AI Record Finder | MyHeritage
AI Record Finder is a revolutionary new way to explore historical records and build a rich narrative of your ancestors' lives through an interactive AI chat
ROLE
Project Manager |
Leading a team of 4
Design
UX research
DURATION
April-October 2023
INTRODUCTION
MyHeritage is a genealogy research platform that enables people to explore their family history through a collection of genealogical records from around the world.
​
The goal of this project is to provide an alternative to SuperSearch (MyHeritage search engine) that can coexist alongside it, offering users a different, more engaging, and user-friendly approach to discovering information about their ancestors using AI capabilities.
-
How might we leverage AI capabilities to create the ideal search experience on MyHeritage?
-
How might we provide users with an advanced way to consume information from various sources about their ancestors?
-
Who are the users who will benefit most from this experience?
AI Record Finder chatbot that assist users in locating their ancestors and unveiling the genuine stories from their family's past
Mobile version
Desktop version
RESEARCH
Interviews
Personas
Competitor Analysis
Thematic coding
User testing
User empathy map
NEED MAPING OF UX CHALLENGES
As a first step, we analyzed and mapped the challenges within the company's existing search engine of MyHeritage.
Among the primary issues discussed were:
-
requires manual editing of various fields.Using SuperSearch (MyHeritage search engine)
-
SuperSearch returns thousands of results for almost every query, making it challenging for users to find specific information. It is not always clear to users what additional information they can provide to reduce the results.
-
Individual information is dispersed among a variety of records. Therefore, users frequently find a single record containing partial information about the person they are searching for.
SuperSearch search form
INTERVIEWS
We conducted user interviews with a total of 10 participants of our target audience; including both genealogy experts, with years of experience in teaching genealogy research, and novice users who are simply curious about their family heritage.
We asked them about their past experiences with SuperSearch, about their strategies while conducting a genealogical research and about other tools they uses in order to discover pain points and opportunities. Following the interviews we conducted thematic coding to identify repeating patterns.
To present the results of the thematic coding, I will illustrate the insights through the lens of three different personas:
The focused investigator
Educated and curious, they dive into intellectually stimulating challenges, approaching tasks with thoroughness.
Jennifer Brown
-
Woman, 69 years old
-
Retired high school history teacher
-
Live in assisted living community in Orlando, Florida
-
Married with three children
-
Deeply connected to her children and grandchildren
-
Hobby: Genealogy, enjoys researching family history and uncovering ancestral stories.
“my Italian family tree is very hard to navigate, because everybody has the same names... I notice that on the Italian side of the family the woman retain their last name, which is not something that common and if you don't know the history of it - you don't even know that you need to look for it differently”
Pain points
-
Jennifer doesn’t sure the records she finds are relevant to her
-
Jennifer is eager to find more than solid facts about her family history
-
Her kids aren't into genealogy, and it frustrates her
Needs
-
An easy way to identify/validate the records she finds
-
A tool that will help Jennifer “connect the dots” and help with building the story of the family
-
A simple, fun and engaging way that will make her close family feel intrigued by getting to know more about their family
The carefree explorer
enthusiastic adventurers who prioritize creating memorable experiences, often juggle busy schedules and diverse interests.
Dan Bedford
-
Man, 47 years old
-
Data Scientist at a High-tech Corporation
-
New York City, New York, United States
-
Married with no children
-
Hobby: (New to) Genealogy research, previously surfing, exploring new cultures, and learning Mandarin.
-
​It is crucial for Dan to understand his family's past, as it gives him insight into his own identity and connects him to a broader historical narrative.
“If I am researching, you know, an ancestor from Ireland that ancestor from Germany, whatever. I don't always understand, if there's Somehow to tie in the piece of social history around it. Whether this is why they were, this is why, you know during this time period it was people commonly immigrated because of such and such...”
Pain points
-
Sometimes, Dan struggles to find the information he seeks until he delves into historical research, which helps him grasp the complete picture.
-
Dan relied on assumptions about his family story that were discovered as not accurate which caused the research to be delayed and misleading.
-
His research is saved on several platforms - can be confusing
Needs
-
A tool that integrates information about both the specific individuals in Dan's family and their cultural background.
-
A tool that prompts Dan with more specific questions to assess assumptions and refine the search by narrowing its focus.
-
Clear and intuitive way that allows to categorize and label information in a manner that makes sense to him.
The family archivist
Elderly, traditional, history enthusiasts who favor simplicity and are not tech-savvy.
Thomas White
-
Man, 75 years old
-
Retired Family Doctor
-
Texas, United States
-
Married for 56 years
-
Five children, 15 grandchildren, one great-grandchild
-
Values family bonds and quality time
-
Dedicated to preserving traditions and desire to learn from the past
“I think when I can add that person and be very confident that can add them to my family tree. And I know they're there, you know, I know they're mine. I know, I've got it. Correct. And just knowing that is a real sense of accomplishment and just brings me a lot of joy. When it comes to building that family tree”
Pain points
-
Thomas is finding it difficult to streamline the information and present it as a coherent narrative.
