Generative AI has swept the imaginations of both consumers and business leaders with remarkable speed. Nearly two years after the change that began with the public release of ChatGPT, businesses are still sorting out how to take advantage of Gen AI capabilities while limiting risk.
Gen AI offers enormous potential for efficiency, knowledge sharing, and analytical insight in the contact center. Many companies are approaching Gen AI cautiously, embarking on use cases that are employee-facing or employee-vetted, rather than presenting generated content directly to customers. This also provides the opportunity for employees to label output to help the AI improve.
There are several ways to work with Gen AI and LLMs, from SaaS applications with embedded Gen AI to custom-built LLMs to applications that bring in Gen AI and LLM capabilities via API. Public and open-source models can be fine-tuned by training with specific domain or organizational data, while Retrieval Augmented Generation (RAG) is a newer choice for augmenting output with proprietary data that remains protected and separate from LLMs.
It’s important to note that Gen AI doesn’t replace what came before. Taking a hybrid approach to AI allows Gen AI to work synergistically with more traditional AI like Conversational AI, Natural Language Processing (NLP), and Predictive AI. Employing these different methods where they make sense maximizes impact, value, and compute efficiency. Companies can continue to rely on trusted methods for critical, high-frequency transactions, while exploring the powerful use cases that Gen AI can enable.
How Gen AI Will Enhance the Contact Center
That’s a very brief explanation of a very complex technology – one that is evolving as we speak. So how can you wrangle Gen AI for risk-mitigated, efficient, and valuable use in customer service and the contact center? A few common use cases are emerging:
- Analysis and insights: Gen AI can help put data analytics into easily readable language, pulling out insights and suggested actions.
- Agent assist: Gen AI can work hand-in-hand with Information Retrieval AI to extract information from multiple knowledge sources to help agents craft simple responses to complex questions.
- Interaction summarization: Gen AI can save precious agent minutes by generating call and text interaction notes that agents can then review for accuracy prior to submitting into their system.
- Social media: Gen AI can generate social media interactions that can be vetted by agents before being posted.
- IVA app development: Gen AI can create sample phrases for new intents, suggest alternative ways to ask questions and provide responses, extract common customer requests from transcripts, and generate code.
6 Criteria for Gen AI Investments
Gartner predicts that by 2025, Gen AI will be embedded in 80% of enterprise Conversational AI offerings, up from just 20% in 2023. Gartner also estimates that by 2025, 30% of Generative AI projects will be abandoned after proof of concept.
So, yes, Gen AI is coming to the contact center. But how can you avoid common challenges and concerns with implementation so that you avoid being part of the 30 percent? Few companies have the time or resources to train their own models, and most will instead rely on vendors for Gen AI capabilities like those outlined above.
Gartner has identified several criteria for selecting a Gen AI partner. Let’s look at these in a bit more detail.
1. Customizability is critical to receiving output that is relevant to your business and avoids outdated and hallucinatory responses.
Ask: How do you enable customization and to make sure the system interacts with my customers in a way that reflects my brand? How do you guarantee that a response will be what I expect? How can I review responses as we’re building the system? Is it possible to create customized workflows?2. Integration extends the value of your use cases, such as by connecting Gen AI capabilities with your contact center platform, CRM, or analytics/BI tools.
Ask: How do you enable integration with our key systems? Is it an API model? Is there integration assistance from your team?3. Privacy, security, and data retention should be top of mind to protect your proprietary data and your customers’ protected information, and to stay in regulatory compliance.
Ask: How will my prompt data be accessed and stored? Will my data be used to train your LLM(s)? Are my prompts shared across clients? Will my business processes be shared or used to develop similar solutions? Are you compliant with the key regulations in my industry (such as HIPAA, PCI-DSS, or GDPR)?
4. Pricing can come in several different flavors, such as subscription or pay as you go. Some features may be included with your overall Conversational AI platform, but some vendors may have additional costs, like fine-tuning and infrastructure. Avoid surprises.
Ask: What is your pricing model? What additional services may cost extra? How do costs increase if the scale increases?5. Deep experience with different types of AI, including Gen AI, indicates that a vendor isn’t just jumping on a trend, but knows how to build hybrid AI capabilities that ensure the right AI technology is being used for the right use case in order to boost performance, accuracy, and efficiency.
Ask: How long have you been working with AI? What does the AI stack powering the app look like? What Gen AI capabilities have you put into production? Do I actually need Gen AI for my use case or is there a better solution? How do your current offerings reflect past roadmaps (i.e., did plans come to fruition)? What is on your road map now?6. Maintenance and support are critical in a long-term engagement. The technology and your business will evolve — and your Gen AI capabilities need to keep pace.
Ask: What does your ongoing support look like? Does it require additional fees? How are the AI capabilities improved over time? What happens when business needs shift, such as new product lines, acquisitions, new locations with additional languages, or changes in scale?
At Interactions, we are working on each of the aforementioned capabilities with several pilot projects. As a pioneer in Conversational AI with 133 patents and counting, we’re excited to bring Gen AI to the contact center for deeper efficiencies, improved agent experiences, and more measurable effortless customer experiences and actionable improvements.
To learn more about how AI is affecting contact centers both today and tomorrow, read our report Shaping the Future of Customer Care: Innovative Self-Service.
