I think you've brushed up against the truth about GAC and calls for its early engagement here, Avri. 
It's a sham at worst, and well-intentioned self-delusion at best. When I was on the GNSO Council 6-7 years ago, there was a period in which the GAC sent a liaison, and while the person tried hard, he contributed little to the policy discussions and whatever role he had in the GAC itself, did not seem to affect the disconnect between the GNSO and GAC. 

In several other working groups, notably Whois, GAC representatives often failed to show up and even when they did, there was no evidence that discussions or consensual agreements reached in the WG had any effect on how the GAC as a whole reacted in the end. Usually this is a problem with the dominating presence of the US representative, who consistently wants GAC to dictate policy in response to political pressures from lobbying groups (trademark, LEAs) or fears about how other governments will react to ICANN policies. 

My other main experience with GAC was the MAPO (morality and public order) cross-community working group, in which a very broad and representative slice of the GNSO and ALAC interacted with a few GAC members. There was only one active GAC member (Bertrand). And in the end, the GAC as a whole completely disregarded any putative commitments their participants made in that working group, and later asserted its right to bilaterally negotiate with the board regarding objections and vetos of TLD applications. (culminating in the stormy Brussels meeting)

Part of the problem is structural, in that GAC participants are not "free agents" who can make binding commitments in email list-based working groups - they are merely agents of a hierarchical principal and must go back up the chain of command to get anything approved. Governments mode of operation is literally incompatible with the bottom up process. Literally. And many of them have multiple responsibilities and are not expert on the policy issues. 

So the record belies any naïve view that earlier engagement with the GAC with solve any of the problems associated with its interventions. 

> -----Original Message-----
> 
> anecdotally, I beleive that people have often tried to engage GAC with
> little success.  I am not sure how true this is beyond the anecdotal and
> so would like to find out more.  How often has the GAC been invited and
> how have they responded?  I beleive it is possible we will find that the
> invitations and the modes of participation are a mis-match and we need
> to explore the issue of how the GAC can participate in the early stages
> of the process.  We have certainly seen over the years, an increase in
> the cross participation between other ACs and the SOs. We have even seen
> some GAC participation, but not at the same levels.
> 
> I think it is ok to ask about my 'ideal outcomes' for ATRT2.  Overall I
> think ATRT1 gave us the impression that the AOC review process might be
> useful and might work.  I hope we come out of ATRT2 with a view as to
> how well ATRT1 really worked and come out with reviews and
> recommendations that represent improvement in Accountability and
> Transparency at ICANN.
> 
> Thanks for the question.
> 
> avri
> 
> 
> On 22 Jan 2013, at 12:19, Maria Farrell wrote:
> 
> > Hi, Avri and Marie Laure,
> >
> > My question is about the Government Advisory Committee's future role.
> >
> > The GAC's report of its High Level Meeting in Toronto said it wanted
> ATRT2 to look at: "Enabling engagement of the GAC as early as possible,
> and at various levels, within the ICANN policy development process".
> >
> > What form do you think greater GAC engagement might take earlier in
> the process, and how would you try to ensure its engagement in the GNSO
> and at the same time protect the multi-(equal)-stakeholder process?
> >
> > I hope this question is within scope, i.e. that it's ok to ask you
> what your 'ideal outcomes' from the ATRT2 might be on this issue.
> >
> > Thanks and all the best, Maria