State of Research Repositories, Knowledge Management & AI

Episode Timestamps

00:00 – Welcome and introduce Emily DiLeo

03:31 – What knowledge management really involves (adoption, governance, ethnography)

06:56 – Why dedicated KM roles are critical (not just side tasks for researchers)

09:20 – Key skills for leading KM: mindset, politics, sharing culture

11:44 – Culture, budget, and resistance: navigating pushback from teams

15:01 – Pre-implementation research: interviews, risk assessment, avoiding tool failures

18:13 – Library skills in UX: curation, description, metadata templates

25:58 – AI + KM: curated sources for bots, avoiding “garbage in, garbage out”

27:23 – Metadata playbook for UX teams (drawing from Library of Congress standards)

29:29 – Advice for researchers: start with problem spaces, learn from past repos

Main Insights From The Conversation

“I think some things I kind of noticed early on when I came into UX—an archivist was very common for me to be given like a collection of like hundred boxes of material from someone famous who passed and so it was my job to create this archival collection quickly. So going through lots of material, curating that material, deciding what was important, maybe what was less important and then writing, creating description for that material.”

For Emily DiLeo, PhD, MLIS, founder of The Current and former UX knowledge management lead at Thomson Reuters, SAP, Amazon, and Autodesk, this archival training wasn’t just academic background—it was the missing link in why research repositories fail.

After spending over a decade as a librarian and archivist at Yale University before pivoting to UX in 2021, Emily brings an ethnographer’s lens to a problem most teams think is about picking the right tool. She’s seen the pattern repeat across organizations: throw money at a repository platform, expect adoption, watch it fail, waste hundreds of thousands of dollars because the organization wasn’t ready.

In this conversation, Emily unpacks what knowledge management really involves beyond tool selection—adoption, governance, internal ethnography, political navigation, and the library skills of curation and description that UX teams desperately need but don’t know they’re missing. She explains why every failed repository is still valuable, how to do pre-implementation research that saves massive costs, what metadata templates drawn from Library of Congress standards look like for UX, and why AI makes good knowledge management more critical, not less.

Her core message? You’re building a library for your bot. If you feed AI uncurated, poorly described content, you’ll get “garbage in, garbage out” at scale. The groundwork—understanding culture and budget lines, co-creating with stakeholders, establishing metadata governance, identifying knowledge brokers, and maintaining ongoing stewardship—is what separates repositories that transform organizations from expensive shelfware.​

From Ethnomusicology to Archives to UX Knowledge Management

Emily introduces her unconventional path to UX knowledge management:

“My name is Emily DiLeo and I’ve been working on knowledge management initiatives in the UX space for about five years. I pivoted into UX in 2021 from working in an archive and I’ll talk more about that later, but I do have a background in library and information science. I have a master’s degree and I worked at the Yale University Library for 11 years”.​

The PhD in Ethnomusicology

“Before that, I did a PhD in ethnomusicology, which is I usually describe it as music and anthropology. But it also relies quite heavily on ethnography. So I do have a qualitative research background as well”.​

The Career Arc

“So I worked for a few companies—Thompson Reuters, SAP. I also worked for a bit at Amazon and most recently Autodesk. Again in knowledge management roles with UX teams”.

Founding The Current

“And last year I founded my consultancy called The Current because I found that even though the market for UX was very tenuous that these knowledge management needs were still there and I wanted to kind of make myself available for folks who could hire me to do like a short consultancy with their organization to help out”.

