Sunday, August 31, 2008

Lessons learned: Educate the management, pt. 2

In an earlier post, I mentioned several things that business managers need to know before you dive in and develop a speech IVR for them.
  • Strengths and weaknesses of speech rec technology. Pity the poor VUI designer who starts a project whose stakeholders have been educated about speech only by slick sales presentations and glossy brochures. I've been there - it's no fun. Part of the challenge of design is to work within the constraints of current recognition technology, like the ability to discriminate between letters in the "e" group and between the letters s and f.
  • Skills required to implement a good speech IVR. Speech IVRs aren't just another IT shop application. The learning curve for speech rec development and testing is very steep. The learning curve for effective dialog design is very steep. Companies that hand complex speech projects to an internal group need to appreciate that and build in a lot of time for training and development of people before the first speech project even begins.
  • Persona isn't just about the system's voice. This one is a little controversial, but I can say from experience that managers tend to spend too much time selecting the perfect "voice" for the IVR in the belief that this "branding" exercise makes or breaks customer satisfaction with the system. The sound of the voice - the way the prompts are delivered, in particular - is important, but the design and proper functioning of the application is equally important.

If a management group "doesn't have time" to participate in a level setting meeting I'd recommend pushing back on that very hard. Spending a little time before a project kickoff talking to management about speech projects in general can save a lot of thrashing later on.

Monday, August 25, 2008

How you know your "Second Life" is out of control

When your "Second Life" virtual girl friend ties up your dog and tries to kidnap you with a taser. Big shout out here to homegirl Kimberly Jernigan of Durham, NC, who drove to Delaware to attempt the kidnapping.

Whew. I guess we can expect to read a lot of stories about the evil influence of social networking sites for a while.

Wednesday, August 20, 2008

IBM speech recognition

IBM has been working on speech recognition for a long time. Still, they've never gained traction with their speech engine in a market dominated by Nuance and its predecessors. This recent article in BusinessWeek about their appearance at SpeechTek 2008 reminded me that they're still there, looking for a market for what may be some pretty good technology.

It's kind of a mystery to me why IBM hasn't made much of an impact on the speech industry. IBM isn't the first company you think of for call center applications, but the engine is available on Avaya's platform. So I don't get it. Any ideas, let me know.

Sunday, August 17, 2008

Flying through Denver

I'm flying in and out of Denver this week because I have business in Boulder. The Democratic National Convention starts on August 25, so I'll catch some of the travellers to the convention. Maybe I'll try to score a convention button or a silly hat from someone.

When I get back I start my 2nd year in the part time MBA program at NC State. So I'm going to be busy, and that translates into shorter blog posts, at least this month.

Monday, August 11, 2008

Remarks on Ahead of the Curve

I'm reading a book that is so good I couldn't wait to finish before blogging it. Ahead of the Curve is by Philip Delves Broughton, an English journalist who quit his job to enroll and pursue an MBA at Harvard Business School. HBS is, by reputation, one of the top MBA programs in the country. Delves Broughton was alternately impressed and appalled by the things he saw there, and tells all in a very funny, ironic way.

When he's writing about the content of the coursework the book sometimes reads like his transcribed class notes. It's when he talks about the attitudes of the students and faculty that make the book worth the read. The present and future Masters of the Universe don't always look very attractive in the author's narrative.

As I read I can't help but think about Philip Zimbardo's classic mock prison experiment at Stanford, in which students were assigned the roles of prisoners and guards in a "prison" set up in the basement of a Stanford classroom building. The experiment demonstrated that participants' behavior was heavily influenced by their environment and preconceptions about how guards and prisoners behave in prison. In Delves Broughton's account, HBR students quickly adapt to the priviledges and perks of being part of the "HBR brand," and display a sense of entitlement.

Excellent read, especially if you're in or considering an MBA program of your own.

Tuesday, August 5, 2008

Persona and in-car navigation devices

This writer for the Washington Post has a problem with the personae of in-car navigation devices, and with speech recognition systems in general. The article a little over the top in its opinions because the writer is trying to be humorous, but it fairly points out that the speech output of devices can become annoying very quickly if not done properly.

Note that the problem as identified isn't necessarily with the sound of the voice, it's the lack of context machines have when giving directions, and the fact that the problem reveals itself over time, as in a long car trip. These things are often not tested well, or tested at all. The article also identifies lack of "naturalness" as a problem with speech systems, but trying to achieve an undefined kind of "naturalness" causes problems as well.

Saturday, August 2, 2008

"Experiential" courses for execs

This BusinessWeek article made me laugh. Companies are sending their execs on junkets to exotic places for leadership training. It's a stroke of genius by some innovative travel agents who have repackaged their expensive tours as "executive development." The travel business in the US is hurting due to the falling economy and rising fuel costs. Big companies have deep pockets and spare no expense in developing their top level managers even if, as the article points out, these "courses" have "dubious educational value." The travel agents are really onto something.

I've decided to jump on this executive development bandwagon and put together my own "experiential learning" course for executives. The difference between my own course and these others is that I think mine would deliver real value. Here's the outline of my course.

  • The exec gets a assignment to write a research paper on a subject he/she is expected to know something about. The first draft submitted is returned with numerous requests for clarifications and a terse note in email from the leader of the exercise who is role playing as the exec's supervisor, "You can do better than this," with no actionable suggestions for improvement.
  • The exec presents his report in a meeting with other "execs," really actors role playing various roles in an organization.
  • The meeting attendees check email and voice mail during the presentation, have side bar conversations with each other on topics unrelated to the meeting, and act generally as if they'd rather be somewhere else. Two attendees at opposite ends of the table text message each other, make eye contact, and giggle after each exchange.
  • The questions during Q&A are unrelated to the content of the presentation. The presenter is frequently interrupted when trying to respond. At the end of the meeting the presenter is thanked for his time and ignored as he shuts down his laptop and gathers his things.
  • The exec receives email from his "supervisor" stating that someone in the audience had some questions about the presentation and that he, the supervisor, answered the questions as best he could. The supervisor doesn't remember what the questions were.

At the end of this grueling exercise the exec learns that this is a not uncommon experience for analysts when presenting to middle managers in many large organizations, perhaps even in his or her own. If a company is trying encourage innovative solutions from everyone in the organization, you can see that this sort of misbehavior doesn't contribute to that goal.

So, if anyone is looking for some experiential sensitivity training for their execs or other managers, drop me a line. I'll do the first class for free. It would be fun.