The story...

An example of how the air defense network normally responds to domestic emergencies is illustrated by the well-reported 1999 case of Payne Stewart's Lear jet. When the golfer's jet failed to respond to air traffic controller communications, F-16 interceptors were quickly dispatched. According to an Air Force timeline, a series of military planes provided an emergency escort to Payne's stricken Learjet starting about 20 minutes after contact with his plane was lost.
http://911research.wtc7.net/planes/analysis/norad/

On October 25, 1999, at 9:33 a.m. air traffic controllers in Florida lost touch with a Learjet carrying golfer Payne Stewart and several companions after it left Orlando headed for Dallas, Texas. Nineteen minutes after Air Traffic Control realized something was wrong, one or more US Air Force fighter jets were already on top of the situation, in the air, close to the Learjet.
http://www.prisonplanet.com/hijackology_101_did_norad_send_the_suicide_jets_on_911.html

According to published accounts, fighter jets intercepted Stewart’s jet in either 15 or 21 minutes after his plane first lost contact.
Page 200
9/11 Synthetic Terror: Made in the USA
Webster Tarpley

Our take...

These accounts tell a similar story: only 20 minutes after contact is lost, and the air traffic controllers realise there’s a problem, Air Force jets are on the scene. This sounds impressive, but unfortunately it isn’t true. A quick look at the NTSB accident report reveals why. Here's the timeline.

"At 0933:38 EDT (6 minutes and 20 seconds after N47BA acknowledged the previous clearance), the controller instructed N47BA to change radio frequencies and contact another Jacksonville ARTCC controller. The controller received no response from N47BA. The controller called the flight five more times over the next 4 1/2 minutes but received no response.

About 0952 CDT,7 a USAF F-16 test pilot from the 40th Flight Test Squadron at Eglin Air Force Base (AFB), Florida, was vectored to within 8 nm of N47BA. About 0954 CDT, at a range of 2,000 feet from the accident airplane and an altitude of about 46,400 feet, the test pilot made two radio calls to N47BA but did not receive a response".
www.ntsb.gov/Publictn/2000/AAB0001.htm

Looks good at first, but read it carefully and you'll notice a change of time zone, from Eastern to Central time. CDT is one hour on from EDT, so contact was regarded as lost at around 09:38, and the fighter didn't get to within 2000 feet of Stewart’s jet until 10:54. That's roughly 76 minutes from the controllers realising there’s a problem, to intercept taking place.

Press reports from the time give more details.

The FAA said air traffic controllers lost radio contact with the plane at 9:44 a.m...

Pentagon officials said the military began its pursuit of the ghostly civilian aircraft at 10:08 a.m., when two Air Force F-16 fighters from Tyndall Air Force Base in Florida that were on a routine training mission were asked by the FAA to intercept it. The F-16s did not reach the Learjet, but an Air Force F-15 fighter from Eglin Air Force Base in Florida that also was asked to locate it got within sight of the aircraft and stayed with it from 11:09 a.m. to 11:44 a.m., when the military fighter was diverted to St. Louis for fuel.

Fifteen minutes later, four Air National Guard F-16s and a KC-135 tanker from Tulsa were ordered to try to catch up with the Learjet but got only within 100 miles. But two other Air National Guard F-16s from Fargo, N.D., intercepted the Learjet at 12:54 p.m, reporting that the aircraft's windows were fogged with ice and that no flight control movement could be seen. At 1:14 p.m., the F-16s reported that the Learjet was beginning to spiral toward the ground.
http://www.washingtonpost.com/wp-srv/national/daily/oct99/crash26.htm

Putting these together with the NTSB report suggests the following points.

First, it takes time before ATC consider they’ve lost contact with a plane. The absence of any radio response was first noted at 9:34, but the controller continued trying to make contact for another four minutes, and the press report suggests contact wasn’t considered lost until six minutes after that, ten minutes after the problem was noted.

And second, NORAD don’t always have the capability to respond in a few minutes. The intercept didn’t begin for another 24 minutes, actually a fast response because the plane was already in the air.

To be fair, if the first fighters had been closer (as they were on 9/11) then the response time would have been better. And 911Research do their best to make even 76 minutes seem an insignificant length of time:

83 minutes elapsed between the time that Flight 11 veered off course and the Pentagon was hit, and 112 minutes elapsed between the time that contact was lost with Flight 11 and Flight 93 crashed.
http://911research.wtc7.net/planes/analysis/norad/

This seems an odd way of accounting. Why does the time Flight 11 veer off course make a significant difference to the intercept time for Flight 77? Or Flight 93? 

Surely a more reasonable approach is to extrapolate from the “likely takeover time” (the earliest time anyone would have known about the hijacking) until the point each plane reached its final target. Which gives us the following elapsed times: Flight 11 (8:14 to 8:47 - 33 minutes), Flight 175 (8:42 to 9:04 - 22 minutes), Flight 77 (8:51 to 9:38 - 47 minutes), Flight 93 (9:27 to 10:04 - 37 minutes). See http://www.9-11commission.gov/report/911Report_Ch1.htm

(Note that some will disagree with these times, in part because we’ve taken them from the 9/11 Commission Report. 911Research, for instance, say that Flight 77 went off course at 8:46 and so it could be argued that we should start the clock then. Look at the full facts and we’re being generous, though: the time we’re using corresponds with the last reported contact with the Flight 77 pilots, and the first suspect action after that, an unauthorised turn south, wasn’t noted until 8:54. It could just as well be argued that this is the first notification of a problem, meaning our timings provide 3 minutes more intercept time than was actually available).

In the Stewart case, as we’ve seen, there was a 10 minute gap between the initial problem being observed (ATC getting no radio response) and a decision being made that contact was lost. An intercept didn’t begin for another 24 minutes, and if planes were scrambled from the ground then that could have taken longer: fighters are typically on 15 minute alert, but we’ll be generous and say it took 5.

If this were repeated on 9/11, then that’s 39 minutes, plus the time it takes the plane to reach the target, before an intercept can occur.  Which suggests there wasn’t enough time to reach Flight 11 and 175 before they hit the towers, and even intercepting Flight 77 would have been very difficult. The Stewart case simply doesn’t support the idea that the 9/11 flights should have been intercepted, then: in fact, quite the opposite.