What Knowledge Management Really Involves

When asked what it takes to set up knowledge management in large organizations, Emily dispels the tool-first myth:

“That’s a great observation which is 100% true that I don’t think a lot of people—even executives and management—understand what is involved in knowledge management itself. And I mean maybe part of that has to do with knowledge management being traditionally kind of siloed in IT and not something that’s taken on formally outside of it”.​

Research Maturity Assessment

“So it’s just an interesting that I’ve found myself as part of a UX research team or research ops team but I’m doing knowledge management. And for me personally, and I think this goes in general, but especially because I have this research background, that means that I’m often in the position of assessing research maturity”.​

Defining Insights Before Storing Them

“Understanding how folks think of their definition of an insight, for example. Because if you’re going to make a database of insights, my first question is well, what differentiates one insight from another? So like how might we tag these different insights? So you automatically then start to have to do that work of defining insights”.​

Adoption Is Everything

“Again like what you were mentioning before, how difficult it can be to get folks to adopt different systems and tools. The same with knowledge management. So adoption is pretty huge area that I work on because any tool you use is only its value is equal to how much is used”.

“So I can talk a little bit more about what I do for adoption. That’s where my ethnographic background comes in handy. But also communication, right? when you roll out a new system. Also kind of actually internal research. I do a lot of researching the researchers and researching stakeholders to find out what it is that they want and expect.”

Governance Beyond Launch

“And yeah, so there is kind of a lot involved in knowledge management. And even after the tool is chosen and everything is kind of rolled out, it’s yes, keeping everything running, keeping it accurate, making sure that the governance is in place and it’s still upheld by the team. So it really is an ongoing responsibility that I think these days tends to fall in research operations teams and folks.”

Why Dedicated KM Roles Are Critical

Emily addresses a universal problem: knowledge management as a side task:

“Oh my gosh. Yeah. It’s so critical. I mean almost every team that I step into as either part of the team or helping out a team has someone who is doing that knowledge management work. And it’s almost always in addition to whatever else they’re doing. You know the research for example.”

When Deadlines Hit

“So when that person gets really busy or there’s a deadline then the knowledge management responsibilities kind of fall by the wayside. But I think once kind of problems start occurring, whether that’s not being able to find things or the stakeholders not being able to find things or not understanding what’s happening and there’s a communication issue, it kind of always comes back to knowledge management. So it really is pretty critical”.​

Advocating for the Work

“And like a lot of my work is actually helping folks who are championing knowledge management at their organizations know how to advocate for this work and explain to their managers and executives why this is so crucial”.​

The Siloed Research Problem

Emily shares a common enterprise scenario:

“Just as an example at one you know enterprise organization and this is super common for organizations that have been around for a while that you know every team that worked on a product was very siloed in their research.”

The Executive Vision vs. Reality

“So what happened then is you know executives said you know I want we want the customers to have a very continuous experience of our product line. Well, we can’t have that if researchers aren’t sharing with one another or they’re not able to you know access another team’s research or see what’s going on with another team”.​

Different Systems Per Team

“And they can be very siloed like every team is using a different system for knowledge management in which case like that that has to be like a pretty big sea change to make something like that happen. And it sounds very simple that oh we just want a common experience across our tools across our products but it’s actually really difficult to carry carry that out”.​

Key Skills for Leading Knowledge Management

When asked about essential characteristics for KM leadership, Emily outlines the mindset:

“Yeah, that that’s a really good question. I think yes of course I mean every team not just UX research teams but like customer experience teams, sales teams like everybody does have to have an eye on knowledge management right like this is a issue for for everybody like anybody who works in a team anybody who has to share knowledge has to have these certain skills and kind of understandings of how knowledge management works and what’s important.”

The Knowledge Management Mindset

“So when I teach a course part of the course is understanding like what is the knowledge management mindset. So for example you know if you’re going to create a shared file system then you all have to align on how you’re going to organize that file system”.​

The Common Failure Pattern

“But usually what happens is one person goes and organizes the system and expects everybody to fall into their organizational way of thinking about things and that that just doesn’t work right.”

