(This is one in a series of 10 articles extracted from the publication Canadian Diversity: Technology in the Settlement Sector (2023). I'll be posting each article as a separate post here on my site.)
Aimee Holmes is Director, Data Governance and Application Services at ACCES Employment. She is passionate about empoweringteams to use virtual tools in service delivery and enhancing how information is gathered, shared, and made available to thosewho need it. Aimee leads the Online Services department at ACCES, a powerhouse team of professionals who lead an integratingan AI-based virtual assistant into service provision, train staff to use over a dozen platforms in programming, develop e-learningmodules, and strategically evolve a Salesforce-based CRM to optimize it for the organization. Aimee has an MA in Global Studies.
In 2020, ACCES Employment launched a virtual assistant powered by Artificial Intelligence (AI) to enhance its online gateway to information and services. In the first two years, VERA (Virtual Employment and Resource Attendant) responded to over 130,000 inquiries about ACCES’ events, programs and other frequently asked questions. VERA offers a viable point of entry to ACCES’ services for job seekers and employers. This article offers insights on the initial investment and ongoing resources needed to build and operate VERA, the unique role that it plays in ACCES’ service delivery journey, and the benefits of using this technology.
In the fall of 2020, ACCES Employment (ACCES) launched a virtual assistant powered by Artificial Intelligence (AI). VERA (Virtual Employment and Resource Attendant), responds to inquiries about our programs and services, helps users self-register for upcoming events, and offers resources that are
relevant to users’ queries. VERA is a chatbot interface on our website and operates behind the scenes within our Facebook Messenger app. She runs on IBM Watson Assistant and Discovery technology.
Between 2016 and 2019, ACCES received two generous grants from Accenture. The first grant gave us the ability to transform our digital infrastructure and offer more online resources for jobseekers. The second grant allowed us to build VERA to enhance our online gateway to service by offering website visitors an interactive experience. Although we started working on VERA prior to the COVID-19 pandemic, the timing for introducing such a tool was ideal as people became more accepting, if not expectant, of being able to access information and programming online.
To guide the design of VERA, ACCES staff provided valuable input regarding the tasks that a virtual assistant might perform. They identified routine tasks at the initial point of service, such as providing an overview of a specific program, conducting a basic eligibility assessment, registering clients for events, and responding to questions about what we offer and our locations.
In the first 2 years, VERA responded to over 130,000 inquiries, which we calculated to offset over 8,500 hours of staff time that would have been spent attending to these requests.
Ultimately, VERA’s success will be reflected in our ability to convert online interactions into client registrations for service. We made two key decisions in pursuit of this goal. First, we structured the conversation to lead users to complete a contact form, so a team member can follow up with them. Second, we selected technology that integrates with our Salesforce-based CRM. The information VERA collects on the contact forms and event registrations is passed directly into our primary stakeholder database. Thus, staff are able to pursue client and employer leads from VERA, alongside the leads that come in through traditional channels. This also gives us the ability to track impact.
Approximately a quarter of all users who filled out a contact form on VERA, went on to become a client. This is a very strong conversion rate when compared to our other light-touch initial interventions. During this timeframe, 12% of all new clients first interacted with our organization through VERA.
During conversations with the virtual assistant, users can ask questions in their own words. VERA uses Natural Language Processing (NLP) to interpret these queries to find the most appropriate response from our library of more than 150 pre-composed answers. Each month, we review a sample of conversations to confirm the accuracy of VERA’s responses and remediate where needed. This also provides insights on trends in user queries and whether we need to add new responses. For example, in 2022, we added a response about how ACCES supports refugees and those who have come to Canada under Emergency Travel Measures. Through Service Delivery Improvements (SDI) funding from Immigration, Refugees and Citizenship Canada, we collaborated with CCS (Catholic Crosscultural Services) and OCASI (Ontario Council of Agencies Serving Immigrants) to add 30 new responses to common questions newcomers have on settlement topics such as short-term housing, and how to obtain a health card or driver’s license. These responses include links to authoritative sources for further information and support.
Another goal we are pursuing with VERA through the SDI funding is to facilitate inter-agency referrals. We will gather clients’ settlement-related needs during the initial conversation with VERA so that our staff will be aware of clients’ needs and can make outbound referrals sooner. Staff at other organizations can use VERA on behalf of their clients to screen them for our services and will be able to make referrals through VERA too.
ACCES needs dedicated resources to maintain and enhance VERA. We have an internal team that ensures VERA is performing well and aligns with our current programming. The team also works with external vendors through enhancement projects and leads the ongoing communication with, and training of, staff to ensure they can capitalize on the leads that come in through this channel. IBM charges a subscription fee to use this technology.
For us, the benefits of this technology and the ongoing investment is clear. VERA plays an important role as a source of reliable information, a viable point of entry to ACCES’ services, and a support to staff by handling routine tasks.
Related from Aimee:
Human-centric AI - some Immigrant and Refugee-serving sector promising practices
I recently shared episode of my Technology in Human Services podcast where I speak with Meenakshi (Meena) Das about not just focusing on AI, but Human-centric AI. It generated some great comments from a couple of people working on AI projects in their organizations, building on and adding to Meenakshi’s insights, which I wanted to share here.
AI, Data and Improving Service Delivery (Metropolis conference 2021 recording)
Published on: April 13, 2021
This session focused on how cutting-edge technology has the potential to support and enhance current models of service delivery through automation and the application of solid data analysis. Presenters shared current and practical applications of AI and other technology innovations in the sector.