メインコンテンツに飛ぶ

We'd prefer it if you saw us at our best.

Pega.com is not optimized for Internet Explorer. For the optimal experience, please use:

Close Deprecation Notice

How email bots process thousands of incoming emails – fast

Matt Healy,
シェアする Share via X Share via LinkedIn Copying...
ログインしてブログを購読する

There is a surging need for fast responses to customer service inquiries

Customer service volumes are up 20% across the board since circumstances began rapidly changing as a result of the COVID-19 outbreak – and the increase exceeds 100% in some industries.

Customers looking to get in contact with a business to change, start, or stop services are seeing the same thing on websites - banners and notices warning of delayed response times. Depending on the means of communication, wait times span hours to weeks.

Organizations are working to succeed with three competing priorities in the face of the surge:

Supporting each customer with excellent service. Organizations need to provide easy, fast, and personalized service to each and every customer in need. With so much going on, customers should not have to wrestle with businesses through high-effort, slow, and disjointed service experiences.

Supporting their workforce through change. Like customers, employees have new circumstances to tackle. As workforces distribute, organizations need a strategy for digitizing work while maintaining visibility and structure.

Adapting with agility. Changes in economic, regulatory, and social environments will require changes in process, channels, and strategy. Organizations need to keep up by deploying changes across their business quickly.

Compounding the problem – an email tidal wave

Unsurprisingly, email service has among the longest response times of any channel. Keeping up with day-to-day email volume is difficult in the best of times. When a major crisis hits, the challenge gets even harder. With inboxes flooding, customers are waiting days to weeks for responses to urgent requests. That's because without the right strategy, email processing is manually-intensive, slow, and error-prone – bad characteristics for high-volume, SLA-driven, multi-faceted business processes. A good strategy will help mitigate the challenges inherent in email.

Emails are highly unstructured. Key information is scattered across email subjects and bodies. Agents looking to triage or process incoming emails spend valuable time parsing through blocks of text to understand the topic, sentiment, and details of the email.

Emails are attachment-heavy. Invoices, contracts, forms, receipts, pictures. Critical information is stored in attached, unstructured, hard-copy documents. Opening, reading, and extracting information from these is cumbersome and slow.

Email is siloed from business systems. Email is completely disconnected from business applications and back-end systems. Initiating or processing work from incoming emails requires manually copying data between systems.

The right email bot strategy and technology can help bail out flooded inboxes and plug the leak

Businesses need to enable smart, fast, repeatable processing of incoming emails to respond faster and eliminate time-consuming manual work. How can business leaders broach fixing their broken email processing process? This is where an email bot, along with the right strategy, are useful.

We recommend approaching email with a strategy and email bot technology that focuses on two things:

  • Understanding emails: Using AI to extract critical information from email and automatically triage work to the right person.

  • Actioning emails: Subsequently taking that understanding and using case management to drive end-to-end outcomes across business units and back-end systems.

With anything however, it's important not to just grab the first solution you see off the shelf, but to understand what characteristics will make these technologies and strategies successful in the long run.

Important considerations for understanding emails with AI

Let's take a look at three important considerations for both understanding and actioning emails that you'll want to take into account when evaluating email automation solutions.

  • Use AI: Seems like a no-brainer. Human conversation is unstructured, complex, and it changes over time. But AI typically has a high barrier to entry – leaving you to either hire a data science team or begin writing static if statements to route your emails. Processing emails based on hard-coded, deterministic rules and keywords is going to leave you with mistakes and huge technical debt as conversations change. You need a tool that is easy-to-configure and can train Natural Language Processing (NLP) to take raw human conversation and intelligently extract topic and sentiment – without relying on hundreds of custom if-then statements or a degree in data science.

  • Use AI built for business processes: Integrating with solution-agnostic, third-party AI and NLP tools will help extract intelligence from emails but will leave you building custom integrations to convert their responses into information that makes sense for businesses. You need a tool that goes beyond NLP to identify and capture key business entities – like account number, location, and date – and make those readily available for employees and case management tools.

  • Use AI that learns and improves: With experience, employees get faster and more accurate. Why should your AI be any different? Stagnant NLP models are like new cars – they deteriorate in value as soon as you drive them off the lot. You need AI with a machine learning feedback loop to increase accuracy over time and keep up with shifts in conversations and context.

How to action emails with case management

  • Start fast, improve iteratively: Rome wasn't built in a day; neither will your automations be. But that shouldn't stop you from getting value fast. Choose a solution that embraces agility and low-code development in the cloud so you can create initial rules for automation and then continue to modify and improve those rules over time.

  • Exceed expectations and optimize workloads: With workforces becoming distributed, it is more important than ever to supplement your teams with business process management tools to drive rapid outcomes and ensure your team focuses on the most important items first – SLAs, intelligent routing, reports, and urgency-based prioritization. But custom-building these tools will be costly and slow. You need a tool with BPM and case management ingrained in every application out-of-the-box to help you stand up workflow processes quickly.

  • Use intelligence to guide your agents: The goal, of course, is to fully automate the lifecycle of every email, end-to-end. This ensures that email is captured, tracked, and resolved. We know, however, that full automation is not going to be achievable on day one. You should still take advantage of the insights extracted from the NLP text analyses and use them to enable your agents to process emails faster. A solution that uses AI to guide agents to the most likely resolution, even for work that may not yet be automated, will help you respond in a quicker, more personalized way.

We expect customer service volume to remain high for the foreseeable future. Email bots help manage volume and support workflow.

By using intelligent email automation to bail out flooded inboxes you’ll be better prepared to keep up with the wave of customer service emails – now and tomorrow. Make sure you use a solution that leverages self-learning AI to help understand incoming email messages and then can action those emails with iterative, structured, and intelligent case management.

Learn more:

タグ

Industry: クロスインダストリー
トピック: AI・意思決定
製品エリア: Pega Workforce Intelligence
製品エリア: プラットフォーム
課題: エンタープライズモダナイゼーション
課題: オペレーショナルエクセレンス
課題: カスタマーサービス

著者について

As a Senior Product Marketing Manager for the Pega Platform™, Matt Healy helps the world’s biggest brands build, automate, and engage at scale with our best-in-class, unified, low-code platform.

シェアする Share via X Share via LinkedIn Copying...
シェアする Share via X Share via LinkedIn Copying...