In a previous post, we stated the obvious. A business needs to service the information needs of existing customers. It needs to generate awareness in the marketplace of a product or service. It needs to educate and from those in market, it needs an in – a sales lead.
We then went on to discuss why:
- It is increasingly difficult to deliver information online and it will only get worse.
- B2B customers are moving away from the internet and towards finding information in trusted libraries that they control.
The technology to facilitate the move to trusted data sources has recently become available. It is driven by the rise of deep learning-based applications.
In this post, we discuss small customised knowledgebases as trusted data sources. We cover how they bridge the gap between customers and suppliers, fostering a closer relationship. Our discussion concludes with a brief look at challenges B2B organisations face when creating, installing, and populating knowledgebases.
What is a knowledgebase?
A knowledge base is a centralised self-service online library of information about a product, service, or topic. Contributors who are well-versed in the relevant subject add to and expand the knowledge base over time.
Knowledgebases facilitate information collection, organisation and retrieval. When a user inputs a query, software attempts to deliver relevant answers based on information in the knowledgebase.
But haven’t We Been Here Before?
Yes, we have.
Intranets first emerged in the mid-1990s. As the World Wide Web developed, organisations realised they could build a localised version for internal use.
The initial goal was to increase employees’ productivity by simplifying access to company documents, reports, product and customer support information.
Various tools were added to intranets over time to encourage employee communication.
The user interface with the intranet was a problem that was never properly resolved. Today, many of the remaining intranets feel old clunky and slow. You can read more on the development of Intranets and user experience issues HERE.
We predict suppliers to large companies with many hundreds of potential users and a distributed workforce will develop the next generation of intranets. These will be powered by Chat-style interfaces.
It will be expensive to implement and run these systems, but for some large organisations, they could pay for themselves several times over.
In this post, we focus on relatively small customised knowledgebases designed for a specific purpose.
How do Large Language Models (LLMs) Help?
As discussed above the intranet concept was valid but had two major problems. One was the user interface (clicking through multiple pages/links to find information). The other was data preparation, formatting and loading,
A user needs an efficient and reliable way to interrogate a knowledgebase and that’s where ChatGPT comes in. ChatGPT (Bard and others) allow a user to interrogate a knowledge base using standard text.
ChatGPT understands words (tokens) and context. It learns over time depending on the data and prompts it receives.
Of course, it is not that easy. Although setting up and maintaining a knowledgebase is potentially simpler than an intranet, developing the instruction set, formatting and training is crucial.
It is not possible to simply upload a mass of data, run training and expect this to change (update) the LLMs (ChatGPT) understanding.
Knowledgebases – The Customer
Now, back to the fundamental problem. Let’s assume the customer has an idea of what they want and have a grasp of which suppliers might be able to supply that product/service.
They want some initial questions answered, but as outlined in our earlier post, the internet is a hopeless research tool.
What if the customer could access separate knowledgebases provided by supplier A, B, C and D and ask questions via a simple text-driven interface? Would that solve their problem?
They could pick up the phone and ask, but they won’t. This is the principle behind inbound marketing.
At this stage, the customer’s questions will be relatively simple as they are nowhere near ready to buy. The supplier could provide a knowledgebase to answer these early-stage questions. With access to that knowledgebase tightly controlled.
Of course, if a customer does not know that supplier E can offer a solution, they are not considered. Knowledgebases do not resolve the awareness (demand generation) issue.
Knowledgebases – The Supplier
From a marketing/sales perspective, there are obvious downsides for the supplier with the above approach. The customer needs to know the knowledgebase is available, be able to interrogate it reliably and trust its output. These issues we will address in our next post.
The potential benefits to a supplier are:
- Small and specific knowledgebases can be tailored to address the unique challenges of a particular industry, or customer segment.
- As information in a knowledgebase is highly relevant and useful to the audience, it leads to better engagement and customer satisfaction.
- The knowledgebase can demonstrate expertise and a deep understanding of the industry or niche. This establishes thought leadership and makes the company more attractive to customers.
- Like the introduction of the internet, those first to the party will have a competitive advantage.
There could be one (or more) customised knowledgebases for customer consumption and one (probably several) for internal use.
Let’s look at the potential benefits of knowledgebases for the supplier from an operational perspective.
- Creating and maintaining a specific knowledgebase can be more cost-effective than building a large, all-encompassing one.
- Small customised knowledgebases can be scaled up, if it makes sense to do so.
- Centralising knowledge in internal knowledgebases can foster innovation and collaboration.
- Knowledgebases equip sales and support teams with up-to-date information.
- A small and specific internal knowledge base can significantly reduce employee time spent searching for information.
Implementing small-scale, application specific knowledgebases is not without its problems, these include:
- Specifying the use case
- Developing the instruction set and training
- Tailoring knowledgebases to the task and/or customer segments
- Curation – Garbage In = Garbage out
- Populating the knowledgebase with content, structure and data management
- Integration with existing systems, for example, CRM
- Data security
The larger the data set and the less defined the use case, the more significant the issues. We will discuss each of the above in future posts.
Knowledgebases are not a new idea. ChatGPT, with its advanced language processing capabilities, is the game-changer. It offers efficient, real-time assistance and enriches knowledgebases by facilitating seamless communication.
From the customer’s perspective, knowledgebases deliver autonomy, immediate answers, and a personalised experience.
From the supplier’s viewpoint, customised knowledgebases streamline operations. They reduce support costs and elevate customer service quality. They enable data-driven decision-making and scalability.