Posts

Building a Cohesive Virtual Agent and Live Chat Solution

By Mandy Reed, Global Head of Marketing

Virtual agents. Live chat. Some of you may remember the days when companies wanting to implement customer support on their website believed they had to make a decision: either a virtual agent or live chat; either automated self-service or human-assisted web chat.

This was the common thinking a decade ago when Creative Virtual integrated their virtual agent deployment for a leading telecommunication’s company in the UK with live chat, creating a seamless handover for users within the same template. Around the same time, an online financial services company in the US collaborated with Creative Virtual to add a virtual agent in front of their existing live chat offering to reduce repetitive questions being handled by live agents. Both of these companies were very forward-thinking in their approach to combining these technologies.

Now in 2022, virtual agents and live chat are seen as complementary tools instead of rival solutions. In fact, it’s become best practice to integrate the two in order to improve digital customer support. In the recent webinar ‘Deploy Chatbots to Meet Self-Service Demands’, Gartner analysts talked about the importance of seamlessly escalating customers from virtual agent to live chat agent to create positive, cohesive service experiences.

Instead of a conversation about which of the two technology options to implement, organisations are having discussions about how to incorporate both into their digital customer experience (CX) strategies. Companies that have already done so are seeing a variety of benefits to both their customer and agent experiences, including:

  • Customers are always supported with 24/7 access to self-service even outside of live chat hours or when all live agents are busy.
  • The most common questions are answered by the virtual agent, reducing the overall number of live chat contacts and the need for agents to answer simple, repetitive questions.
  • Virtual agents do the discovery work and pass a complete history of the customer’s conversation to the chat agent, creating a seamless experience for both user and agent.

Adding a virtual agent to an existing live chat deployment

If you already have live chat available to customers and are ready to add a virtual agent, you can jumpstart the project with transcripts from those conversations. By analysing your live chat transcripts, the virtual agent vendor can identify which questions can be answered without human involvement. This analysis can also identify the percentage of live chats that can be deflected with successful automation. It can assist you with identifying key metrics and help build your business case.

Many traditional live chat providers are now offering virtual agents and chatbots as an add-on solution. Unfortunately, often these vendors maintain their focus on driving usage of live agent support. Because of this, the functionality of their virtual agent tools is limited. It’s also not unusual for them to use a pricing structure for self-service that’s similar to their live chat, such as charging by transaction. This makes the virtual agent both poor performing and expensive.

Self-service tools from virtual agent-first vendors deliver better user experiences and more reasonable pricing models. A sophisticated conversational AI platform will support integration with a variety of different live chat providers. This means you can deploy a highly functional virtual agent with seamless handover to your existing live chat solution, providing both successful self-service and easy escalation to human-assisted support.

Adding live chat to an existing virtual agent deployment

If you already have a virtual agent available to customers and want to add handover to live chat, you should start by talking with your virtual agent vendor about your options. Confirm that your existing conversational AI platform supports integration with live chat to handover users within the same template and pass a full conversation history to the live agent. If it doesn’t, then it’s time to find a better virtual agent solution.

Ideally, you’ll already have both a virtual agent platform with flexible integration options and a vendor you trust with expert insights. If they offer their own live chat product, explore that option first. Ask them about their integration experiences with different live chat providers and how their joint solution is better. Ask them for live examples of other similar deployments and for existing client references to get additional insights.

You can use your virtual agent transcripts and metrics to assist with setting up custom rules and triggers for users to be escalated from self-service to live chat. How users have engaged with your virtual agent should inform the set-up of live chat to ensure you are adding it in a way that will deliver the best experience possible for your customers.

Adding both a virtual agent and live chat or changing providers

If you need both virtual agent and live chat technologies, begin your selection process with the virtual agent. Customer engagements will start with your virtual agent, so you want to ensure you are delivering a positive self-service experience even if escalation to a live agent is necessary. A ‘bad’ virtual agent frustrates users and makes the job of your live chat agents even more difficult.

Use the advice shared above and in this Guide to Selecting a Virtual Agent or Chatbot Vendor for selecting a virtual agent platform. If you have an existing virtual agent, either live or in-progress but never deployed, be sure to ask how a new vendor can reuse it so you don’t lose that investment. If you have live chat or contact centre transcripts, also ask about using an analysis of those to jumpstart a new virtual agent.