Political Issues Around Sharing

“But then of course that touches on political issues about sharing. Like some folks don’t want to share knowledge. they don’t want to share you know that’s job security or that’s you know their hard-earned learnings that they maybe don’t want to share for political reasons and you know I I have to sometimes you know address those situations and understand what’s going on in an organization frankly or because if it’s a big issue then then that’s going to be very difficult for me to work in that in that environment”.​

Culture, Budget, and Navigating Resistance

Emily explains how culture impacts implementation:

“So you know by culture I think what I mean is like what is valued right like what kind of behaviors are valued in that particular organization like what are people rewarded for doing or not doing and that yeah that’s pretty huge in terms of implementation too”.

Budget Lines and Skin in the Game

“It’s like also budget lines like I have to understand you know where is the money coming from you know who I mean it sounds but it’s very important to understand like who has kind of skin in the game here and who doesn’t like because that’s going to make or break an implementation or funding for a tool right if if departments need to share funding for a tool or if they all have to contribute that can get pretty tricky.”

One Team’s Problem vs. Another’s Non-Problem

“You know for the implementation and anything to do with the implementation will also affect adoption right like who maybe one team thinks this is a problem and another team does not and so my job is to understand why like what why is that so I really you know maybe here I can talk about the kind of research that I do when I come in”.​

Pre-Implementation Research: Saving Hundreds of Thousands

Emily describes her most valuable intervention:

“I mean, I can say that I think it saves organizations a ton of time and money because most are just like, ‘Okay, we’ll throw money at this tool. We’ll purchase this tool and just go’ and then it fails and they’ve just wasted hundreds of thousands of dollars on a tool when their organization wasn’t ready for it”.

The Research Process

“So the kind of work I do as a consultant is to come into the organization and interview the stakeholders, interview the researchers, who’s using the research, what and whatnot. And then you know I provide like a real I take a deep dive into those and like what are the themes that are coming out? Are folks ready? Is the organization ready? Is funding ready?”

Risk Management Assessment

“And then kind of you know do some kind of risk management kind of work and and just say okay here’s the bottom line you know okay maybe you all are really ready for this change but these certain things have to be in place so the tool will fail unless you have these things also in place and I think that’s that’s really valuable because I’m it it just happens so frequently I just see it so much even with small teams as well”.​

Navigating Pushback: UX Research Within UX Teams

When asked about resistance from other teams during implementation, Emily reframes the challenge:

“Yeah. I mean I think sometimes there’s work that I think sometimes UX researchers are so passionate about the research they do and the value of the research that we also have to pay attention to okay I’m going to give you this like it has to just be more of an exchange right like you must use this system to get access to what I’m doing no we’re actually going to co-create a system together.”

Applying UX Principles Internally

“It’s kind of like UX research within UX team, right? Like who are your stakeholders? How do they work? How would they like to access your insights? Like what’s the easiest way for them to do this? Like what’s the least friction? And a lot of UX teams don’t do this, follow this line of inquiry. That’s mostly because they just don’t have time or maybe they don’t have the bandwidth to do this”.​

Observing Informal Knowledge Systems

“But you know another thing I do is notice and observe the kind of like human connections that happen when there isn’t a system that works very well. You know we have humans have a lot of ingenuity right they they come up with these ways of communicating like oh yes I always talk to this person when I want to find something out about research they know a lot right so that person has a lot of organizational knowledge.”

Knowledge Brokers

“And so they set up these kind of knowledge informal knowledge systems some folks even become what I call knowledge brokers where they consider part of their position is communicating research to stakeholders and that that an act in itself is very valuable to them”.​

Co-Creation Is Key

“So I think again it’s kind of like finding out who who are the players and what do they do and what do they value, right? Straight up ethnography like how do they think they think about what they’re doing and how can we make a system that benefits both the UX researcher and the stakeholder.”

Cross-Functional Collaboration: UX + Marketing

Emily highlights promising integrations:

“And maybe the last thing I’ll mention is that I have seen collaborations like between UX and marketing teams which I think is great, right? Yeah, of course. Let’s put all the research let’s put both of those research assets into one tool. I mean that’s kind of ideal, right? So that there’s less places for stakeholders to check when they’re looking when they’re looking for research”.​

Library Skills in UX: Curation and Description

When asked how library background helps with knowledge management, Emily makes the connection explicit:

“So I think some things I kind of noticed early on in when I came into UX and I just to be specific I I was an archivist rather than a librarian which you know it it’s it’s the same degree just the things that archivists do is a little different.”

