Consistency and Accuracy in the Wisdom of the Crowd Era

By Monica Kohn   |   May 16, 2017
Consistency and Accuracy in the Wisdom of the Crowd Era
Our democracy is based on the wisdom of the crowd and always seems to us as (relatively) the best system of government. In recent years, systems, start-ups and innovative ideas such as Waze, eToro and many more are based on the wisdom of the crowd.

It seems very common to us that decisions and answers will be taken by the crowd.

Is it such a good idea?

Do you want your doctor to analyze your situation and give you a diagnosis based on the wisdom of the crowd?

Do you want your car to navigate based on crowd wisdom?

Do you want that the answers to your customers in your site, in your chat, or in the call center will be based on crowd knowledge?

The answer to these questions – is NO.

In all these cases we want accurate, consistent and professional answers to our problems.

You owe your customer the most accurate answer to your question, and consistent answers in all channels.

From other hand we still want to enjoy the fruits of the wisdom of the crowd? How can we do this?

The answer for this, we want to gain from the wisdom of the crowd but not only.

You need to have the ability to build your answers based on your existing knowledge that resides in a proper knowledge base, up-to-date and accurate with the ability to strengthen it by the crowd.

Here are some examples of on how we do it:

In Lighthouse we use Natural Language Processing (NLP) and Semantic Search techniques as a baseline for our search results and Bot interaction answers.

We improve it by leveraging several ecosystems, all of them use human know-how.

First of all, feedback. Feedback can be implicit and explicit.

Implicit feedback can be – pressing on first result 2 seconds after querying a specific question or moving to page 3 and only then choosing the result.

Implicit feedback can also be trying to search over and over again with small differences in text – which Lighthouse can learn easily that the answer doesn’t exist or search is not working correctly.

Explicit feedback can come from end users during our chat feedback, or from agents that can’t close CRM tickets since they were unable to provide the customer the right answer.

You can learn and tune the results based on this, and the upcoming answers that the system will give will automatically improve by this.

Lighthouse Insight gives more details on which areas should be strengthened and which areas are a lack of attraction. By using the tool you can ask the content managers or any relevant, professional member in your organization to improve it.

Even integration with a CRM can be your wisdom of the crowd by understanding which areas are the hot areas that customers are looking for. By that we can enrich the knowledge and the attractiveness of the items in order to have more traffic, which means at the end of the day more sales.

Some people will call it machine learning, others will call it artificial intelligence – at the end this is the ability to combine existing content and knowledge with feedback and interfaces with human and crowd directions to improve the results.

Using all of the above and much more technical and behavioral techniques that resides within our product, transform Lighthouse from a static, common knowledge base to a critical and live system within the organization that continually improves during the implementation.