Then once you have found the conversational AI platform that best fits your organisation and goals, explore the live chat technologies that integrate well with your virtual agent choice. Use the expertise of the virtual agent vendor to help with your selection. They will know from experience which live chat systems deliver the greatest results for your industry, use case, etc. when integrated with their self-service tools. Test some existing joint deployments and talk to other companies using both technologies about their experiences.

Developments in conversational AI over the past decade have enabled a more seamless integration of automated self-service and human-assisted support. With the right technologies, organisations can take advantage of these advancement to deliver improved end-to-end experiences for both customers and agents. Cohesive, convenient customer service is key to building brand loyalty and reducing customer churn. It can provide real business value today and give you a solid foundation for the future.

Set your Chatbot up for Success with the Right Budget

By Mandy Reed, Global Head of Marketing

I think it’s safe to say that we all love to get a good deal – or at least feel like we are! – and save a little money whenever we can. I think we can also all agree that are certain times when going with the cheapest option or skimping on a line in the budget doesn’t pay off in the end. You get what you pay for, right?

A conversational AI solution is one of those projects that your organisation should be careful not to underfund. Without the proper level of financial support and ongoing funding, you will never achieve success with a conversational AI project. A chatbot or virtual agent that is treated like an unimportant side project not worthy of dedicated resources will perform like one. It will provide a poor experience and drive users away.

If you want to have a conversational AI tool that increases customer satisfaction, contributes to cost savings, generates new revenue, and improves efficiency and productivity, then your company needs to make a commitment to invest in those goals. That commitment starts with properly budgeting for the cost of the technology, the cost of developing and implementing your customised solution, and the cost of ongoing maintenance.

Budget for the technology

Earmarking a portion of your budget for the technology itself should be a no-brainer. Without a working conversational AI technology, you have no working chatbot! However, the amount you allocate for this really depends on how and where you plan to deploy your solution.

You should take into consideration your initial project plan as well as how you may want to expand and scale it in the future. Identify your integration points, calculate how many concurrent users you anticipate, estimate how large of a knowledgebase your content will require, and select the deployment channels that best serve your users. All these elements will impact which technology is a best fit for you and how much you will need to budget for that technology. An experienced conversational AI vendor or consultant will be able to provide guidance to help you scope out your technology requirements.

Budget for the development and implementation

While there are conversational AI solutions on the market that can be deployed straight out of the box with very little configuration, they will provide a very generic, basic engagement. To really create a positive experience and be successful, a chatbot needs to be customised for your organisation, use cases, users, and goals. This customisation should include integrations with other systems (such as your CRM platform, ticketing systems, or live chat) and conversational flows tailored for your users. You also want to ensure that the chatbot can respond to questions about your products, services, and procedures with specifics unique to your business.

Unless your organisation has a team with experience creating successful chatbots with the technology you select for your identified use cases and/or channels, attempting the building and implementation of your tool internally will be a mistake. Working with an expert vendor is more cost-efficient because they already know what they are doing so you aren’t paying them to figure it out. This also means you cut down on the development time and get better, quicker results.

Budget for ongoing maintenance

If a conversational AI provider tells you that you can configure and deploy a chatbot with their platform and then leave it alone to do its thing, cross them off your list immediately! Companies that invest in those solutions quickly learn that they have wasted money on empty promises. The truth is that the ongoing maintenance of conversational AI tools is what enables long-term success.

Newly implemented chatbots need more attention than well-established ones, so that needs to be reflected in your budget. During that initial period, engaging the expert vendor’s team is recommended for the same reasons you should work with them during the building and implementation step. However, after that you should have options for moving all or some of your chatbot maintenance in-house. If you choose to do that, factor into your budget costs for those internal staff members and any related trainings or licenses.

 

If the price tag of a quality conversational AI solution creates some hesitation within your organisation, consider the cost of deploying a chatbot that delivers a negative, frustrating experience for users. Putting time and money into a tool that your customers or employees won’t want to use – even if it is just the bare minimum investment – is a misuse of resources. Not only are you wasting your budget, but you are harming your digital experience and eroding confidence in your business.