The Archivist’s Job

“So an archivist it was very common for me to be given like a collection of like hundred boxes of material from someone famous who passed and so it was my job to create this archival collection quickly. So going through lots of material, curating that material, deciding what was important, maybe what was less important and then writing, creating description for that material. What is this like who is it for? Why is it important?”

The UX Parallel

“So when I came to UX and just starting to work with research repositories I was like ah yes this is exactly this is exactly the kind of work well that researchers do every day they just don’t know it right is curation right deciding you know why a study is important like what what are our best studies and and why is that and kind of reflecting on that and thinking about that also description.”

Archival Description Standards

“So There’s like an entire handbook that’s dedicated to archival description. It’s all kind of rules and of ways of describing things like as objectively as possible. And so like that plays into so many aspects of work in general, not just UX, but like when we describe you know a file in our SharePoint, when we you know use descriptive metadata on our reports, which I can talk about later, which I think is really important right now with AI”.

Single Source of Truth

“And creating a central point of access which is called the what is it the shared single source of truth. This the kind of same idea as in library science is this central point of access where you can go and you know that’s the place where you can find everything. So all those concepts I realized were translatable.”

AI + Knowledge Management: Curated Sources Matter

Emily addresses the AI question directly:

“I mean this has been kind of a theme of of my work for the last year it’s been okay this company that that I’m working with wants to set up you know they want to use AI to solve all their knowledge management problems. And my my answer back is always well yes AI can help. But you there’s a lot of groundwork that we have to put in place before AI can really do what it does well.”

Garbage In, Garbage Out

“So you know when you create this kind of library for your bot to draw upon, make it a curated area. I mean, don’t don’t give it garbage, right? Because that’s just garbage in, garbage out”.

Magical Thinking vs. Reality

“Because I think I think there’s a lot of magical thinking you know with AI and a lot of kind of aspirational ideas about how you know it can it can completely change the landscape and yes it has the potential to completely change that landscape but unless there’s like you still need that groundwork underneath right.”

Alignment on Sources

“Not just in how you set up the AI but like what what you’re giving it and what and are you all aligned on what it is that you’re going to give it? Right? So, it’s like you can’t get around that work of that that you want to do with knowledge management”.​

Metadata Playbook: Drawing from Library of Congress

Emily introduces a key artifact she’s developing:

“Sure. So I am creating this kind of playbook. I’m calling it the like KM plus AI for UX playbook for teams who are have been you know directed to start using AI or they want to try to use it. And I’ll include that the metadata template in there”.​

Not Secret, But Standardized

“It’s not it’s not anything secret. It’s I’m drawing upon library cataloging. Again, archival processing and archival cataloging where like, oh yes, we should definitely use these models because they’ve were been developed by like the Library of Congress and around for a long time”.​

Team Alignment on Metadata

“But again, it’s also the metadata is something that has to be aligned on as well, right? like everyone has to decide okay do we do we need the title of this project in our metadata? Not really. I mean because the title may not mean much in in that in the company. The date might mean is probably means a lot but like what kind of date format do you want? Do you want to say what quarter it was in? Things like that.”

Basic Things Create Big Problems

“So even basic things like that I think sometimes get missed which creates problems later”.

Advice for UX Researchers Setting Up Repositories

Emily offers encouragement and practical guidance:

“I do. And oh my gosh, people have done such amazing work in setting up research repositories. And even if they haven’t been used and even if they fell flat or no one adopted them, I mean, please don’t despair. They are going to come in super handy in the next iteration of whatever that repository is going to be.”

