Happy Friday!
What We Shipped This Week
Contradiction flagging got more thorough, especially on details that are easy to miss. This is the semantic layer you read about last week, now applied to those pesky stakeholders who use “batch” and “real-time” interchangeably.
Coming next week: Discovery Questions (beta)
I talk to proserv teams and SIs all day, every day. Almost universally, at some point in the process, there is a list of questions for the client to answer. Sometimes it’s called an input form, sometimes a scoping template, sometimes an operational readiness assessment - sometimes good old fashioned discovery questions. Sometimes the sales team owns the questions, sometimes the delivery team (often, they each have their own set of standard questions).
Today, these questions are static, manual, and inconsistently answered/inconsistently asked. Which is a shame, because they’re also super important; they’re the questions you NEED to know before you can scope well.
Static: by definition, it’s a list. It doesn’t help you dig into the nooks and crannies or surface areas of risk. It grows only when someone decides to change the template - usually, in response to something that was a total miss.
Manual: either the client is filling out a form by hand, or you’re asking them questions directly. In both situations, it takes manual, human time to go through the list and check things off.
Inconsistently answered: You give a client an intake spreadsheet and you get: skipped cells, extreme verbosity, one-word answers, cells highlighted for unclear reasons, sentences you have to read out loud slowly to understand, client-specific terminology that you have to ask them about later. You give a client an interview, and you’re good! Until you talk to the other stakeholders :)
Inconsistently asked: But it’s no fault of the question-asker! In a live discovery conversation, it’s hard to both engage in the conversation and keep track of your list - things get covered out of order, you skim through the list. You (the question-asker) are busy prying open interesting tidbits that they dropped that open up whole other areas of concern.
Discovery Questions fixes all of these things. It starts from your existing list of questions, automatically marks them as answered as project context comes into Glossa, surfaces the open ones so you never miss anything, and suggests how to go deeper. And, based on your project history, it flags not just how to go deeper, but where, so you’re always scoping better than the last time.
We're starting with presales. So if you're an SE with a list of questions you always ask, I want to talk to you - just reply! And for everyone else, hold onto your horses until next week, when we’ll be testing it out in beta.
Let’s hang out at Databricks Data & AI Summit next week
I'll be at the Data + AI Summit at Moscone (June 15–18). If you'll be there too, I'd love to grab coffee or say hi in person - reply and let's find a time.
Have a great weekend!
- Ali