Companies have a wealth of data that they could use to share with employees in order to improve customer service.
This data is often buried in different systems, so staff must wade through tons of irrelevant data to find the information they need or figure out the exact search terms needed to access it. This is a time-consuming process, which means that customer queries are not resolved as quickly as they should be. Recent advances in generative AI (GenAI) offer firms an opportunity to offer a AI can improve customer experience in knowledge management (KM) .
According to Forrester Research, knowledge management software is software that allows employees to answer questions and customers to find answers through web or mobile self service. In the case of the customer experience, employees act as customer service agents. According to Kate Leggett of Forrester Research, vice-president, principal analyst, CRM and customer service, a firm can reap multiple benefits with a solid KM framework.
“KM reduces operational costs by empowering the customer to get rapid answers to questions, which values a customers’ time,” she says. It helps agents resolve issues faster, in a more personal way. Both of these benefits drive CX.” GenAI can be added to KM for additional benefits.
AI is revolutionising knowledge maintenance and creation by automating and improving knowledge practices. Leggett says that knowledge creation, sharing and improvement can be integrated into core CX processes. “GenAI-powered KM makes it easier to get comprehensive information from different sources in real time. GenAI can also be used to co-create knowledge, which helps generate new insights and speeds up the creation process.
Adobe Population Health improves CX with AI-powered KM
Adobe Population Health already uses AI to support knowledge and improve customer experience. The healthcare provider provides virtual and in-home care to more than 400,000 customers. The organisation provides a combination of social welfare and medical care to promote health and wellbeing.
Adobe Population Health provides medication and healthcare advice to many of its members, who are from underserved areas. Adobe Population Health also helps them with housing assistance, home modifications such as wheelchair ramps and food.
The firm is now using Salesforce Agentforce, mainly to focus on prompts, to support its clinicians. Before using the agentic AI platforms, clinicians would spend 15 to 20 minutes prepping for each member’s visit. This included checking for the latest information, such as recent hospitalizations or new medications, and verifying important medical details using past records by accessing multiple systems.
With the PromptBuilder in Agentforce, Clinicians can create templates to pull data from Health Cloud and patient portals, databases, and MuleSoft via MuleSoft, to generate care summaries within seconds instead of up to 20 minutes.
Adobe Population Health has begun to look at the next phase of the technology which will integrate knowledge management and the AI platform. The company has an extensive library of educational materials in its systems and website and would like to find ways to share them with members when they are most needed.
At the moment, when clinicians work with a member they ask a series questions about their health and social needs. If a member tells the clinicians that they smoke three packs of cigarettes per day, then a knowledge article, or educational material, could be sent to them.
Our clinicians and social work are amazing. They have a lot in their heads, but a person can only hold so many resources and information. Alex Waddell is the CIO of Adobe Population Health. “I’m thinking about how I can elevate our tech because, right now it’s either the clinician chooses, ‘I want to send this to them’, and ‘I will pull this up myself.'”
We’re working to have an agent who can say, “You just said this person smokes. Here’s some information about smoking cessation. Here’s how you should communicate this, offer them this education material.”
So that we can, to use a more appropriate term, attack the issue at that time. “Because when [the patient] leaves they may just throw away an educational material,” he says.
Waddell is also looking at how this could affect the business processes.
If we put our standard operational procedures in there, would you like an agent to tell you, “You’re going to be in for this type visit, here are the things you need to cover and here is some material on how you can do it?” You might be going to a cancer screening and want to know how to use a device you have. We provide training, but it’d be great to be in a position to give them more information.
Waddell met with members of the Salesforce team recently to discuss an agent that would help clinicians. They discussed transcription and summary from transcription which would make a big difference for the doctors and nurses working for Adobe Population Health. They prefer text boxes to electronic medical records with dropdowns and pick lists when they are consulting a member.
Waddell says, “It interferes with the focus on the members.” “We had discussed the possibility of building an agent who could listen and send data in the right places. If I put 129 over 89 in a member’s blood pressure, Salesforce will send out different messages, knowledge articles, or education material.
