6 Technologies for Personalized Messaging at Scale
In an era when 97% of business calls go unanswered, and nine out of 10 consumers prefer instant messaging over talking on the phone, enterprises are under immense pressure to change the way they communicate with consumers. To meet these growing demands, 80% of 800 decision makers from well-known brands intend to deploy chatbots by 2020, according to a 2016 survey released by Oracle.
However, even with the introduction of chatbots, companies still struggle to deliver that "personal touch" at scale. Automation has historically come at the cost of personalization. To deliver hyper-personalized messaging experiences, businesses need to seamlessly integrate six fundamental technologies.
- Channels -- SMS, Facebook Messenger, WhatsApp, LiveChat, Alexa, Slack... There are so many communication channels to choose from, each with slightly different user behaviors and capabilities. No matter what channels customers want to use, businesses must be able to build a solution once and have it operate consistently across all channels. Simply stated: Build once and seamlessly deploy to any channel.
- Automation -- Messaging solutions must operate at scale, which inevitably means at least a basic level of automation. Some kind of artificial intelligence engine will be required for natural language processing (NLP). Amazon Lex? Google DialogFlow? Microsoft Luis? A custom bot solution? There are so many options. Basic, pre-programmed keyword automation is one thing, but what about getting your AI engine to locate data from your CRM and then act based on your customer's response? What about the training data needed to seed the AI engine? How do you improve automation automatically?
- Structured Data Exchange -- Depending on the AI/NLP deployed, you should be able to solve basic intent discovery and automate responses to common requests for static information such as "How late are you open?" Beyond that, bots struggle when they have to collect or exchange structured data, quickly leading to customer frustration. Take a simple example:
Bot: What's your first name?
Bot: What's your last name?
Customer: oops Phil
Bot: Hello, Philll oops Phil, how can I help?
AI/NLP isn't the right technology or paradigm for exchanging structured data like forms, checklists, wizards, or rich media content. But customer experiences will inevitably need this capability. You need to provide a seamless transition between unstructured conversation into bi-directional structured data exchange. Ideally, the structured data can be validated and then sent back to the bot to further the messaging flow and drive results.
Certain channels allow a richer media experience than others. Facebook Messenger is one, RCS messaging promises another, and Apple is touting its Business Chat. Microsoft's Adaptive Cards allow businesses to present data in a variety of visual formats, and Oracle licenses software to support bidirectional structured data with its bot platform.
However you choose to tackle this problem, your "instant" apps will ultimately need to work on 99%+ of mobile devices, ensure security of data, be deployable to any channel, handle bi-directional data flow, and be engineered for automated build and delivery. They need to present native-like mobile-optimized experiences, without download or install. They're purpose-built mini apps that can be inserted into the conversational flow for a real-time, dynamic exchange of information with a specific customer at a specific moment in time.
- Integrations -- Messaging solutions need to integrate seamlessly with backend data sources, CRMs, and APIs. Well-formed RESTful APIs are a rarity in the business world, and companies often underestimate the challenge of pulling together all the information needed from their siloed systems.
If a customer asks "How much to renew?" a bot may need to hit your CRM to find the customer's current plan and then another database to find the latest pricing for that plan. Without tight data integrations, messaging solutions are largely relegated to toy-like use cases for which they merely surface generic information, unable to act on it.
- Agent Tools -- While automations will be first responders in many messaging experiences, inevitably you'll need humans to take over the conversation -- for VIP customers, transcript verification, training, or agent escalation. You'll want the entire bot-based transcript available on demand for agent review. A robust solution will also surface data collected during structured data exchange and actions taken as a result.
- Analytics -- Finally, you'll need to track and measure how your messaging experience is performing. What queries are we successfully handling? How many agent handoffs are required? What's driving higher conversions? How are our CSAT scores trending?
Bringing It All Together
Businesses building for the future will need to stitch together these six key technologies into one seamless, on-demand experience. Successful companies will deliver hyper-personalized, robust messaging-based experiences to their customers at scale. And your customers will love you for it.