So in last week’s post, we talked briefly about the synthetic interview our client wanted us to do. As mentioned, we’re a bit hesitant about this idea, given the tendency of synthetic interviews to fall flat and not answer the user’s questions sufficiently. Even more worrying was that our client wanted us to use a prerecorded video, which wasn’t exactly the most well-filmed thing.
The existing video:
This week, our client dropped another bombshell on us: he wanted the synthetic interview to be able to handle open ended questions. This means the user could key in any question in the world, and our app will have to answer it as best as it can. For questions without an answer, the app was to notify our client so that they could insert a temporary slide with the relevant information, until they got around to filming it.
As you can imagine, this got us pretty worried, especially from a technological standpoint. One of our advisers, Scott, tried reassuring us, saying we just needed a weighted algorithm to power the search, and that it’s really simple to implement. All other technical faculty we talked to disagree on the difficulty level though.
One of the faculty we spoke to also brought up a very good point: many people, let alone senior citizens, often don’t know what they want to ask, so how can we expect them to know what to type into the search box? For example, if the user has no idea what a stroke is, how can they ask what their chances of getting a stroke after the surgery is?
After a lot of debate and a very helpful consultation session with Jesse, we decided to come up with another deliverable, one which would hopefully address our concerns: a hypertext FAQ. Our idea was to compile a list of FAQs based on the materials our client gave us, as well as the interviews we conduct. Next, we’ll match them up to relevant portions of the existing video, or note down if nothing fits. Finally, we’ll link each question to its video answer.
Of course, on top of that, we’ll still have the open ended search bar, but we believe that the hypertext FAQ will be a lot more effective and well received by the users.
Recent Comments