question

I think there are two things here. The first is that Pandora is not a third-party skill, it's a direct integration with Alexa, so they get all sorts of special UI features that normal skills won't (at least not in the near term, I'd guess). The second, though, is that I think you can probably get most of what you want with what IS available. See this doc here: https://developer.amazon.com/public/solutions/alexa/alexa-skills-kit/docs/linking-an-alexa-user-with-a-user-in-your-system

As Galactoise has said, there is a major division in features. There are those provided by Amazon (Pandora, shopping list, and various home automation integration) and those provided by 3rd parties (skill tab, done with the ASK). There is not a level playing field between them. Built in features work intrinsically. They are addressed as part of the Echo's base functionality. ("Alexa, add underwear to my shopping list") 3rd party skills have to be invoked via deliberate (and often awkward) syntax. ("Alexa, open Knock Knock", or "Alexa, ask Tweet Poll about Hillary Clinton") I'm pretty sure some of this is deliberate. Amazon don't want to be held liable for misuse of features that are not theirs. (E.g. the flagrant abuse potential of Ask My Buddy.) Drawing a clear distinction between the two will look better in court. Then, having 3rd party skills less streamlined, also means that they don't compete with anything Amazon chooses to introduce as an intrinsic feature. It's a *lot* more natural for a user to say "Alexa, tell me a knock knock joke" then "Alexa, open knock knock". My skill can't compete with theirs (other than in quality). If you truly want to go the Enterprise route, and you think you have enough of a market, I think you should be talking directly with Amazon. If you think you can move 100-1000 units, they would probably be willing to make firmware changes, put your branding on it, and put it out there as our own.

@jjaquinta -- about this: "If you truly want to go the Enterprise route, and you think you have enough of a market, I think you should be talking directly with Amazon." That would be a dream come true, but I suspect that we need to prove ourselves first. We also run into this (the need to prove ourselves) when we talk with venture capital firms. We've been asked to come back for a follow-up conversation in January. Many of them want to see 2 things: 1.) will the Amazon Echo be popular this Christmas? 2.) can our app generate real excitement from executives inside large enterprises? If we can go back in January and answer "yes" to these 2 questions, then I think we can manage to get some funding, and then that will help us continue with our experiment to see if the Echo can be made to answer some urgent needs inside the Enterprise. And then if we can do all that, I hope to talk to Amazon. Certainly, we will apply to their $100 million fund for the Echo.

@Galactoise -- thank you again. The documentation that you link to looks like exactly what I was looking for.

> will the Amazon Echo be popular this Christmas? Was the Amazon Echo popular on Black Friday? Who can tell? Amazon haven't released any numbers on Echo sales. On my little metrics chart, many of our skills look to be doing great. But without knowing the size of the market, there is no way to tell in absolute terms. How do you plan on sourcing numbers to impress venture capitalists after Christmas?

@jjaquinta -- "How do you plan on sourcing numbers to impress venture capitalists after Christmas?" We don't have a plan. We will improvise something. If Amazon releases numbers then we can use those. (Perhaps if the numbers are good, then Amazon will want to boast?) If Amazon does not release numbers, then we will proceed entirely on the basis of how many executives buy an Amazon Echo, or seem excited by our demos. In the end, that is the part that matters most. All executives struggle with the fact that their companies have vast amounts of information trapped in databases. How to get the data is a constant source of conversation. If a new interface appeared that made it easy, then executives would certainly be interested. I don't think anyone can dispute that. The only question then is whether the Amazon Echo offers that interface. There are 2 problems: the lack of support for secure WPA2 Enterprise wifi, and the frequency with which Alexa misunderstands the company name. That second problem we might be able to address ourselves, but the first issue entirely depends on Amazon to fix.

Exciting executives is about proving business value. Everyone wants to "unlock the information trapped in their databases". My company is heavily invested in delivering on big data. But a new interface is not going to solve the problem of synthesizing business value from vast reams of raw data. I hope you have a solution to that, and are just using voice for your presentation layer. But, even if that is the case, I think you have a lot more problems than the two you listed. When our Tweet Poll application launched, it used IBM Insights for Twitter to extract some quite detailed information about up to the hour trends based on Twitter sentiment. You could compile and present the information in quite detailed ways. For a nation obsessed with an election that was still a year away, it was a great tool. Only no one used it. You could not build a voice interface that allowed you to say "Alexa, ask Tweet Poll to tell me about positive tweets about Republicans ranked by state". You had to break it down into constituent parts. "Alexa, open Tweet Poll... Positive... Republican... By State." Very powerful. But too much for people to grasp. Right now it's being migrated to a much simpler, spoonfed interface. It's annoying for me, because you can get the very detailed and specific breakdowns that the other format allowed. But there is no point in a powerful tool that no one uses. I think with Alexa in its current form, this is a problem that will play out each time it is used to try to wrangle large data sets or complicated search and query type things. [Here's a demo of the old interface: https://www.youtube.com/watch?v=Gd5HKlbSpo4 You can use the skill itself to see the transitional interface.]

As to the question of whether the Echo did well this holiday season: http://www.theverge.com/2015/12/1/9826168/amazon-echo-fire-black-friday-sales

Yeah, that was the "report" I was indirectly referencing. No numbers. Nothing to impress anyone who isn't already an enthusiast. My Echo Programming book was #1 in its category for six weeks after launch. Am I going to put that on my resume? Damn right. Am I going to get a book deal with a publisher without giving actual numbers? Not a chance.

@jjaquinta -- I think Tweet Poll is awesome, in terms of showing what an open ended session allows via Alexa, however, it is mostly giving people information they already know. Anyone who watches the news for 5 minutes has a reasonably good idea of how well Clinton is doing, or how well Donald Trump is doing. A Vice President of Sales would engage in a dialogue with Salesforce exactly because the information does not exist anywhere else. And Alexa does offer a great interface for doing that kind of open ended exploration, as you have demonstrated.