Sherpa.RunningLate 2

Here’s yet another personal assistant app in a space that is becoming fast crowded — and where others have died before.

Empowered by the idea that mobile will make everything different, there are some notables: Google Now, Tempo, Grokr, and of course Siri. Add one more to the mix: Sherpa, an iOS app that launch today in closed beta. It also announced $1.1 million in funding from Andreessen Horowitz, Google Ventures, InterWest Partners, and angels.

It’s a tough market to crack, especially when users have become evangelical with their bots of choice. Google Now has a rabid following and Tempo has had success early on, despite a hiccup with the onboarding process, where some new users reported waiting for hours while the service integrated with users’ calendars.

On a general level, Sherpa (not to be confused with this personal assistant of the same name) does many of the things the other personal assistant services do: gets traffic information, gives you flight information, and gives weather updates for where you’re going. For example, if you’re in San Francisco driving to a meeting in Palo Alto and traffic is heavy, Sherpa will proactively push an alert to you asking if you want to send an email to your colleagues telling them you’ll be late. It also tells you if it will start raining in Palo Alto in the next 15 minutes.

But as these bots become more common and the wow factor wears off, nuance and usability will be the key to getting traction. After all, others have promised such wow features before– included the ill-fated Palm Pre. The big differentiator, founder Bill Ferrell says, is basing the service completely around location, instead of something else, like the calendar. It’s an odd pitch, considering all of the other bots, including calendar-centric Tempo, to some degree give cues based on location.

But this is where it gets interesting: What he means is that he eventually wants to turn Sherpa into a platform for developers to build location-based personal services. Ferrell says that people like to compare the service to the other apps that are in the marketplace because of similar consumer capabilities. “But we’re excited about powering location functionalities for other developers,” he says.

For example, if you are driving home, Sherpa can ask you if you want to turn the heat on in your house or unlock the door when you’re a few minutes away. There are, of course, companies already working on those individual tasks, like Nest and Lockitron, respectively. And this functionality has worked in cars for years. But Sherpa hopes to be the platform that brings all of these disparate kinds of services together.

Ferrell also gave the example of getting your morning coffee. If Sherpa knows you go to the same café everyday – based on geo-data it’s already gathered from you – the app can ask you if you’d like the coffee waiting for you when you get there. This may be a little harder to pull off when thinking about scaling to ubiquity, because then you have to partner with individual merchants.

Ferrell’s idea reminds me a bit of IFTTT – a fellow Andreessen Horowitz portfolio company. IFTTT (short for “If This Then That”) allows users to create little digital “rules” for various situations. For example, one rule might be, “If it’s going to rain, send me a text message.” But while that platform empowers users, Sherpa wants to hand the reins to developers.

It’s an interesting long-term vision, given all the hype over connected devices. If executed properly, a good platform could be a main artery for developers trying to inject the Web into every part – and every thing – involved in people’s everyday schedules. The real test will be if Sherpa’s infrastructure for developers will be strong and inviting enough to attract them. Sherpa will also need to convince people to stick by it as it works toward this long-term vision. Because, as absurd as it sounds for something still so new, Sherpa is already late to the personal assistant party. And it will need to keep users not only content, but wowed, in the meantime.