Exploiting AI to augment experience outcomes
The great channel shift:
2020-21 crisis has caused a great acceleration in the adoption of digital with interaction, commerce, and service all moving to online channels.
Respondents to a survey by McKinsey mentioned that they are “three times likelier now than before the crisis to say that at least 80% of their customer interactions are digital in nature”1.
The rate of increase has been unprecedented and has caused stress and increase in costs and investments in sales and service channels creating a strong need to embed AI-led automation interventions in the value stream across Experience, Sales, and Service.
Here are a few examples of how companies are applying AI to deliver Intelligent Connected Experiences across different aspects ranging from Marketing, Search, IT Services, Digital Assistants, and AI enabled CX transformation:
1. Delivering personalized experience at scale
With the dramatic increase in digital channel usage, marketers are scrambling to get their companies’ unique memorable message out to the world. Marketers and are trying to evolve new strategies to ensure customer loyalty but have limited resources. Using technology at scale to deliver real-time dynamic personalization is one way to adapt to this new normal.
AI technology embedded in CMS coupled with a strong CDP (Customer Data Platform) today helps capture the interaction between the customer and the content with the relevant engagement data such as customer attributes, the outcome of an engagement, content characteristics that evoked the response (color, topic, headline).
AI enables the discovery of these correlations and then delivers the optimal tailored combination to the customer on multiple channels to drive conversions in online Commerce. It is not possible or cost-effective to do this at scale using traditional methods. According to Adobe, these technologies can deliver upwards of a 15% engagement increase in engagement on personalized content2.
Our partner Adobe has designed Intelligent Services to provide marketers easy access to these advanced capabilities in the form of Customer AI, Attribution AI, Journey AI, and Content AI. For more information on these capabilities, please explore Intelligent Services by Adobe by clicking here.
CMS/DAM and AI:
Machine learning models trained on a large set of data enable auto-classification for existing and fresh content.
This is useful in CMS for
Automatic tagging of images
Automation tagging of content (inferring topics/themes from content, and suggesting classifications) & SEO recommendation.
NLG techniques (GPT-3) to enable Assisted/Augmented Copy Writing3.
GPT stands for generative pre-trained transformer and is developed by Open AI utilizing deep learning techniques to generate outputs such as content or even code (for restricted domains). For example, Headlime is a tool for marketers to write better copy (or generate landing pages to improve conversions) faster with AI assistance. The marketer can specify the product type, audience, creativity hint, and tone of voice (professional, informal, etc), description of the product (100 words), and then ask the tool to generate copies.
Intelligent search as part of Digital Asset Management software (DAM): Searching content within videos, automatic (and accurate) voice-to-text transcription of video content, or identifying moments of high interest in videos. As the volume of digital content increases, AI will be an essential component to handle these challenges at scale. For instance, this broadcaster had implemented “Contextual moments” ad-tech that helped advertisers place their ads in the context of a unique scenario or content in a program that is relevant to their brand. As per the broadcaster, this feature resulted in improved brand performance metrics including uplifts in spontaneous awareness (+34%) as well as positive brand perception (+12%) and purchase intent (+13%)4.
AI enabled service at Contact Center: Analysing the tone and sentiment of content and suggesting improvements.
Coforge company SLK Global Solution uses AI on customer conversations providing an agent with a live view of metrics such as customer satisfaction, agitation, compliments, rudeness for both customer and agent. This is achieved by using an integrated streaming pipeline of real-time transcription (speech to text) and sentiment analysis powered by a domain knowledge graph.
2. Using AI in Enterprise Search:
Financial institutions have challenges around data accessibility as a large amount of data is unstructured. These institutions want to exploit the value present in their data so their employees, such as customer service agents, can find the information they need quickly.
Traditional approaches have worked with consistent formats and matching keywords directly to their appearances in enterprise documents. In the last few years advances in OCR and computer vision have allowed for scanned documents to be indexed and made searchable. However, as firms moved to digitize, they had to migrate a large volume of structure and unstructured content. Parsing and applying metadata manually was making this a non-starter.
Today, a combination of AI and Semantic technologies (Industry ontology, semantic search) helps in automating content categorization, metadata inference (content subject/topics, main keywords in context of concepts), and application of metadata at scale across both structured and unstructured content.
We have built an AI solution, “EasyAnswers” that customers can bolt on top of their current content management and search infrastructure to provide meaningful and precise answers instead of frustrating search results. Our solution forms a dynamic persona of the visitor by analyzing the content viewed (links visited) by the user before invoking the search query. This helps EasyAnswers to understand the intent of the user with better contextual understanding. EasyAnswers is easy to bolt on top of the current CMS (such as Sitecore or AEM) and Search infrastructure (such as SwiftType). It augments the search results in real-time using the dynamic persona insights and NLP techniques to display “easy answers” or a “highlighted span of text that matches the intent” instead of the traditional search results that need to be further analyzed after clicking a search result.
