I have long been a supporter of running multiple parallel pilots for major IT projects. The cost of funding the additional pilots should be recouped in better specification and a more fit-for-purpose delivery. I think this was used for the congestion charge (first time around). There are some details of the Technical Design Study at 6.3 / page 56/57 of this long background report (pdf) on the charge itself.
I’m not sure how the conservatives think that this approach will weaken the reliance on major vendors: they are often the only players with the required scale. That’s a function of the projects, rather than of the vendors.