Documents
Tips for a Successful Quick Reaction Capability
Jan. 19, 2018
TOP SECRET//COMINT//NOFORN//20320108
Friday, October 26, 2007
(U//FOUO) Tips for a Successful Quick Reaction Capability
by
(U) Collaboration is a way to get the job done.
(TS//SI//NF) Collaboration is key when you are tasked with a high priority, laborintensive project. When the Iran Government Branch (S2E41) and the Iran Leadership
Branch (S2E43) were asked to participate in a Quick Reaction Capability (QRC) for
Iranian President Ahmadi-Nejad’s late September 2007 trip to New York, we went into
action to coordinate an effective approach to the task. In our case, QRC meant a shortterm increase in special collection to provide time-sensitive information to a defined
customer set. Fortunately, we were not new to QRCs, and our experience has taught us
the value of collaboration. We had learned the three most important tips for successfully
carrying out a QRC: plan ahead, delegate, and share.
(U) Plan Ahead
(TS//SI//NF) In response to the QRC request, we immediately began to coordinate across
the division with the efforts of various other partner offices throughout the Agency as
well as with other Intelligence Community agencies. To prepare, S2E41, S2E43, the
BLARNEY office, and Tailored Access Operations began working with Oversight and
Compliance and the Office of General Counsel to ready the necessary legal documents
and obtain authorization for special FISA collection on the Iranian delegation during their
stay in New York for the 62nd United Nations General Assembly.
(TS//SI//NF) Concurrently with obtaining the “ok” for the special collection, we had to
ensure that the proper procedures would be in place to be able to efficiently tackle the
anticipated influx of traffic. The Human Language Technology (HLT – S23) office
provided numerous technologies to aid in the coverage:
• Speech Activity Detection was run on all incoming traffic to help identify cuts
with little-to-no speech to prevent our having to listen to dead air.
• Speaker Identification was run to help pinpoint people of significant interest,
including the Iranian Foreign Minister.
• VoiceRt phonetic keyword searches were developed and run to help uncover key
target information, such as the passing of email addresses and discussion of
prominent individuals.
TOP SECRET//COMINT//NOFORN//20320108
TOP SECRET//COMINT//NOFORN//20320108
Friday, October 26, 2007
(U//FOUO) Tips for a Successful Quick Reaction Capability
by
(U) Collaboration is a way to get the job done.
(TS//SI//NF) Collaboration is key when you are tasked with a high priority, laborintensive project. When the Iran Government Branch (S2E41) and the Iran Leadership
Branch (S2E43) were asked to participate in a Quick Reaction Capability (QRC) for
Iranian President Ahmadi-Nejad’s late September 2007 trip to New York, we went into
action to coordinate an effective approach to the task. In our case, QRC meant a shortterm increase in special collection to provide time-sensitive information to a defined
customer set. Fortunately, we were not new to QRCs, and our experience has taught us
the value of collaboration. We had learned the three most important tips for successfully
carrying out a QRC: plan ahead, delegate, and share.
(U) Plan Ahead
(TS//SI//NF) In response to the QRC request, we immediately began to coordinate across
the division with the efforts of various other partner offices throughout the Agency as
well as with other Intelligence Community agencies. To prepare, S2E41, S2E43, the
BLARNEY office, and Tailored Access Operations began working with Oversight and
Compliance and the Office of General Counsel to ready the necessary legal documents
and obtain authorization for special FISA collection on the Iranian delegation during their
stay in New York for the 62nd United Nations General Assembly.
(TS//SI//NF) Concurrently with obtaining the “ok” for the special collection, we had to
ensure that the proper procedures would be in place to be able to efficiently tackle the
anticipated influx of traffic. The Human Language Technology (HLT – S23) office
provided numerous technologies to aid in the coverage:
• Speech Activity Detection was run on all incoming traffic to help identify cuts
with little-to-no speech to prevent our having to listen to dead air.
• Speaker Identification was run to help pinpoint people of significant interest,
including the Iranian Foreign Minister.
• VoiceRt phonetic keyword searches were developed and run to help uncover key
target information, such as the passing of email addresses and discussion of
prominent individuals.
TOP SECRET//COMINT//NOFORN//20320108
TOP SECRET//COMINT//NOFORN//20320108
(TS//SI//NF) One of the branch’s senior language analysts, an Excel spreadsheet guru,
put together a detailed list of all 143 delegates, including not only their names and
positions in the Iranian government, but their passport numbers, dates of birth and more.
In addition, she included the Iranian President’s and Foreign Minister’s schedules.
Collection information and identifications of the various hotel rooms and cell phones
were also updated as soon as they were discovered. The analyst kept it updated
throughout the day and provided it to everyone involved with the QRC, including the
FBI.
(TS//SI//REL TO USA, FVEY) Meanwhile, language analysts from S2E41, S2E43, and
S23 combined forces to take on the challenges of the significantly increased collection.
The voice language analysts relied on one senior voice language analyst who had a knack
for finding the way to best organize the traffic using NUCLEON.
