Because tagging breaks down in familiar ways, Frame offers a way to create a new source of truth with your existing data. Your team tags customer conversations in order to learn and communicate . Maybe they’re calling out challenges and victories for a weekly review, categorizing feature requests to share with product teams, or flagging competitor mentions for your sales and business teams. In each case, tags exist because modern CX teams do more than just help customers succeed: they help their company succeed by understanding the customer. Unfortunately, tagging also tends to turn into kind of a mess. Let’s say you could freeze time. You meet with sales, product, compliance, and every other team about insights they need from customers. You review past discussions and create the perfect, manageable tagging scheme. How long would it stay perfect?
In the real world, things change. New customer groups arrive, bringing new concerns and new ways of talking about them. Conversations spike around releases or other events, taxing your ability to tag as you go. Product areas and business priorities shift and merge. And as they do, your perfect ontology gets… stretched. Tags get duplicated, with slightly different names. New tags get applied by some people and not others, so that every analysis requires asking yourself “were we tagging those like that back then?” Both (a) tagging conversations and (b) using the tags to do research become a chore. The good news is that you’re not alone. For months we’ve been asking customer-facing teams about their biggest tagging challenges. We used their answers to put together a bestiary of troubled tagging schemes, including a few below. See if any sound familiar!
The Void
You stared into the challenge of setting up a useful schema, and it stared back into you. Now tagging conversations is perpetually on the backburner, because the challenge to implement has only grown with your customer base and inter-team needs. You cross your fingers that any high priority request about your conversations comes with the time and the research budget to do a lot of reading!
The Mosh Pit
Wouldn’t it be nice if everyone got to organize tags for themselves? Many have tried, and this is what happens: a wild free for all. Alice tags password reset question “access” while Bob uses “login.” Janet marks when someone asks about “pricing,” which sort of overlaps with Evan’s tags about “upgrades.” Nobody even remembers what “Project Cucumber” was, but it has its own tag. Sometimes your tags are useful, but they can never be trusted, so important research requires — you guessed it — manual effort.
The Junk Drawer
Years ago, when you had a dozen customers and two people doing part-time support, Alice handcrafted an ornate tag schema. Now you have hundreds (or thousands) of customers, a team of reps, and no one can find Alice. Your team uses the filing system she built, but mostly shoves everything into a drawer called “Other”.
The Archaeological Dig
Your tagging system has been created — and re-created — one project at a time. Remember when you were first building a knowledge base, and your team came up with 20 potential sections before settling on six? Those are in there. And that one bot that attached an incomprehensible version tag to every conversation for a week before you shut it off? Recorded forever. Someday you’ll tell the story of the company with these tags, but for now every time the auto-complete opens is a game of “remember when?”
We can do better.
These situations don’t happen by accident, or because people don’t care. They happen because tagging is hard . It’s hard to do accurately in the moment, it’s hard to do as your company scales and needs change, and it’s hard to do when your help desk system treats tags as just one more tool for surviving and then archiving a conversation. At Frame, we think that customer conversations are an asset that can help the whole company make better decisions — if they’re easy to sift through and report on. So we’re on a mission to make tagging easy.
Creating an adjusting your ontology should be easy. So Frame’s “Elastic Tags” let you view, rename, split, and merge tags you already use — keeping your history, but creating a view that serves the way you think about customers today.
Reporting to stakeholders should be easy. So Frame lets you group tags and generate reports on trends and correlations most relevant to each team that depends on you.
Applying a new tag to past data should be easy. So Frame gives you a powerful, cross-channel search system with rapid bulk-editing, so you can scan through and label conversations in a fraction of the time.
Tagging new conversations consistently and accurately should be easy. So Frame trains a neural model for any tag in seconds — allowing it to surface suggestions of where you may have missed the tag in past data. When you’re satisfied with the model, you can even turn on auto-tagging as a supplement to or replacement for your manual process.
Reporting on tags should be easy. So we built clear, configurable visualizations that respond to your searches and tags, letting you support your customer analysis with data. And if the view you want isn’t here, every search and visualization is exportable to your favorite spreadsheet.
You can get started right now.
If you’re looking for a better way to really understand the content of your conversations, we can help. Reach out and let us know what questions you’d like to answer about your customer conversations, or just get yourself started at frame.ai. Tag preview in Frame with a neural model enabled.
