I actually have to second ebbor's suggestion, in part; a vital aspect of having a weapon of mass destruction is making sure it only goes off when and where we want it, and/or that we have the ability to disable or control it in some way- that is, so that it doesn't work on us.
-If our objective's profit at any cost, then we won't have much overhead to upgrade security measures, but do so inasmuch as budget allows- if there's an outbreak in any of our facilities, then we should at least have the ability to completely section off that area at a moment's notice.
-Mandate that any facility that possesses an Artificial Intelligence is required to have at least two such units on site, that all such units may only have partial (but overlapping) control of facility... facilities, and that all such units in any given facility must be subject to immediate remote shutdown by the proper access codes and/or security scans of the Director or Chief of Security of that facility.
-In case of loss of primary and emergency power in a facility, or full system reboot, the default state of blast doors and similar security measures should be "Closed"; such devices should not automatically open/disengage until access codes on the Director/CoS level are input.
-Begin development of a separate strain of the virus; attempt to strain out the neurological effects of T-β while retaining the regenerative properties.
-Begin development of a proprietary chemical compound that targets the T-α virus to rapidly break down the structures of infected cells.
ED: several small edits for clarity