The Value of Failed Attempts

“Not to mention that you’ve learned a lot about what the assets that you do have and that knowledge in itself is very valuable”.​

Start with Problem Spaces

“I would say you know to set up a repository or repository 2.0 right now is to again think about your problem spaces. What do you think is the most urgent problem about what you have? You can think about it in terms of kind of your internal team. Is it an internal knowledge management issue? Is it a, you know, communicating insights out to stakeholders issue?”.​

The First Company Story

“The first company I ever worked at in 2021, before I even really knew much about what was happening in UX, I recommended a tool that wasn’t a UX tool at all for their repository because I said I I don’t think the problem is that you are not, you know, organized enough. you’re expecting stakeholders to come to you and they’re not going to”.​

Key Takeaways: Repositories Are About People, Not Tools

Emily’s journey from ethnomusicology PhD to Yale archivist to UX knowledge management leader reveals crucial lessons:

1. Knowledge management is never just about tools

It involves research maturity assessment, insight definition, adoption work, communication, internal ethnography, governance, and ongoing stewardship.

2. Dedicated KM roles are critical, not side tasks

When researchers do KM “in addition to” their main work, it falls away during deadlines—exactly when it’s needed most.​

3. Pre-implementation research saves massive costs

Most organizations throw money at tools before understanding readiness, culture, budget dynamics, or who has skin in the game—wasting hundreds of thousands when repositories fail.

4. Political issues around sharing are real

Some people hoard knowledge for job security; others won’t share hard-earned learnings; ignoring these dynamics dooms implementations.​

5. Co-create systems with stakeholders

Apply UX principles internally: understand how stakeholders work, how they prefer accessing insights, what creates least friction—don’t impose researcher-centric organization.​

6. Observe informal knowledge systems

Humans create ingenious workarounds when formal systems fail; “knowledge brokers” emerge who communicate research to stakeholders as part of their identity.​

7. Curation is deciding what’s important

Like archivists curating hundred-box collections, researchers must reflect on which studies matter most and why—this work happens whether you acknowledge it or not.

8. Description follows archival standards

Entire handbooks exist for describing materials objectively; UX teams need consistent metadata drawn from Library of Congress and archival cataloging, not ad-hoc tagging.

9. Failed repositories are still valuable assets

They surface what research exists, where silos live, and what organizations are actually ready for—knowledge that informs better second attempts.​

10. AI makes good KM more critical, not less

You’re building a library for your bot; uncurated, poorly described content fed to AI produces “garbage in, garbage out” at scale.

11. Magical thinking about AI misses groundwork

Yes, AI has potential to transform landscapes, but you can’t skip alignment on sources, metadata governance, and curation just because an LLM can summarize.​

12. Cross-functional collaboration reduces search burden

UX + marketing research in one tool creates fewer places for stakeholders to look; integration beats proliferation.​

Final Thoughts: The Library You’re Building for Your Bot

Emily’s thesis is deceptively simple: knowledge management in UX fails because teams think it’s a tool problem when it’s actually a people problem.

After 11 years at Yale Library processing collections, conducting ethnographic research, and developing archival description standards, Emily recognized the work UX researchers already do without naming it—curating studies, deciding importance, describing context. The difference is professional archivists have centuries of developed practice, standardized metadata, and governance structures that UX teams reinvent poorly each time.

Her pre-implementation ethnography saves organizations from the most expensive mistake: buying repositories before understanding culture, budget dynamics, political resistance to sharing, stakeholder workflows, or whether teams even have aligned definitions of “insight”. The pattern repeats: executives mandate tools, researchers comply, no one adopts, hundreds of thousands vanish, and six months later someone new tries again.

But Emily reframes failed repositories as valuable. They surface assets, expose silos, reveal maturity gaps, and teach what “ready” actually means. Repository 2.0 should start not with vendor selection but problem space definition: Is this an internal KM issue or insights communication challenge? What do informal knowledge brokers already do that formal systems could support?

