Freshchat is a live chat software that lets businesses converse with customers. Agents typically deal with a high volume of conversations and struggle maintaining context.
The current inbox search allows users to search for conversations, and contacts. This project showcases the enhancement of the search experience in the agent portal.
DESIGN PROBLEM
The aim is to improve the current search capabilities by alleviating current user pain points and designing a more advanced search that improves the discoverability of conversations for agents
The primary persona is the agents- ensuring that they can access conversations easily. Freshchat is primarily used for usecase- support or sales, hence the personas would be customer support agents and sales agents
Agent Persona
In order to evaluate the existing system and its shortcomings in terms of usability, I evaluated the existing search experience against a set of heuristics.
Heuristic Evaluation
Research Methods Used : Competitor Analysis & User Inteviews
The first step of the process was to conduct research to identify user needs, behavior, pain points, validate assumptions and motivations.
You don’t always have to reinvent the wheel and hence, I conducted a competitive analysis to assess the currently available products that deliver a similar experience like Drift, Zendesk, Intercom etc. Through this analysis, I was able to understand the pros and cons of each and took these insights back to our drawing board to tackle this problem with a different approach.
One aspect of working with a mature product is the access to a pool of user feedback via tickets and forums. I collated this data to create a repository for all user feedback. I further categorised this
To understand what users typically search for and the context behind search. To validate the assumptions presented.
Contextual questions
1. How often do you decide what you are looking for and search? State one example scenario
2. When are you likely to search for something ? {Situation}
4. When do you search for contacts?
5. Do you search for keywords within the same conversation ?
Qualitative
Feedback
1. How efficient is the search currently ?
2. What are the positives and drawbacks of the current search mechanism ?
3. Would an advanced search be helpful ?
5. How often do you create custom views ? Can you give me a situation where this is used.
6. What have customers requested ?
7. Can you give me some instances when you’d need parameters to narrow/ specify your search ?
8. Is there a difference between how agents would use and the supervisor ?
Contexts & Usecases from interviews
Notes from user interviews
Empathy Map
Pain points
- Support agent that convert their conversations into support tickets, cannot find conversations easily - they search for conversations on their ticketing platform because the search on FC is poor.
- Duplicate conversations pop up whenever I search.
- Can't sort conversations based on date
- Doesn't highlight where the keyword has highlighted.
- Doesn't indicate if an exact match or close match has been found
- Can't filter conversation by label, date, contact
With ample data and information at hand, I was able to deconstruct the workflow into smaller components.
Phase 1
User Story : Entering key word
I made sure to incorporate plenty of cues to make sure the system is intuitive
Some key features that were included - a recently searched suggestion and an auto complete feature segmented by type
User Story : Displaying Results
Brought in information like total results, options to sort by relevancy and chronology and even added a filter option.
User Story : Contextual search
Introduced an option to search through conversations thereby letting users search for keywords within conversations as well
User Story : Filters
Based on the user interviews, the key ask was to include a filter section. Based on the research, I classified the filters into ones that will be frequently used and ones that aren't used too often.
Phase 2
Due to product roadmap reprioritisations and timeline constraints, search was redone with the intention of making it compatible with the Freshdesk Omnichannel.
Some of the constraints that we had to work with :
- Make a system that was scalable which would work with Freshdesk Omnichannel (the search should be able to search for tickets, contacts and solution articles in the future)
- Filters were not technically feasible and had to be moved to phase 2.
This project is currently in development phase and once it's done, the results will be updated on this page.
