We have all seen that AI systems are impressive. We have also seen they can mess up. There are many stories of wrong information being shared – some humorous, some dangerous and some having severe financial implications. The intelligence in AI has its limitations. That limitation is human intelligence (HI). 

Human AI partnership

The days of speaking in terms of AI and humans competing or AI taking over from humans have passed. Conversational AI solutions are most effective when humans and machines work together in partnership. Leveraging and balancing the complementary strengths of AI and HI unlocks the full potential of AI with safety, ethics, and trust as the overriding conversation principles. When humans are actively involved in the AI process the outcomes can be trusted and aligned to brand values and needs. 

Three human AI relationships

In conversational AI there are three main human and AI relationships: human-in-the-loop (HITL) and human-on-the-loop (HOTL) and human-out-of-the-loop (HOOTL). In each of these there are different levels and types of human involvement. 

Human-in-the-Loop

Creative Virtual has HITL systems integrated into its V-Studio platform which supports highly customized brand aligned solutions. When HITL is part of the conversational AI solution process, there is active and continuous human involvement in every stage of the AI process flow, with the end-user only seeing the results once passed human review. 

The continuous human input guiding the AI model, results in future output and processing being more reliable, accurate and relevant. HITL is particularly important in use cases where the highest levels of accuracy are required, for example in financial services, healthcare, and legal services. Organizations that rate accuracy, reliability, and ethical considerations as priorities need HITL tools and systems to be available as a standard conversational AI platform feature. 

In addition to HITL, Creative Virtual’s conversational AI solutions allow for HOTL to be adopted in instances where speed is important and the system is handling many tasks by automation, such as in the contact center with Agent Assist deployed. 

Human-on-the-Loop

With HOTL humans only intervene when necessary and act as a monitoring service. The results of the input are presented directly to the end user and humans only intervene afterwards unlike HITL where nothing is presented unless pre-approved by a human. Whilst humans play more of a supervisory role in HOTL, and corrects output errors after generation, the corrections are fed back into the system, to allow for retraining and used for future outputs to ensure continuous accuracy improvements. 

Human-out-of-the-Loop (Agentic AI)

HOOTL are systems that have no human intervention. These systems take completely independent actions based on training and data and make decisions autonomously. HOOTL conversational AI is most used when the interaction is considered routine or structured, and when there is low risk of harm or misinformation. It is important to note that when using HOOTL it is not without humans’ involvement. Humans will have been involved with setting up the system, parameters and content, and there will be an embedded action to escalate to a human if required. 

HOOTL conversational AI solutions include chatbots helping reset a password, or a technical process. HOOTL chatbots handling billing queries, top up plans or plan changes are often used by telecommunication companies. E-commerce companies often use HOOTL chatbots in order tracking or returns and refunds. And in financial services, banks often use HOOTL powered conversational AI in balance and transaction history type enquiries. 

Human and machine harmony

AI is not taking over from humans. AI is not competing with humans. AI and humans are partners, working together harmoniously to create better customer and employee experiences. 

Human-first approach

At Creative Virtual we have, for over 20 years, taken a human-first approach. Our V-Studio platform is designed to be user-friendly, our V-Person chatbots, voicebots and virtual agents are designed to create better customer, employee, and contact center agent experiences. We ensure that human needs and values are integrated into the process to ensure transparency, trust, inclusivity, accessibility, empathy, and ethical AI. We enable humans to always be in control of the AI, and AI defers to humans when it is unsure. 

Creative Virtual’s human-first approach ensures better experiences, higher trust, and ethical and responsible AI use.