Skip to content

9 challenges in implementing an AI Virtual Assistant in your organisation

With the proliferation of AI-assisted devices like Siri and Alexa in the home and chatbots online, there’s no wonder people in business are wondering what AI-powered Virtual Assistants can do for their organisations.

Virtual Assistants do offer huge potential, but it’s not enough just to introduce one and think that productivity and performance will go up, costs come down, and training become irrelevant.  To get the most out of a Virtual Assistant you need to prepare.  It’s not a case of install it and they will use it.

Primarily, it’s not really a question of the technical installation.  It’s more about considering the implication for procedures, practices, and ways of doing business across the organisation.  It’s more of a cultural than a technical change.

To help you prepare, here’s a guide to 9 key challenges of introducing an AI Virtual Assistant into your organisation and how to face them.

1. Overcoming resistance

Anyone reading the popular media might easily conclude that AI will transform the workplace, but often in ways that pose a threat to jobs.  There’s a widespread notion that many tasks will be automated and there’ll be no place for humans.

But that needn’t be the case.  Much of the promise of AI is about relieving employees of the repetitive, drudge work and freeing them up to be more creative and to perform to their real potential.  So, while from a budgetary point of view, AI might represent cost savings, from a more general performance point of view it’s about enhancing an organisation’s capability by a better, more efficient division of labour between the Virtual Assistants and the humans.  That’s the message you need to send to your organisation.

2. Securing the real benefits

It’s easy to make the assumption that simply introducing a VA you’ll enhance your organisation’s levels of performance and productivity while saving costs.  Introducing a VA isn’t particularly technically difficult, as they’re flexible enough to sit and operate within existing systems.  Integrating them with your practices and procedures, however, is a different matter.

Before deploying an AI assistant, you need to construct robust and credible use cases.  These should identify and justify the business uses and benefits of a VA.  Ask yourself where it makes sense to deploy your AI assistant.  And make sure you’re clear who the key stakeholders are and who has ownership over the project.

Consider where it can make the most impact.  You might start in a specific area – training, HR, marketing or customer services for example – and then aim to roll VAs out across the wider organisation.  Part of the analysis involves knowing your organisation’s limitations and those of the VA.  VAs in themselves are only as good as the environment in which they’re placed.  Imperfections in systems will only be magnified by an ill-chosen deployment of a VA.

3. Choosing your assistant

You ought to consider your Virtual Assistant in the same way you consider a human agent: as an employee asked to perform specific tasks.

Not all assistants or bots are alike.  The simpler ones operate on rules while more complex bots incorporate machine learning.  The difference between the two is expressed in degrees of predictability, responsiveness, and adaptability.  The basic ones respond to set commands.  But try and express a query in a way for which they’re not programmed to respond, and they won’t be able to adapt.

The more advanced chatbots employ machine learning and they can learn from each interaction or conversation, detecting preferences and making recommendations based on past requests.

It may be that you want a VA only to perform a very specific and straightforward set of tasks and then hand more complex tasks over to an employee. In that instance, the simpler version may suit your organisation.

4. Training and onboarding your VA

To pursue the analogy between virtual and real assistants, Virtual Assistants also need to be taught and supported.  It’s not just about loading them with information or giving them access to resources.  To be really effective, a VA needs to know the context and its audience.  It has to be fully integrated with existing processes.  This is a continuous process.  If processes, procedures or business goals change, the virtual assistant will require retraining, just like its human counterparts.

Who’s going to do this work of educating the VA?  This is one of the higher-order tasks that require human intervention.  The coaching of the VA can, however, be turned from an additional overhead into a business advantage.  In integrating the VA, you can at the same time make your business processes clearer, more robust, and transparent.

5. Employing a Virtual Assistant

What does it mean to work with a Virtual Assistant?  The fact that you have an assistant 24/7 that never tires, takes holidays or falls ill may seem like an obvious gain, but where does the division of labour fall?

Take an area where VA chatbots excel: fielding customer service queries.  What happens when the bot’s ability to help fails when a query arises which it’s not been taught to handle?  Here a human agent needs to step in and that handover between the two needs to be seamless.

There’s also the question of perception when you allow customers to interact with a VA.  Do you humanise and personalise the experience by giving the VA its own identity and personality, by using avatars for example?  Are you up front with customers by telling them they’re interacting with a bot and give them the option to switch to a human agent?

From a marketing perspective, you might also consider branding your VA to reflect your organisation’s identity.  You might decide your VA should be corporate or more fun or whimsical.

6. Using VAs for learning

Virtual assistants offer great potential for placing learning directly in the workflow, providing information and training at the point of greatest need.  But again, it’s not just about installing a VA and switching it on.  You need your L&D people to plan for this by providing training content in a format that’s more accessible to a VA.  Chunking content, considering micro-learning opportunities, making your resources open to searching and interrogation by a VA is vital.

Intelligent reuse and repurposing of content can bring real benefits and make training more relevant and give it greater impact.  With the right intervention from L&D, learning and training using a VA become more effective.

7. Integrating VAs into your existing systems and apps

Technically this integration isn’t especially difficult.  For example, chatbots can operate easily inside work management tools like Slack.  This enables the VA to have an organisation-wide presence.  There’s obvious application in the areas of HR, L&D, PM, Marketing and Communications.

Again, the real benefit comes in usage.  The VA offers performance support, but it has to be part of the workflow.  It needs to be embedded into day-to-day operations and working with a VA should be regarded as best practice.

8. Handling data

One of the benefits of using a VA is its ability to record and process data quicker and more effectively than human agents. But the VA’s data handling capability presents challenges. What do you do with it? How do you analyse it?

The potential for better understanding of clients’ and employees’ needs are considerable, but it will need proper consideration – a task to which you might assign employees freed up by the deployment of Virtual Assistants.

The other aspect is privacy and data protection. With new EU legislation in the form of GDPR, you will have to include VAs in your data auditing and security measures.

9. Future-proofing your VA

You’ve faced and overcome the challenges of introducing and integrating a VA into your organisation.  But don’t rest on your laurels yet.  The process of educating and monitoring your Virtual Assistant can’t stop if you want it to remain relevant.  There’s the necessary, cyclical process of analysing performance and receiving feedback and then updating VAs as policies, requirements and, business needs change.The upside of this overhead, though, is that the VA will become more efficient over time and provide increased performance benefits to your organisation.  And it can play an important role in the refinement and evolution of those changes.  This list of challenges isn’t exhaustive.  Different organisations have different needs and contexts vary.  But with careful planning and well-informed use cases, you can face down the challenges and reap the considerable benefits of deploying a Virtual Assistant in your organisation.

AI
artificial intelligence
chatbots
learning and development
Otto
virtual assistant

Got a learning problem to solve?

Get in touch to discover how we can help

CTA background