunnell’s version provided to Mr. Hunter’s office seems to be bolstered by a survey conducted for the Army’s Rapid Equipping Force, which provided the new Palantir platform.

In May 2010, two months before the Stryker brigade left Afghanistan, one of its company commanders complained in a report: “It is a shame that the Army had not allowed us to use Palantir from the beginning, I think we would have had a more successful deployment across [the brigade].”

The captain, whose name is being withheld at the request of an Army official, said that when his company arrived, it inherited no intelligence or historical data in its assigned province.

“Our efforts in targeting were extremely hindered by our lack of historical data,” the officer wrote.

In Kandahar, a special-operations detachment exited in November, leaving all its data on intelligence sources and the enemy. The Stryker unit had to sift and file through the material.

“This took hours, if not days, just to target certain IED cells operating in the area,” the officer said. “Palantir would have cut the research time down to just a few minutes.”

How it works

In one example of how Palantir helped, intelligence reported hearing a Taliban commander saying he would be crossing a river into Stryker’s area with “pots” — a code name for IEDs. The company inserted the information into Palantir and immediately saw that another company had posted information on an IED “facilitator” who lived near the river and helped transport “pots.”

The entire brigade knew the Taliban commander was on his way, and there was a facilitator on his route.

Palantir has given my [intelligence unit] the ability to organize raw intel data, analyze what that data means, edit and organize possible cells, offers possible links to other enemy organizations and helps us focus on certain individuals if you are trying to track a certain flow,” the captain wrote in a report.

Another soldier, a brigade intelligence analyst, noted, “The only complaint I have is that we were not able to deploy initially with Palantir. We were trained and used the software successfully at [the National Training Center] and thought we would have it available to us upon deployment, so it was upsetting that we didn’t get it until much later in the deployment.”

A third soldier who specialized in acquiring intelligence from Afghan sources wrote, “There is no better system for ingesting all forms of electronic data and linking it all together. Nothing out there comes close.”

The Washington Times first reported on an unseen bureaucratic battle within the Afghanistan War.

The Times detailed how soldiers asked for Palantir and got rebuffed by staffers at Army headquarters who think the common-ground system, and other software, was all they needed. An internal Army email from a senior official asked colleagues about the best way to “turn off” a request for Palantir from the 82nd Airborne Division, where officers, as with Col. Tunnell, were hearing good things about the software widely used by special-operation forces, but not the conventional Army.

The Times reported on how the Army ordered the destruction of a favorable field evaluation of Palantir completed this April and replaced it with another test report critical of the software. The Times also obtained a copy of an operational test of the Army’s common-ground system. The report to Gen. Raymond T. Odierno, Army chief of staff, concluded that the network suffered “significant limitations,” was unreliable and vulnerable to cyberattack.

The House Oversight and Government Reform Committee has launched a probe into the entire sequence of events.

Rocky year in Afghanistan

Even without the Palantir snafu, the 5th Stryker Brigade, named after the eight-wheel Stryker armored vehicle, was destined to spend a rocky year in Afghanistan.

Most of its preparation was for Iraq. It learned late in the training game that the new destination was Afghanistan — and the south region where the Taliban rose up in the 1990s and still counted on plenty of allies.

During the brigade’s tour, five of its soldiers were charged with killing innocent Afghan civilians. A subsequent Army investigation harshly criticized Col. Tunnell for displaying an aloof leadership style and for failing to follow the command’s new counterinsurgency strategy to protect the civilian population. He focused instead on killing insurgents.

The Army general who conducted the investigation in 2011 said he would have fired Col. Tunnell if he still commanded the brigade. Col. Tunnell, a West Point graduate, was transferred to an Army recruiting command at Fort Knox, Ky., and is seeking to retire.

Other officers who followed the 5th Stryker brigade in Afghanistan came to the same conclusion that the Army needed a better computerized intelligence network to identify the enemy.

They included Army Lt. Gen. Michael Flynn, then the top intelligence officer in Afghanistan.

As the Stryker brigade was preparing to leave Afghanistan in July 2010, Gen. Flynn sent an urgent memo to the Pentagon.

Gen. Flynn did not single out a specific system for criticism, but Pentagon officials say he was bemoaning the performance of the Army’s common-defense system. He demanded that the Pentagon provide a system able to process, link and display data for all units to see.

“Current tools do not provide intuitive capabilities to see the relationship between a wide variety of disparate sets of information,” Gen. Flynn said in the memo, first reported by Politico.

“The enemy is able to take advantage of his ability to hide in plain sight in the population because we have been unable to fully exploit the information/intelligence we already have. Detainees with existing connections to the insurgency have been released because we could not fully understand or exploit the information we held.”

Gen. Flynn, named director of the Pentagon’s Defense Intelligence Agency last month, was seeking the kind of analysis for which Palantir was becoming famous.

Surprised by IED threat

In December 2009, the Army Times published an extensive article by Sean Naylor on the 5th Stryker brigade’s first six months in Afghanistan and the hits it took from IEDs.

That October, seven soldiers died when an IED blew up a Stryker vehicle. By December, 21 soldiers from a single battalion had been killed, the highest for any Army battalion in Afghanistan.

“The extent of the IED threat was a surprise to us all,” the Army Times quoted one commander as saying. “The enemy we faced in the Arghandab [district] adapted to our TTPs [tactics, techniques and procedures] faster and more effectively than anyone expected.”

An Army spokesman provided The Washington Times with this Palantir timeline: Col. Tunnell’s official request was submitted July 23, 2009; the Army approved it two months later; a contract was awarded Dec. 21; Palantir arrived in the country on Feb. 18, 2010, the brigade’s eighth month in combat.

Mr. Kasper, the congressional staffer, said Col. Tunnell stopped asking for Palantir until he arrived in Afghanistan, where a request from the field would have a better chance of approval.

“He made a deliberate decision to request in theater and use the theater as the backdrop for the request,” Mr. Kasper said.

Rep. Hunter added, “Eventually, the Army had no other option but to give in and say ‘yes,’ even though it probably didn’t want to. The circumstances called for different technology and the command was asking for it, repeatedly, in fact.

“Bottom line: the command should have got what it needed, when first requested, to better protect soldiers in harm’s way.”