Check out the Guide to Enterprise Conversational AI Pricing for more insights on budgeting as well as typical pricing models, average costs, and calculating your return on investment.

Resolve to Make Your Conversational AI Project Healthier this Year

By Mandy Reed, Global Head of Marketing

The new year is here and that means it is time for New Year’s resolutions. The most common personal resolutions are focused on being healthier – exercising more, eating better, improving fitness, losing weight, stopping smoking. People join the gym, sign up for weight loss programmes, and download meditation apps.

But what about your conversational AI project? Does it need a New Year’s resolution to be healthier in the new year, too?

If your organisation already has a conversational AI project, then you don’t need me to wax on about the importance of digital customer support. You get it. However, if you’re concerned that your current conversational AI tool isn’t up to the task of improving your digital support experience in 2022, then it’s time to make a resolution for change.

Even the best laid plans sometimes take a wrong turn or need to be tweaked as customer expectations and your organisation change. The start of a new year is the perfect time to take a step back and re-evaluate your conversational AI project and strategy. If this review leaves you dissatisfied with what you find, you’re not alone. Here are some common reasons other organisations have given for being unhappy with their conversational AI projects:

  • I can’t expand my solution to support my growing business and customer base.
  • I have limited integration options to create a seamless and personalised experience.
  • I started my project with an inexperienced start-up that isn’t able to provide the technology updates and support I need from my conversational AI vendor.
  • I am struggling to manage multiple chatbots across different business divisions or departments.
  • I am unable to staff my chatbot project with internal resources with the necessary knowledge and experience.
  • I don’t own the user interface or training data with my current chatbot provider.

The good news is that none of these common issues are dealbreakers that mean you must scrap your current virtual agent or chatbot project and start over. Like any New Year’s resolution to be healthier, you just need a plan that starts where you are and takes you to your goal of creating a successful, valuable, and healthy solution.

Your first step should be to download the new ebook Conversational AI Issues & Solutions: Transforming Ineffective Chatbot & Virtual Agent Projects. It takes an individual look at each of the common issues listed above, explaining how they can negatively impact your conversational AI project and exploring ways they can be solved.

When you’re ready to work out the details of your plan for a healthier chatbot or virtual agent and put it into action, the Creative Virtual team is ready to be your personal trainer and coach. Contact the team here to learn more about the expert consultation and technology that’s helping brands around the world deliver reliable and valuable conversational AI solutions.

This new year, resolve to transform your conversational AI project into a healthier, more effective customer service solution. Make 2022 the year your customers, employees, and company experience the full benefits of a successful chatbot or virtual agent.

Analysing Customer Queries to Improve Customer Service

By Maria Ward, Account Manager & Knowledgebase Engineer

I’ve worked in the chatbot field for over 15 years helping companies deliver better customer service, experiencing the technology both as a client and a provider. One of my favourite parts of my job is being able to use all my years of experience to steer my clients through the minefield of options to deliver a conversational AI solution that is both effective and efficient.

One of the reasons Creative Virtual has so many long-term customers is that we really get to know them and their products and services. This approach allows me to collaborate closely with my clients to identify opportunities that will specifically help them improve their chatbot experience.

Earlier this year I was working with one of my clients on expanding their customer-facing chatbot to provide self-service on more topics. One area we agreed would provide significant benefit to customers was addressing error codes they might encounter. However, due to the wide range of different machines and models customers might be using, the list of possible error messages they could see was very, very long!

Given the knowledge I had about their business and my experience with developing chatbot content, I knew right away that attempting to address every single error message would be wasted effort. I quickly steered them away from that pitfall and instead suggested our first step should be to go directly to the source: their own customer queries.

I analysed the previous year’s data collected by their chatbot to look for trends. What were the most common error codes users were asking about during that period? The goal of this analysis was to better understand which error messages customers were actually needing help with and how they were asking about them.

Once I identified the trends, I liaised with the client to select the error messages that should be added to the chatbot first. We looked at which ones customers were most frequently asking the chatbot about and which ones the client knew were likely to be the most common. We also looked at which errors could be explained and resolved best with a self-service approach. The error messages on this list were the ones that would deliver the biggest impact on their customer service.