He says, “It is just removing the system and charting so that doctors can focus on each other and solve problems.”
Lloyds Banking Group improves CX with NICE
Lloyds Banking Group has a major digital project underway to enhance its products and services. In order to improve customer service, the company uses NICE’s GenAI Technologyv, internally known as Athena. This technology helps staff answer questions from customers quickly and efficiently.
In just 12 months, Athena was able to support 12,000 customer service colleagues who answer more than 25,000,000 customer calls per year. The tool helps staff with a variety of customer interactions including personal banking, fraud and disputes and bereavement. The company will now extend the tool to 43,000 employees by 2026.
Athena provides staff with a user-friendly tool that simplifies complex questions. It uses generative AI, which summarises information and procedures from detailed articles that would otherwise take colleagues time to read and digest.
By speeding up information retrieval, Athena reduces the time required to complete these tasks by half, says Suzanne Ellison. She is head of product-consumer relationships at Lloyds Banking Group. This allows our colleagues to focus on what really matters – meeting the needs of our customers and providing excellent customer service. By using Athena we combine human expertise with GenAI’s efficiency.”
Gina Whitty is the director of product development at GoTo Connectis a smart assistant that uses AI to provide the most up-to date information to human agents to help them meet customer needs quickly, accurately, and effectively. It does this as advanced receptionists streamline front-line service by automatically handling simple requests using insights from catalogues and articles.
AI-supported scripts can offer further CX improvements, as they are more precise and use keywords and intents. Whitty explains that they established a formula for text-based chats or live calls, and human performance was measured by how closely agents adhered to defined phrasings and protocols.
GenAI’s ability to learn quickly and in a free-flowing manner allows it to break away from this rigidity. “Companies don’t have to spend significant time and resources mapping out the right’ approach for each possible scenario, which is driving down a dramatic decrease in the time to value,” says Whitty. “Projects which would have taken at least a half-year to complete, if they did not take longer, can now start, be fine-tuned, and rolled out within three months.”
Whitty says that these benefits have led a rapid increase in experimentation with GenAI, and that this will only continue to grow as more businesses trust and appreciate its flexibility.
GenAI’s agile adaptation offers businesses CX improvements, but its fundamentally data driven nature also offers consistency. The parameters that smart agents operate within are determined by the information companies enter in knowledge catalogues and articles. Whitty says that smart assistants will deliver consistent services because they use the same core database. This also allows for greater scaling. “On top, it’s much easier to ensure that advice is always accurate and current by simply adjusting these data inputs.” Whitty says, “And with the confidence that error rates will remain low.”
Virgin Atlantic is aware of the potential for AI-driven knowledge management but has not yet reached a point where it can benefit.
The carrier is using Adobe Real-Time to collect knowledge materials and sees an opportunity to combine that with data from social media platforms, which are constantly generating vast amounts of content. This could help to better understand customer sentiment. Simon Langthorne is the head of CRM for Virgin Atlantic. While it’s too early to test out GenAI in knowledge systems, he can see the point where the company develops on the fly knowledge material that then leads to a personalisation offer or a change of content across the website. “From my perspective the future is when we create personalisation use cases on the fly and you can manage those experiences and those experiences are created for the customer’s benefit,” he says.
At the moment, GenAI’s ability create accurate materials is the biggest barrier. “It could create anything.” Langthorne asks, “How can I trust it enough if I don’t know what it is doing?” It’s knowing the guardrails when it comes to that, how do we build the guardrails? Knowledge is often dispersed across different platforms, databases, or departments making it difficult to locate. Leggett says AI systems need to be trained to retrieve relevant knowledge quickly to reduce the time staff spend searching.
“Teams often restrict the flow of information through access permissions and personal knowledge stores. Openness encourages critical thinking by allowing agents to evaluate AI outputs critically and strengthen decision-making.