The human ingenuity she observes—researchers who become go-to sources, informal sharing channels, knowledge brokering as identity—shows systems emerging whether leadership acknowledges them or not. Good KM design co-creates with these realities instead of imposing researcher-centric organization and expecting adoption.

Her library background surfaces constantly. Curation isn’t nice-to-have—it’s deciding which hundred boxes matter from a thousand donated. Description isn’t optional—it’s the archival handbook of objective, consistent metadata that makes future discovery possible. Single source of truth isn’t buzzword—it’s central point of access, a library science fundamental.​

And now, with AI, these disciplines become urgent. Emily’s repeated warning: you’re building a library for your bot. If you curate nothing, describe inconsistently, govern haphazardly, and feed LLMs whatever sits in folders, you’ll get synthetic confidence about garbage. The magical thinking that AI solves KM problems without groundwork leads to automated hallucination at enterprise scale.

Her metadata playbook drawing from Library of Congress standards offers teams what they’re missing—tested, standardized ways to describe research that align teams and make content AI-ready. Not title fields that mean nothing, but date formats stakeholders actually use. Not tags invented per project, but controlled vocabularies that enable comparison across years.

The political dimension matters too. When knowledge hoarding protects job security, when sharing feels like giving away hard-won advantage, no tool overcomes that culture. Emily’s work mapping who has skin in the game, why one team sees problems another ignores, what behaviors get rewarded—this ethnography surfaces implementation-breaking dynamics before they waste budgets.

Her advice to researchers tasked with “repository 2.0” after a failed first attempt: celebrate what you learned. You now know your assets, your silos, your readiness gaps. Start with urgent problem spaces. Decide if you’re solving internal organization or external communication. Don’t expect stakeholders to come to you—understand their workflows and meet them where they are.​

The UX + marketing integration she advocates shows another path: fewer systems, broader research types, less stakeholder burden. When customer experience, sales conversations, and UX studies all surface in one place, the repository becomes hub, not silo.​

Most powerfully, Emily positions KM as ongoing responsibility, not launch event. Governance doesn’t maintain itself. Accuracy doesn’t preserve itself. Adoption doesn’t sustain itself. Teams need someone whose job—not whose “also responsibility”—includes curating, describing, governing, and evolving knowledge systems as organizations change.​

Her consultancy, The Current, exists because even in tough UX markets, KM needs persist. Organizations still generate research. Insights still get lost. Stakeholders still can’t find what matters. And now, as everyone rushes to plug AI into repositories, the consequences of poor groundwork scale from frustrating to dangerous.

Emily’s archival training taught her that description standards, curation practice, and governance structures aren’t bureaucracy—they’re how knowledge survives beyond individuals who created it. UX teams doing this work without training, without time, without dedicated roles, without aligned metadata, without understanding their own politics—they’re not failing because they’re bad at KM. They’re doing an entire professional discipline as a side task.​

The repositories that work, Emily suggests, are the ones that acknowledge KM is ethnography, political navigation, co-creation, ongoing stewardship, and yes, library science applied to research. The ones that fail are the ones that think it’s about picking between Dovetail and Aurelius.​

And in the age of AI, that distinction determines whether your LLM amplifies organizational intelligence or industrializes organizational confusion at scale.

Thank you for reading!

If Emily’s insights on why repositories fail, the value of archival thinking, pre-implementation ethnography, and building curated libraries for AI resonated with you, share this article with UX teams, research ops professionals, and anyone tasked with “fixing our repository problem.”

Have questions about metadata playbooks, knowledge broker dynamics, or preparing repositories for AI? Connect with us at hi@uxarmy.com

Special thanks to Emily DiLeo for sharing two decades of expertise from ethnomusicology research, Yale Library archives, and UX knowledge management roles at Thomson Reuters, SAP, Amazon, and Autodesk. Learn more about her consultancy at The Current.

And to all of you, thank you for being part of the User Insights community.

⚡ This podcast is brought to you by UXArmy, an all-in-one UX research tool.