(U) Delegate
(TS//SI//REL TO USA, FVEY) With a limited number of resources and an abundance of
traffic to scan, we had to find a way to divide the work so that nothing was missed or
duplicated. We were scanning approximately 2000 voice cuts a day, so organization and
delegation were key. The team chief created MORPHEON rules, shared queries, shared
filters, and tabs to organize and search the voice traffic. The team agreed to divide the
traffic by using the filters and creating tab groups in addition to using color-coding, and a
limited set of NUCLEON dispositions, e.g., if the SRI (signal-related information) data
for a voice cut was red, it meant “to do,” and if blue it was “new,” and if black it had
“already been done.” That way, no matter what query or filter a language analyst was
using, everyone else could “see” what had been done and what was still left to do.
(TS//SI//REL TO USA, FVEY) Some analysts scanned and prioritized while others
transcribed and gisted…everyone moved between the two jobs throughout the day as
needed. This method made for very efficient scanning, which is paramount when five or
six language analysts are working the same traffic at the same time.
(U) Share
(TS//SI//REL TO USA, FVEY) All the language analysts kept in contact throughout the
day and passed on their findings, analyses and highlights via email, phone and often
impromptu gatherings. With the group working extended hours, from 0400-2300, the last
analyst to leave would gather together the highlights from that day’s findings and share it
with the other offices involved. Sharing was crucial because no two language analysts
scanned the same traffic, so we had to collaborate in order to provide the most
comprehensive information to the customers.
TOP SECRET//COMINT//NOFORN//20320108
TOP SECRET//COMINT//NOFORN//20320108
(TS//SI//NF) One of the branch’s senior language analysts, an Excel spreadsheet guru,
put together a detailed list of all 143 delegates, including not only their names and
positions in the Iranian government, but their passport numbers, dates of birth and more.
In addition, she included the Iranian President’s and Foreign Minister’s schedules.
Collection information and identifications of the various hotel rooms and cell phones
were also updated as soon as they were discovered. The analyst kept it updated
throughout the day and provided it to everyone involved with the QRC, including the
FBI.
(TS//SI//REL TO USA, FVEY) Meanwhile, language analysts from S2E41, S2E43, and
S23 combined forces to take on the challenges of the significantly increased collection.
The voice language analysts relied on one senior voice language analyst who had a knack
for finding the way to best organize the traffic using NUCLEON.
(U) Delegate
(TS//SI//REL TO USA, FVEY) With a limited number of resources and an abundance of
traffic to scan, we had to find a way to divide the work so that nothing was missed or
duplicated. We were scanning approximately 2000 voice cuts a day, so organization and
delegation were key. The team chief created MORPHEON rules, shared queries, shared
filters, and tabs to organize and search the voice traffic. The team agreed to divide the
traffic by using the filters and creating tab groups in addition to using color-coding, and a
limited set of NUCLEON dispositions, e.g., if the SRI (signal-related information) data
for a voice cut was red, it meant “to do,” and if blue it was “new,” and if black it had
“already been done.” That way, no matter what query or filter a language analyst was
using, everyone else could “see” what had been done and what was still left to do.
(TS//SI//REL TO USA, FVEY) Some analysts scanned and prioritized while others
transcribed and gisted…everyone moved between the two jobs throughout the day as
needed. This method made for very efficient scanning, which is paramount when five or
six language analysts are working the same traffic at the same time.
(U) Share
(TS//SI//REL TO USA, FVEY) All the language analysts kept in contact throughout the
day and passed on their findings, analyses and highlights via email, phone and often
impromptu gatherings. With the group working extended hours, from 0400-2300, the last
analyst to leave would gather together the highlights from that day’s findings and share it
with the other offices involved. Sharing was crucial because no two language analysts
scanned the same traffic, so we had to collaborate in order to provide the most
comprehensive information to the customers.
TOP SECRET//COMINT//NOFORN//20320108
TOP SECRET//COMINT//NOFORN//20320108
(U) The Iranian President signs a guest book at the UN, September 24, 2007 (Reuters)
(U) Results
(TS//SI//REL TO USA, FVEY) Our QRC efforts had several positive outcomes:
1. The discovery of several new Iranian Foreign Ministry and leadership leads.
2. Several reports on topics ranging from Ahmadi-Nejad’s sentiments about his
speech at Columbia University to last minute plans for the delegation to go to
Latin America.
3. Several language analysts are now working with the Social Network Analysis
office to better understand the connections between many of the Foreign Ministry
and Leadership delegates.
(U//FOUO) Although we are still delving into all of the data gleaned from the QRC, and
there are still many uncertainties, one sure thing is it would not have been nearly as
successful without the collaborative spirit.
TOP SECRET//COMINT//NOFORN//20320108
TOP SECRET//COMINT//NOFORN//20320108
(U) The Iranian President signs a guest book at the UN, September 24, 2007 (Reuters)
(U) Results
(TS//SI//REL TO USA, FVEY) Our QRC efforts had several positive outcomes:
1. The discovery of several new Iranian Foreign Ministry and leadership leads.
2. Several reports on topics ranging from Ahmadi-Nejad’s sentiments about his
speech at Columbia University to last minute plans for the delegation to go to
Latin America.
3. Several language analysts are now working with the Social Network Analysis
office to better understand the connections between many of the Foreign Ministry
and Leadership delegates.
(U//FOUO) Although we are still delving into all of the data gleaned from the QRC, and
there are still many uncertainties, one sure thing is it would not have been nearly as
successful without the collaborative spirit.
TOP SECRET//COMINT//NOFORN//20320108