WHY INFORMATION ARCHITECTURE

I am a professional product designer and information architect often tasked with understanding a complex problem and making sense of it. Research and building a knowledge base using IA principles is a great way to have a deep understanding of a topic. Research can be the foundation for many purposes including improving findability, designing or improving an enterprise system, research for a UX project or as the foundation of an AI project.

Information Architecture (IA) creates the foundation for Artificial Intelligence (AI) to be effective through the quality, context and accessibility of data. As a former boss, Seth Earley, has explained in his book ‘The AI-Powered Enterprise’, ‘You can’t have AI without IA’. (Note: As an Amazon Associate I earn from qualifying purchases.)

RESEARCH AND IA FOUNDATION FOR KNOWLEDGE BASE

As mentioned, the research and IA foundation developed is critical for UX, product design and other disciplines but it can also simply create a Knowledge Base or a Personal Knowledge Base (PKM) for deep understanding of a topic which I will explore in this article.

METHOD FOR SURFACING AND STORING INSIGHTS

The following is one effective method of surfacing insights from a topic and storing those insights in a connected knowledge base.

I will be using the new AI search engine Perplexity to return detailed information from internet sources and Capacitiesto store and connect insights and for building a knowledge base foundation. Another way to extract insights is to use an AI Chat program that works with research PDF papers or books in PDF format but that will be another article.

PERPLEXITY SEARCH

If you haven’t yet tried Perplexity, be prepared for a new search experience that presents real answers along with sources.

“Perplexity merges search and AI capabilities, providing a more integrated experience. Unlike traditional Google searches, Perplexity uses advanced language models like GPT-4 to not just return search results, but to provide insights, analysis, and answers to queries. Perplexity cites the sources of the information it provides, allowing users to assess the reliability of the results.”

There’s a great in-depth article about Perplexity if you want to find out more. Also, as with any search engine, be careful about entering sensitive information in the prompt as there is not clarity about how they will store and use the information.

The standard Perplexity search presents excellent chat type search results that will guide you through asking appropriate ‘follow-up’ queries. The results also include organized images, videos and resources presented as linked citations. The Pro version presents smarter, more updated and detailed results at a cost of $20/mo. and includes access to their API. A limited number of Pro searches are offered each day so it is easy to try and see what it has to offer.

The Perplexity UI enables a copy of the search results by clicking an icon at the bottom of the results and also has a Share icon that can be used to copy a link to the query. These allow you to easily create a record of the results, within Capacities, which will be shown in the below example.

CAPACITIES IS A BETTER KNOWLEDGE SOLUTION

Capacities.io states that it is ‘A Studio for the Mind’ and from my experience it is that, and much more. The storage of insights is based on Capacities with the Believer or Pro level that enables AI through their AI Assistant, see Capacities pricing for more information. Capacities has a different take on the data structure using objects instead of folders. This structure creates a huge advantage in intuitively building a connected knowledge system.

Note: Links to Capacities.io are affiliate links that provide a commission to me if you purchase a subscription, there is no additional fee to you. As an incentive if you purchase a subscription through my affiliate link you will receive from Capacitiess the first month free and appreciation from me!

PERPLEXITY SEARCH — UNDERSTANDING BAMBOO FOR CONSTRUCTION

For the example of researching a topic, I have chosen ‘bamboo building and construction’ which is a complicated topic but good for demonstration. Bamboo is an amazing sustainable material whose time has not yet fully come in the architecture world but is starting to be given serious consideration.

The first thing I ask Perplexity is simply ‘Bamboo building”

The results from Perplexity display:

At the bottom of the results there are related questions that will add to the search results if clicked, or you can ask any follow-up question. This is a great way to refine and add more context to what you are really looking for, similar to Google’s experimental Search Generative Experience (SGE).

Perplexity related questions for follow-up searches

Perplexity also displays images and videos in the upper right corner keeping them separate from the results.

In addition, Perplexity lists sources with short descriptions when the ‘View More’ link is selected.

STORING SEARCH RESULTS IN CAPACITIES

To build a knowledge base around the topic of ‘Bamboo Construction’ we need to store the search results and associated content. Capacities enables us to store the results from Perplexity, build the necessary metadata and see a graph view showing the relationships to information.

CAPACITIES OBJECT STRUCTURE — SEARCH PERPLEXITY

In Capacities I have defined an object called ‘Search Perplexity’

Capacities object settings for ‘Search Perplexity’

Beyond the normal Title, Tags and Content I define:

Search Query (Text) — used to store the actual Perplexity search
Pro (Checkbox) — to indicate if it is a ‘Pro’ search
Summary_AI (Text with AI enabled) — used to build a summary that is used within the Card view. The instruction for AI is ‘in english write a two or three sentence summary for the content of this record’. This instruction works well for me but modify this as you see fit.
Tags_AI (Text with AI enabled) — used to suggest tags if they don’t seem obvious. The instruction for AI is ‘In English create a string of the top 5 keywords separated by commas that describe this record. Only write the keywords and commas, do not add anything extra.’ This is certainly optional but can be helpful for some records.
Results Link (Text) — for storing a link to re-execute the search.

Within the object settings for ‘Search Perplexity’ I also customize the card view by setting the ‘Customize card view’ as follows:

With this object structure in place in Capacities I will execute a search in Perplexity through the follow-up question of ‘what are the challenges of building with bamboo’. In the results, I copy them by clicking the icon at the end of the results:

The copied results from Perplexity can then be pasted into Capacities. After creating a new Capacities object record, select below the properties in the record placing the cursor in the ‘text’ area and typing ⌘ + v (on Mac).

I also save the link to the Perplexity search in case I want to re-run the search by first clicking the Share icon.

selecting the Share at bottom of results

Then select the copy icon in the upper right screen area which puts the Perplexity search link in the clipboard.

Paste this link into Capacities within the ‘Results Link’ property field

Capacities results link for Perplexity search

Make sure to click the ‘Fill’ button on the right side of the ‘Summary_AI’ property after the content is in place to have AI create a short summary.

When looking at an overview of the records, the optimized Card view presents a nice overview with the short title, the tags, a summary and the collection shown. The knowledge base you create will most likely have tens or hundreds of records which will offer the opportunity for filtered views or specified queries to be set up.

Capacities card-view showing Perplexity searches

Note that for the Title, I use the format of showing the detail first before the more generic terms. In the example ‘Challenges’ and ‘Regulations’ are used before ‘construction — bamboo’. This is important because in the Graph View the characters in the Title will be truncated so it is best to use the specific terms in the beginning of the Title.

Title created using specific terms first

After only six records recorded the Graph view begins to present a nice connected overview. (Remember to click on a tag that is common to all records, in the example it’s ‘Proj_bambooConstruction’ to enable the Graph view icon in the header).

initial graph view of bamboo construction

The process to build a knowledge base around a topic is not difficult. As more data is added the connections and insights become more apparent.

This example used the free version of Perplexity but the Pro Paid version takes search to the next level.

I hope this helps people understand one method to build the information architecture and eventually build a good knowledge base. Please share your comments and let me know if this is effective for your use case.

Again, Capacities is amazingly simple and yet powerful, see Capacities pricing for more information and to give it a try. Thank you for your support!

Note: Links to Capacities.io are affiliate links that provide a commission to me if you purchase a subscription, there is no additional fee to you. As an incentive if you purchase through my affiliate link you will receive the first month free from Capacities.

Knowledge Research using AI Search, IA and a Knowledge Base Program