The 4 key elements missing from your conversational AI strategy

How chatbots fall short in the enterprise, and how to catch up

By | 5 minute read | April 8, 2021

Chat bubbles

In 2021, enterprises who wish to implement a chatbot, virtual agent, or conversational AI have no shortage of options. Informed buyers will take the virtual agent’s feature set into account. These products will range from simple chatbots that regurgitate pre-programmed dialog in a single channel, all the way up to expansive conversational AI platforms that operate across the entire organization. Given the glut of options available, business leaders may be tempted to find a range of solutions that meet their technical requirements and base their decision largely on price. But this could be a critical error. In the long run, less robust conversational AI solutions can cost organizations more.

There are four broad considerations to make when your organization is looking to adopt conversational AI.

1. Compliance and data privacy from day one

Governments and industry oversight committees impose continuously shifting regulations on businesses, and for good reasons. When interacting with customers and their data, your organization will need to think about all the potential points of regulatory exposure involved.

Some countries maintain strict requirements on how organizations store and use basic identity information, web data, health data, and much more. Some conversational AI products already have baked-in compliance with regulatory frameworks like GDPR, HIPAA, and SOC2, and others don’t. Even if your initial use case doesn’t require data privacy, as additional use cases emerge, the cost of extending an existing conversational AI platform will be significantly lower than implementing a more robust solution later — not to mention the cost of potential violations, audits or reputational damage.

In addition to compliance, also ensure you understand how the solution uses and where it houses customer data, as this can also present potential issues given regulations like GDPR and other regional data privacy laws. And few conversational AI providers have options for data isolated single tenant environments and on-premises deployments.

 2. Compatibility and integrations for maximized value

Many organizations focus their initial conversational AI use case on fielding customer queries, since that is often the biggest pain point or represents the highest value. With that in mind, the platform must be able to integrate with all the channels your customers are using, and to maintain that conversation in the customer’s channel of choice without forcing them to switch to an alternative channel. You need the ability to respond to each unique customer with the most relevant information about them and their accounts. Seamless integrations with your knowledge base and backend systems, as well as customer care platforms and ticketing systems, will help them find answers and get help without waiting for an agent.

To do this, you need an open conversational AI platform that is designed to connect to these systems to solve all kinds of customer problems, such as locating a delivery, or getting product instructions, or creating a travel alert on their bank account. This is also a great opportunity for improving employee experiences: providing contact center agents with faster responses based on the customer conversation, or integrating with HR systems to onboard employees and manage vacation requests, or creating IT service outage alerts through Slack.

Over the next few years, organizations will embed conversational AI as a user interface within other applications in hundreds of ways. But not every conversational AI is designed to integrate seamlessly with in-house or third-party applications. You’ll want a solution that is easy to plug into other systems if you want to maximize your value — not just for the use cases you have today, but the ones you haven’t even thought of yet.

3. Scalability to enterprise demands

If 2020 taught us anything, it’s that unexpected events can occur with short notice. Systems designed for “business as usual” can be overwhelmed. But in a world defined by uncertainty, you’ll want a flexible conversational AI platform that is able to quickly scale up and down based on current demand. Your platform must withstand extreme spikes in call or messaging volume.

Your platform also needs to effectively categorize, respond and route customer issues appropriately. If you don’t solve a customer’s issue the first time, they will keep trying to contact you — or take their business elsewhere.

To ensure you are resolving and routing issues appropriately, ensure that your conversational AI platform has industry-leading intent detection: that can decipher what a customer is trying to achieve, clarify the request if unclear and either respond or route appropriately. Based on the volume of interactions in your enterprise, you also need a platform that can quickly analyze and improve with each customer conversation, ideally without manual training or tuning.

4. Standardization for consistency and cost-effectiveness

Instead of implementing distinct chatbot solutions throughout the organization to handle small tasks, you can capture more value by introducing a single, all-inclusive conversational AI platform. The platform approach enables your enterprise to still maintain multiple virtual assistants, for various internal and external purposes, ranging from HR and IT support to customer self-service and product assistance.

Many organizations initially took a more fragmented approach, adding dozens of chatbots to serve the different needs of each department. While this allowed each internal organization to solve its own need, it also created inconsistent experiences and difficulty with oversight and maintenance.

Instead, standardize them all on a single platform that can be used and updated across divisions. This will not only reduce maintenance costs, but also generate significant value by collecting a wider swath of organizational data in one place. Such a system can apply analysis to that data, providing insights that reveal ways to make the entire organization more efficient and effective. A holistic approach has the potential to grow with your business, paying dividends over time. More data means smarter AI.

A platform for long-term success

Every organization has unique needs and technology requirements. But thinking beyond the narrower scope of the initial conversational AI project will highlight many potential use cases, and potential risks. If an enterprise can take the long view, a true conversational AI strategy can build upon these initial projects and make subsequent efforts easier without the need to consolidate or deprecate multiple chatbots in the future.