The multiple stages of chatbot maturity

Having a Siri-like bot for your business will be less and less impressive as time goes on. Leading brands are now investing in bots to solve real customer issues rather than offering simple answers. Building a conversational bot is a challenge, but it is the right thing to do. We are on the verge of a new era in user experience, and you cannot afford to replicate the tedious IVR experience into your bot. You just can’t!

Bot objective and life cycle

You can deliver customer care in messaging without a bot. However, you cannot scale up without it. A straight-through bot may be good for the time being if all your customer needs is a simple commodity, like ordering a pizza delivery or booking a table at a restaurant. However, if you deal with customers who have real questions and issues, you need to go beyond the gimmick — you need to provide your customer with a hybrid bot experience (HBE).

Today, we are at the infancy stage of bots, and it’s OK to experiment with half-cooked UI during this stage. But what builders and brands may not fully realize is that messaging doesn’t behave like any other technology. It took 10 years for the web to reach 1 billion daily users; it took three years for messaging. Just a little while ago no one was talking about bots, but nowadays there is hardly any serious organization out there without a bot project.

The biggest challenge for enterprises is the speed of change. The guys from Facebook and such are moving fast — very fast! What took the web 20 years will happen with messaging within four. It means that the bot you envision for 2020 must be available by 2017. A year from now all of the current, infant, no ROI bots will disappear.

Once mature, proactive bots will generate value that is significantly greater than the value currently coming from passive websites. Call centers will fade away (well, almost — in some situations we will still want to get a human on the phone) and new AI Natives Brands will replace traditional, outdated web hippopotamuses. This is our projection for the bot life cycle:

• 2016 Infancy
• 2017 Youth
• 2020 Maturity

Bots currently have three significant barriers to break through to be able to solve customer issues:

1. Natural language understanding (NLU)
2. Integration with self-service APIs
3. Decisioning in a wide variety of situations

The AI challenge

Consumers expect the customer representative, whether bot or human, to have a solid understanding of the product discussed and the context involving the customer. Consumers do not want to guide the representative, either to say obvious things or to repeat their requests more than once. The context of a given incoming message (ICM) includes all of the ICMs/OGMs (outgoing messages) that happened before the current message, past interactions, stored customer data (from CRM, billing, etc.), customer history, and general public knowledge.

The complete AI challenge includes 9 levels of context:

Close context (current)

1. Last ICM (the current message)
2. Last OGM
3. All prior ICMs/OGMs within current interaction
4. Current customer profile
5. Current policy (current offering catalog, other policies)

Near context (history)

6. Past interactions with the customer
7. Historical customer data

Far context

8. Industry knowledge
9. World knowledge

Infancy

The infant bots of 2016 are easy to play with, but their business impact is very minimal. As with humans, in the critical infancy stage, the primary objective is to learn the language and gain skills. Infant humans cannot do much, but they are amazing learners. Unfortunately, as of 2016, most of the bots out there don’t carry the needed AI foundations, and they don’t learn anything. Brands will have to put in place the needed AI foundation for continuous learning; otherwise, infant bots will not grow up to become effective business machines.

Youth

If in 2016 an infant bot knows how to deliver simple A, B, or C scripts, the young, smart bots of 2017 will tell you: Thanks for reaching out, how can I help you?  They will understand your request in 95 percent of the cases and will execute the corresponding business flow effectively. A young smart bot doesn’t ask its consumers to provide information already provided during the conversation. It understands the near context, and it is capable of answering FAQs in each step of the conversation.

Maturity

If a young bot is expected to handle around 20 percent of the common customer issues, a mature bot is expected to know it to the level that cultivates valuable customer relations. The mature bot is expected to build loyal relationships with consumers and to allow the brand to harvest a yield beyond the here and now. A mature AI bot needs to demonstrate mature AI technology that blurs the demarcation between a bot and a human. Ideally, the consumer should not be able to tell if he or she is talking to a bot or a person.

Conclusion 

Your customers don’t care about your technological challenges, and you should and can provide an amazing bot experience from day one. If you provide time, love, and care from day one, your customers will use your bot as their primary channel to communicate with you.

If you use the correct, causal language and provide time, love, and care across the entire customer journey, you will have endless power in your hands. Messaging is not just another channel. Messaging carries the potential to transform your relationship with customers, especially digital-native customers.

After you provide outstanding service via messaging, something magical happens: Your customer considers you a friend. It happens in messaging faster than in any other form of communication. This channel gets very personal very quickly!

This article appeared originally at ServiceFriend.com