With technological advances and complex products, customer support is caught in a perennial dilemma. Customers desire quick resolution but qualified human support is incredibly expensive. Companies throw documentation at users, invite them to post to community groups and create tiered levels of support; all in an attempt to balance quick resolution with expense.

Will AI finally get us out of the conundrum? First, let us first look at the roadblocks:

  1. Large language models (LLM) are trained on publicly accessible data which is often not current. 
  2. LLMs hallucinate and will provide inaccurate responses.
  3. LLM responses can be generic. Even when the response is accurate, it may not be useful for specific issues.

There is a plethora of chatGPT wrappers that all suffer from these well known issues.

There are products that claim to resolve these issues. Not one of those claims is accurate. LLMs and their responses are rooted in statistics, therefore an absolute guarantee is impossible. Techniques such as prompt engineering and similarity based filtering can reduce these problems not eliminate them.

The goal is to maximize the likelihood of a response that resolves the issue. Context is critical to resolving customer support issues. That is what SureSteps is designed for. This is why our AI response is so effective. That is also why our approach enables the best human response!

The SureSteps Approach

Organizations can upload user manuals, guides and any relevant documentation. This proprietary data is then automatically incorporated into the AI response. We can even implement an internal pipeline that automatically keeps the documentation current.

Furthermore, SureSteps enables the use of AI at every step of a step-by-step procedure. In addition, the steps can be configured for users to provide additional information pertaining to their environment,  All of this data is made available to the AI assistant. As a result, the assistant gets the most precise context and provides an accurate answer.

The Human Element

The current support paradigm expects our users to discover the solution themselves. We throw links to documents and tools to help users resolve the issue. Only after the user has exhausted their own effort do we want them to talk to a human. When the indomitable user finally makes it to a human, they are faced with the daunting task of explaining the problem and what they have already done.

The SureSteps approach tackles this problem too. The exact procedure followed by the user is visible to support staff who then knows what the user has completed and where they have stumbled. The support staff now has the context to provide a precise answer.

Not only does the SureSteps approach maximize the likelihood of success of the AI assistant, the same approach simplifies the job of human support!

Latest posts