We all have junk drawers that are messy and full of clutter. They contain things we don’t know what to do with, but just can’t let go. Within Factor, we asked around and junk drawers included old magazines, outdated devices, broken pens, books we promised to read, receipts, and even a to-do list from 1987. As uncomfortable as it is to find a to-do list from 1987, there is also gold to be discovered. Like a gift certificate to a favorite pizza place or gourmet cupcake stand.
Large organizations have junk drawers too.
Unlike our personal junk drawers, enterprise organizations are less confined to one easy-to-find drawer. Enterprise drawers often contain the bad and/or misplaced data that has accumulated or aggregated, through intent or accident. In the case of your enterprise, the junk drawers contain every conceivable information delivery channel to your constituents (i.e. customers, employees, investors, and more). The data seems too valuable to delete, but nobody can figure out what to do with it or get any value out of it. So into the junk drawer(s) it goes.
It might make sense to ask – don’t we have technology that can do this now? Don’t we have AI, ML, NLP or GPT or some other acronym that can help us magically clean out the junk and extract the valuable stuff?
Unfortunately no, because AI and even much-celebrated large language models (LLM’s) like ChatGPT are reliant on the quality of the information they are fed. If we feed AI a diet of junk, it will return a result made of junk. To be super blunt: Garbage in; Garbage out. The simple reality is that AI cannot automagically transform your junk into gold. It needs help. Using finely tuned humanity-based scientific processes, we sort through the junk to find the gold and then maximize its value. This is Information Architecture.
Information Architecture is the art and science of creating meaningful structure(s) for data. It includes metadata, or taxonomies, or even (jargon alert) ontologies or knowledge graphs. It hasn’t always been recognized, or well-understood, but it’s so critical that…
“There’s no AI without IA.”
This phrase was first popularized by the Watson team at IBM. They’ve been working on AI probably longer than anyone, so they should know. Even sophisticated LLM’s can’t generate meaning out of randomness. They perform much better when they have structured, intentional, inputs. Less junk mines more gold.
Most, if not all, large organizations have this challenge. We’ve seen this pattern over and over. Pardon our French, but over ten years at Factor, we have come to call this the “Oh sh*t, we’re screwed!” pattern. There may be more technical terms for this, but this sums it up perfectly. It usually goes like this:
- Generate lots of data.
- Figure out what to do with it later.
- Get new software to save you (it won’t).
- Hire a data scientist to save you (they won’t).
- Oh sh*t, we’re screwed!
This approach leads to the following outcomes:
- Lots of data; little insight
- Missing data
- Low quality (untrustworthy) data
- Data needs lots of wrangling and clean-up
- Data is difficult to access
- See #6 above.
At Factor, we propose a different approach. It requires more work up front, but that work will save you from heartache and legitimate conflict over time. It goes something like this:
- Evaluate business strategy and goals
- Align (internal and external) user goals
- Identify the data that supports those goals
- Explore technical capabilities and limitations
- Create models for your data that is usable by both humans and systems
- Develop a sustainable governance and maintenance plan for your data.
This is how you move toward a meaningful return on investment in the foundation. We have guided organizations through this process and with the new tools we have available, we’re just beginning to scratch the surface of what’s possible. Tools such as ChatGPT can’t do this work for us, but they can help us speed things up along the way. However; it’s important to not to forget that this relies on human interaction. It’s built upon intensive research and deep engagement via interviews, surveys, and discovery. Through this process we create authentic connections from individual experiences in order to achieve collective success.
If you think about your junk drawer–this deeper discovery element makes perfect sense. A random stranger (human, robot, or AI) can’t know the difference between garbage and gold. Upon further review, that 1987 to-do list might include something worth saving–maybe it has “buy engagement ring” listed on it. Only the most informed would know.
So, where do we start? First of all, we’d like to hear about your messes, your junk drawers, your clutter. We’d like to help you get rid of the useless decades-old to-do lists, make peace with the unpaid bills, find the gold, and put it to work.