“But why would we need that?”
– said one innovation director of a large corporation me and my two colleagues visited at their offices. He said that after seeing our demo of knowledge management system that we’ve been building at Untrite in 2021. It was already utilising Natural Language Processing (type of AI that ChatGPT and other LLM systems are build with) – a revolutionary step allowing you to write a question in a natural ‘human’ language, while computer understood the context and the meaning behind it. You no longer needed to know the exact keywords of what you’re trying to find.
It wasn’t the first time.
And it was like experiencing “We’ll call you” moment (when they never call back), or reliving Groundhound Dog movie day over and over again and wondering what’s going on.
Not many saw Untrite’s value proposition, as if they are perfectly fine with SharePoint and knowledge management systems terrorising them with 000s of irrelevant results since 90’s.
After years of working around data problems with clients, we kept seeing the same pattern:
organisations with legacy systems and specialised knowledge suffer from data silos and fragmented information.
We felt that what we’re building is really needed, but we couldn’t articulate it in a simple enough, yet powerful way. We knew also from our own experience that in service fields in particular, people are often spending time searching for an answer instead of drawing from collective experience and wisdom. They keep reinventing the wheel. Or they don’t know how to even express what they’re looking for. We wanted to focus on helping businesses with critical products and services to serve their customers / citizens (as in case of police that we work with) – better and faster.
But nobody seemed to be seeing it as a ‘must have’ solution, so we kept thinking; do we speak alien language? Are we crazy? What do they know what we don’t? And why can’t they simply tell us so we build it for them? Or the world simply doesn’t care?
Turns out, we may have just been a bit too early.
Often you need to gain a critical mass, for the world to start noticing the need (or you need to have strong network that can land you a spotlight and a couple of first big clients, but that’s a topic for a separate post). Last year AI and the AI assistants around knowledge have finally got their spotlight. And I’m glad to see that the world (both the consumers and the investors themselves) start believing that this is the future; Perplexity landing series B (valued at $520m), Phind, Bard, ChatGPT, Komo, Bing and other LLM assistants seem to be the new favourite, indispensable tool.
However, many of these tools are designed to sound like they know the answer, while giving you lots of false positives and AI hallucinations (made up answers). They are disguised as an authoritative source, and you can’t be 100% sure which one of them is real. (like this lawyer who used ChatGPT in court and cited fake cases 🫣).
Who do you trust?
Verification and reputation are essential for building trust and credibility in human interactions. They help us reducing risks associated with transactions and decision-making processes. To put it simply – we people don’t like to think too much and we rely on others to help us make a decision. If someone tried something before us and it worked for him/her, we will take it.
Socially, reputation acts as a form of valuable capital, influencing us relationships and interactions.
That’s why Google PageRank was so revolutionary – because it allowed consumers to focus on sites which were ranking higher, as Google was doing the leg work for verifying which sources were more reputable and referenced more.
Similar mechanism contributed to Amazon’s success – namely, product reviews from verified buyers (at least until review farms were created and other ways sellers started using to hack the system).
For us at Untrite, providing verified and curated sources for our users is a big deal. I would even say, the biggest. When you have thousands of $$$ or worse – lives at stake, you can’t afford to provide hallucinated, unverified information.
Because internal company knowledge is by design self-contained and protected from the outside world like Google (that’s the secret sauce differentiating company X from company Y), all you should want to do is to take advantage of building on top of what’s already known and cross-checked. With help of technology, you should be able to easily find:
- What do I need to know to do my job better?
- Who can help me find a solution or show existing ones across my or other teams I don’t know of?
- How can I serve more clients in less time?
- How can I be up to speed with new product launches and improvements?
…and the list goes on.
We call it collective wisdom or collective intelligence. Dictionary definition says that collective Intelligence can be best understood as the improved capacity and knowledge that is created when people work together, often with the help of technology, to mobilise a wider range of information, ideas and insights.
In other words, if something has been solved and written somewhere by a reputable person, you should at least have easy access to it to help you do your work better.
What do to when the phone doesn’t ring.
But this is not the post about it. It’s about what you should do if you’re not gaining traction from the wider market.
It’s true, sometimes market signals are quite confusing, because even if your product does provide value but market is not ready for it, you won’t get product market fit. The best course of action is to adjust your messaging or find a niche in an industry that is more inclined to try new tools. Specialise in that niche and speak their language and users will tell you themselves what they “must have” to do their work, vs. what would be “nice to have” in the future. Find your ‘early adopters’ that will love your product and become your advocate. Word of mouth is your best ally.
That’s what worked for us, at least.
Ah, and as for those companies that initially rejected us and now are reaching out asking for help – we’ll gladly do so. But this time, we’ll quote them a 10x higher price 😉.
Pic. Joshua Tree, 2023, 40th surprise party for my friend Dave
