Once I begin, I'm sure I won't know where to end.
Have you managed a survey project before, and do you know how to construct a valid survey? I'll assume the survey you've created is fairly good and measurable (and please, don't be offended...I don't know you or your work...most homegrown surveys simply do not make the grade).
You are right to be wary, but not necessarily for the reason you state. You are planning to impact over 700 people, and with all due respect, you better have a well-defined project plan in place.
Communication is absolutely vital. Who are your stakeholders, and do you have their support? Have you and your stakeholders sold this project throughout the organization? If not, start now. You need for managers, department heads, etc., to buy into the project. Then you feed them the information they need to inform their staffs and to garner their support. If your communications are spot-on, you don't have to worry about this turning into a complaints forum.
What are your milestones? How long do participants have to complete surveys? How will you do completion follow-ups? How many surveys need to be completed for you to be able to generate viable, valid data?
What kind of data do you want to see? Executives vs Managers? Managers vs Individual Contributors? Peers vs Peers? Male vs Female? How are you going to collect this kind of demographic data? Are you asking people to sign their names? If so, remember that if people believe their answers will be identifiable, they will not be as truthful in their responses. To look at norms, you need a minimum 25% response ratio of any reporting group. Of course your "norms" will not be able to be drawn against any industry norms, as you are doing homegrown.
Why is your survey internal only? If you want ideas on client service, should you gather feedback from your clients, too? (Not trying to be snide at all...there are very valid reasons for both including and excluding external people from the feedback process...just wanting to make sure you've thought about this).
Your survey structure: be sure to level the playing field by offering definitions of your response categories. Your definition of "good-better-best" is different than my definition. So define your scale, something like "Good = meets expectations, Better = exceeds expectations, Best = knocks it out of the park every time."
Make sure you don't have any compound questions. Those're questions that are ill-thought out. They contain two different ideas that make it impossible for the raters to answer honestly. For example: We deliver superior customer service and we build lasting relationships.
If your raters believe the first part (we deliver service), but not the second part (we build relationships)...how're they going to convey that?
How are you going to weight your responses? There are many ways to accomplish this, depending on what you are doing, what you are looking for. As an example, you may collect data on how your organization delivers service, markets service, and measures service. Which result is most important for your organization to act on? Or will Executive data be more meaningful to the organization than Individual Contributor data? Or will it be both action and data that is important?
What kinds of reporting will you be able to achieve once the data is collected? Do you have a tool to use? Are you going to dump it into Excel? And look for....what?
How're you going to report back your findings to the organization? Is your organization ready, willing, and able to move on the data once it is analyzed? This is absolutely critical, especially if you would like to continue surveying people to find out what they are thinking. If you collect information too often and do nothing with it, pretty soon no one will want to offer you information to collect.
I could probably go on for days. I just tried to hit some of the high notes you need to be addressing. I cannot urge you enough to create a project plan and follow it, and to take the time you need to learn about successful survey construction and deployment (if you don't have this in your background). If I could, I would encourage and urge you to scrap the process before you get too invested, and go out and hire the experts to conduct this for you. They have done extensive research to develop measurable, valid surveys. They know how to deploy to gain results. They know how to interpret the data. They have industry norms.
Good luck,
ksb | 