3. Using AI in Tech in IT Services: Reimagining a familiar workflow
Wireframing/Prototypes: Uizard is an AI-powered wireframing tool that seeks to democratize the capability to create interactive prototypes quickly and effortlessly for web and mobile apps.
Uizard uses techniques from machine learning and computer vision to translate low-fidelity wireframes to high fidelity sketch mock-ups or generate front-end code automatically (supports both web and mobile targets).
Code/App Generation (Prototypes): Taking the GPT-3 technologies and the above use case forward, AI today can help generate layouts of the UI in code. This does not replace a developer but augments the developer experience so that UI development effort is reduced, and the developer can focus on delivering business functionality.
4. Using Digital Assistants: Automate the routine. Provide an “easy button” to adopt recommendations.
PDA (Personal Digital Assistant) is an old term that was ahead of its time. While RPA today serves the enterprise market, it is possible to imagine background agents automating the routine for all individuals. Today, Microsoft has an agent that summarizes your day and items of focus in your inbox, much like an assistant. Gmail has agents to recommend automatic un-subscription. SMS applications on a consumer’s device today serve up reminders (on late bills) and other recommendations on basis of data the consumer has agreed to share. Robo-advisory services help provide automated investment advice to subscribers and rebalance their portfolios.
5. Using domain-specific AI capabilities to transform experiences:
Coforge partner Google provides a niche solution, Lending DocAI, to transform the home loan experience by automating document processing for mortgages.
Clients can customize parsers (such as 1003,1040,1099, etc - mortgage lending and tax parsers) and help reduce errors and faulty information during intake. This solution will help reduce cycle times and provide concrete benefits in a digital lending solution with benefits such as faster loan processing times, fewer document intake errors, and lower origination costs5.
6. Transforming customer experience using machine learning:
A Property and Casualty Insurance client based out of the US transformed the shopping experience for home loans by reducing the time taken for a quote from an average of 15 minutes to 5 minute, leading to a 300% increase in quotes, and a six-fold increase in its volume of new business.
In the previous solution buying home insurance for this client showed stress points for the agent and the customer due to the multiple interactions needed to capture all the property characteristics and then a further dependency on the underwriter's approval.
The new solution involved combining internal first-party data with 3rd party data from a data provider (for properties in specific cities) and an offline predictive analytics model to pre-score the properties in a city.
This helped the client reimagine the online purchase experience by providing single-click, pre-underwritten quotes. Additionally, the marketing team could also run targeted outbound campaigns for 12% of the households identified as good risks.
The key takeway:
The examples provided above demonstrate how AI can be utilized across Marketing, Search, IT Services, Digital Assistants, and Revenue Channels to augment experience for Customers, Employees, and Partners.
Key themes to reflect on: Scale, Democratization, Reinforcement
AI helps to do things at scale through automation. Because of the scale involved, automation is essential as tool-aided manual approaches will not scale. The examples in the first point on Digital Marketing and CMS/DAM/AI elucidate this point.
AI capabilities delivered as an API or a Service democratize this capability to drive usage at scale. Affordable universal access to advanced capabilities will drive innovation or reimagination:
What could not be done till yesterday can be attempted today
What is being done or delivered in a specific way today can be reimagined for tomorrow
The example mentioned in the previous section around transforming insurance shopping experience using machine learning & third party data provides an example of reimagination of CX.
As usage increases. more data is made available to improve the service. This will improve AI models and benefit the entire ecosystem, driving continuous reinforcement and greater adoption till the service finally approaches “utility” status.
We can help you implement Intelligent Connected Experiences:
In addition to assets and frameworks, at Coforge, we have a dedicated team to help deliver ROI on AI initiatives for clients wanting to apply AI in the space of transforming Customer, Partner, and Employee Experience. This is an emerging space where the benefits may not be clear before the actual application. With our catalog of real use cases and actual case examples (both internal and from our partner ecosystem), we can help refine your problem statement, crystallize the benefits, and provide concrete solutions to your problems in a collaborative model.
Please reach out to Our Digital Engineering team to discuss your needs.
- Adobe announces AI Powered Customer Experience Capabilities
- Headlime AI Copywriter uses GPT3
- Broadcaster announces global first artificial intelligence tv advertising
- Roostify reduces mortgage processing times with google cloud