Ask HN: Solving workplace knowledge fragmentation with AI | Hacker News Reader

Ask HN: Solving workplace knowledge fragmentation with AI

2 pointsearcar3 comments

I'm researching the challenge of workplace knowledge fragmentation – the problem of important information being scattered across multiple platforms (Slack, email, wikis, GitHub, etc.), making it time-consuming and difficult to find what you need.

Initial survey results (n=30) suggest that:

- Most knowledge workers spend 30+ minutes daily searching for information

- Top pain points are scattered information, difficulty finding specifics, and dealing with outdated data

- There's high interest in AI-assisted information retrieval and synthesis

I'm exploring the concept of an AI assistant that could search across platforms, understand context, and provide relevant information quickly. However, I'm still in the problem-validation stage.

Questions for the HN community:

1. How significant is this problem in your work?

2. What solutions have you tried? What worked/didn't work?

3. What would be your concerns about an AI-powered solution for this?

4. Are there aspects of this problem I might be overlooking?

I'm also running a more detailed survey (link in comments). If you're interested in the problem space, I'd appreciate your insights.

Thanks in advance for any thoughts or critiques!

Here's the link to the survey: https://tally.so/r/wzqDq8

This sounds like https://www.glean.com/ ? I'm sure there is many other SaaS like that. It's just the one that came to mind first.

The problem you highlight is real in any large enough enterprise. But the issue is not just about where the info is:

- How do you know you have the latest view of a given topic, ie. the files you've ingested represent the "head" of your company's knowledge tree?

- How do you deal with that important piece of insights that never made it to a digital storage, because it's only in Joe's head and by the way he moved team so unless you know who Joe is in the first place, you don't know what you don't know?

- How do you deal with the fact that some desks really don't like sharing info with each other and a big part of a Project/Product Mgr's job will be to painfully wrestling info from other teams (very real in banking/finance)?

Workplace knowledge fragmentation is an interesting business problem, mind you. I'm just saying it's not just a software problem...

I'd really start with a market survey of the existing competitor landscape if I were you. And maybe niche-down to knowledge fragmentation for a specific role-type (QA or DevOps or Marketing, etc...). And pinging a few existing founders for insights. I don't know, just spitballing.

Thanks for the thoughtful reply!

You're right, there are other solutions out there. Glean's the pricey one aimed at big companies, but from what I've seen, the others are either still in development or not quite hitting the mark. My research shows this is a problem for smaller outfits too, and I think there's room for a solution that really works.

On your points:

1. Latest view on a topic: Yep, this is a big issue. I'm looking into ways to integrate metadata to help with this.

2. Knowledge in people's heads: Good point. Actually, I'm hoping my tool could encourage more knowledge sharing. The idea is that if it's not written down, the AI can't retrieve it, which might motivate people to document more. Still brainstorming on how to make this process smooth and not feel like extra work.

3. Information silos: Definitely a thing. My approach is to mirror existing access restrictions - the tool can only access what the user can. But I'm thinking about how it could still help improve cross-team communication within these limits.

Your idea about niching down is interesting. I'm casting a wide net right now, but focusing on specific roles could help tackle unique challenges more effectively.

I'm curious - what's your experience with knowledge management systems? What's worked well for you, and where do you see the biggest gaps?

Really appreciate your insights. This kind of feedback is gold at this stage.

Reply on news.ycombinator.com