“The ROI case for Re:infer was very easy for us to make. Just from the initial analysis, the platform was able to show us a detailed overview of where our highest-volume requests were coming from. We were able to identify very quickly where we could get the most bang for our buck!”
<small>Service Enablement Manager<small>
The challenge
Our client is one of the leading providers of payroll, HR and benefits outsourcing services in the US. The company pays 1-in-12 of private sector employees in the country. To properly serve its 700,000 customers, the company relies on a large service workforce of around 5,000 payroll and HR specialists.
However, the scale of this undertaking created significant operational and communications challenges. Clients predominantly used email as the primary engagement channel for requests and queries. This meant that, in a single year, the company’s service team had to process more than 40 million emails.
A lack of visibility into these communications, coupled with inconsistent manual processes, led to subpar customer and employee experiences. The volume of emails coming through to the service team was so high that the company often struggled to maintain satisfactory response times. This became a compounding issue that degraded service quality and strained client relationships that weren’t receiving the support they needed. Employee turnover and customer churn were running high.
In a single year, the company’s service team had to process over 40 million emails.
The solution
The client had previously invested in robotic process automation (RPA) tools, but were unsatisfied with the results. Due to the unstructured nature of email and communications, the client’s RPA tools failed to identify issues and opportunities in its service function. The client didn’t know where to start and could not see how they would be able to automate any of their client responses.
That’s why the payroll provider partnered with Re:infer. After a rapid deployment, the team was able to use Re:infer Communications Mining to quickly understand and analyse its email data. As part of phase 1, the client has used the Re:infer platform to turn over 1 million production emails into structured, actionable data in near real time.
Within days Re:infer provided detailed analytics and dashboards of the trends it identified. Using the latest advances in AI and NLP, Re:infer was able to identify and understand the drivers of inbound demand and categorise the types of emails coming in.
The result
Through Re:infer, the client has been empowered to identify negative sentiment in its service function and uncover high-volume opportunities for automation.
It discovered that the client service team was receiving a high-volume of Yammer emails that were only meant for internal communication. This allowed the client to intervene and quickly deflect these emails from agent workflows - no RPA required.
Re:infer also identified high volumes of ‘thank you’ emails from clients. While appreciated, these emails remained a distraction for agents who found themselves wasting time to process and close out the connected requests before moving on. By automating the closing of thank you emails, the client was able to reduce its total email volume by 9% and save its agents precious time.
Most importantly, Re:infer has given the client a new capability to automate and improve customer escalation. Re:infer rapidly identifies negative sentiment and emotion in inbound emails, enabling the client to immediately auto-route this demand to its specialist customer retention team.
In only 3 months, the client was able to keep 202 customers from churning, saving the company $420,000 in annual revenue as a result of this improved escalation system.
Key Results
- 202 controllable loss opportunities identified - $420k in annual revenue saved in just 3 months
- 9% of total email volume automated from end to end
- Yammer emails removed from Upstream workflow tool
“Re:infer is super easy to use. I’m not the most tech-savvy person, but I was easily able to train a model by myself and quickly get it into production. Re:infer trains based on the content and meaning of messages rather than their words. This makes it very simple to train the platform to recognise certain themes and topics without having to go through long training sessions.”
<small>Service Enablement Manager<small>