-
Thomas often feel lost when reaches a dead end and feels uncertain about what is the next steps to take.
-
He experiences uncertainty regarding the spelling of his ancestors' names, and when he discovers records that are unreadable to him, it further adds to his confusion.
Needs
-
A simplified approach that enables him to delve into his family heritage without becoming overwhelmed or confused by complex genealogical processes.
-
A tool that guides him in determining the most suitable next step based on the information he already has.
-
A tool that translates information and suggests culturally relevant name alternatives.
MAIN INSIGHTS
People aim to feel confident about the records they find while identifying the relevant records for their searches
People are motivated to discover the stories behind their family's historical decisions
People desire to involve their children and grandchildren in the family history research process
DESIGN PROCESS & ITERATIONS
×´You can't always get what you want,
But if you try sometimes, well, you might find you get what you need” Mick Jagger
While developing the feature, we encountered various challenges and explored multiple solutions. In this process, we made the decision to either discard certain approaches or opt for simplified versions, keeping practicality in mind for smoother integration during development.
The following are some solutions I've explored throughout the design process
STORY
One of the most strong themes from our research was that people desire for a complete story, supported by a thorough understanding of history and culture.
In an initial prototype, we introduced a 'story message' feature. After a user found the relevant person, they could access his/her full story in an engaging and up-to-date format. This interaction also provided users with the opportunity to share the story and involve more people.
The development of the 'story' required significant resources across multiple areas within the product. To prioritize our efforts and manage risks effectively, it was crucial to determine the necessity of this feature. Therefore, we decided to release a leaner version of the feature so we could study its demand and usage.
The alternative we chose
Instead of presenting a story, we chose to offer users a PDF divided into the following sections:
-
Personal details
-
Biography
-
Historical Context
-
Last Name Origins
This approach enabled us to achieve a shorter and more cost-effective development time, allowing us to gauge the demand for the story.
“FAN FACTS”
To help users identify and refine relevant results, we proposed including "fun facts" in the conversation. These facts would guide users toward more effective searches.
However, the idea was rejected due to its significant development demands and lack of clarity regarding its value.
INTERACTIVE TAGS
Research also shows that users may not always know what additional information they can provide to refine results. This occasionally leads to reaching a dead end, where users are unsure of the next steps, resulting in loss of motivation.
For this reason, we proposed the inclusion of tags in the chat, which could help facilitate conversation by displaying potential responses based on the chat's context. Additionally, we implemented the ability to reload other tags if the displayed ones didn't meet the users' needs.
Similarly, this idea was rejected due to its substantial development requirements.
”If at first you don't succeed, then dust yourself off and try again”
Aaliyah
One of the significant challenges in the project revolved around how to present records to users and assess their relevance. Throughout our journey, we experimented with various approaches, and I will now outline two of the strategies we explored.
PRESENTING RELEVANT HISTORICAL RECORDS
As part of this project, we had to figure out how to present the records to our users and know which ones are most relevant to them.
I will present 2 of the attempts we explored along the way.
SIDE PANEL
One of the design options we tested was to display the records in an interactive side panel that would change dynamically based on the conversation
We chose to drop this option because we were concerned lacking a singular focus point to focus on could cause confusion and disorientation. Additionally, adapting this design for mobile wouldn't have been ideal due to the limited screen size.
CHOSE MULTIPLE RECORDS
​It technically requires several records to produce a narrative about a person.
In response, we created a design allowing users to select multiple records from displayed choices. Next, click the 'done' button to get the narrative, or the 'continue chat' button if no relevant records were found.
As soon as this behavior was developed, we performed user testing to discover that the usability of the behavior was lacking. Below I will list the issues we encountered:
​
-
Copy problem - users did not understand the difference between 'done' and 'continue chat'
-
It wasn't clear to users that they had to scroll horizontally in order to see more records
-
There was a lack of understanding on what users had to do in order to progress
-
A lack of understanding about what the end result might be​
Our solution involved leveraging MyHeritage's knowledge of 'related records'. In this simplified approach, users only needed to choose a single record, click 'view,' and they were able to generate an 'AI Biography'. With this simplified approach, confusion was reduced, and users were better able to understand their ancestors narrative.
TAKEAWAYS
Although we set out to create a more engaging way for users to uncover their family history, the feature ultimately fell short of expectations. While users did receive historical records, these were often irrelevant or too fragmented to be meaningful. The algorithm prioritized output over accuracy, often returning poor matches in an endless loop. Additionally, because we chose to "start small," the experience lacked the storytelling and emotional connection users were hoping for — many expected rich narratives about their ancestors, instead of historical documents. This mismatch in expectations led to confusion and frustration.
In some cases, users chose this feature over the main search flow, which unintentionally decreased overall engagement. After three months, we decided to sunset the feature.
This experience taught me a lot about the gap between technical feasibility and user value, the risks of launching before a core user need is fully met, and the importance of aligning product goals with emotional outcomes — especially in sensitive domains like genealogy.