Once we narrowed down the list of error messages, the next step was to identify whether each needed a new answer added to the chatbot or if there was existing content that was relevant. I also looked at how conversation flows could be used to guide customers to the specific information they needed to deal with their error code.

In the world of customer service, surveys are common tools for getting customer insights. However, you should never underestimate the value in analysing customer queries. The data collected by a self-service tool like a chatbot provides an honest, unfiltered look into your customers’ needs and what they are really asking.

My tip this Customer Service Week for improving your organisation’s customer service is to analyse your customer queries. Whether that’s transcripts from your chatbot, live chat, or contact centre, you can gain priceless insights that allow you to align your customer service updates with the actual needs of your customers.

Conversational AI Doesn’t Have to Be a Risky Investment: Step 2

By Mandy Reed, Global Head of Marketing

Innovative, cutting-edge, ground-breaking – these are all words used regularly to describe conversational AI technologies. Being the organisation that deploys an innovative technology typically requires being comfortable with a high level of risk. However, most companies don’t have the financial flexibility or company culture to take that degree of risk, whether real or inferred.

Deploying conversational AI solutions like chatbots and virtual agents can be risky but doesn’t have to be. You don’t need to be an early adopter of innovations to benefit from the technology. These solutions have been used by businesses for over two decades as part of their customer engagement and employee experience strategies, and you can take advantage of those learnings to deploy reliable, successful projects.

In this three-part blog series, I’m sharing three steps for achieving conversational AI success while minimising the risk for your organisation. Last time, we delved into Step 1: Be selective when deciding on a vendor and technology. If you missed that post, I recommend you read it first before moving on to the second step:

Step 2: Build a business case with realistic goals.

Embarking on any business project without identifying the goal is always a risk, so it is essential that you have a realistic business case and clear objectives for your conversational AI project. An experienced vendor will be able to assist you with this process by performing a textual analysis of your existing data, such as live chat or contact centre transcripts, to identify what queries can and should be automated with conversational AI.

Starting with this analysis immediately reduces risk because your business case is being built around your own data. It’s combining the vendor’s expertise directly with the information that is unique to your customers, employees, and company. Instead of guessing your users’ self-service needs or taking a generic approach, your business case is customised to you and your pain points from the start.

Follow that initial analysis with a consultation workshop to review the results and collaborate with the vendor to identify your key performance indicators (KPIs) and set realistic goals. These business objectives will directly inform how your chatbot or virtual agent is built and implemented. Having clear goals and deciding how you will track progress and measure outcomes minimises the danger of investing in a project that won’t really meet your needs.

The key in this step is to build your conversational AI business case around realistic and obtainable goals. Being practical about what you are automating and setting sensible targets for your solution creates a solid foundation for your project. It keeps your investment focused on reliable, reproducible outcomes and business benefits.

In the third and final instalment of this series, we will talk about starting your conversational AI project with a pilot and the best approach to minimise risk while rolling out a full deployment. A great resource for better understanding the financial investment needed for a successful virtual agent or chatbot is the Guide to Enterprise Conversational AI Pricing: Calculating the Cost of a Successful Chatbot or Virtual Agent. Even if your company isn’t at the enterprise-level, this guide provides valuable insights into budgeting and calculating ROI that’s useful for all organisations.

Conversational AI Doesn’t Have to Be a Risky Investment: Step 1

By Mandy Reed, Global Head of Marketing

In the technology industry there tends to be a focus on being innovative, cutting-edge, and ground-breaking. Industry awards, conferences, and articles frequently showcase and reward vendors for technological innovations. Analysts and expert speakers regularly highlight case studies of companies that are early adopters, deploying technologies in inventive ways, or finding success by taking a chance on something new and unproven.

Innovation is essential to the advancement of technology but doesn’t automatically equal practical business benefits. Having companies try out new technological developments and deploy existing solutions in creative and unfamiliar ways is important for finding practical applications for new innovations. However, being the organisation that deploys an innovative technology typically requires being comfortable with a high level of risk.

Most companies don’t have the financial flexibility or company culture to take that degree of risk, whether real or inferred. For them, proven and reliable results are more important than being innovative and flashy. Projects that get budget approval and management backing are ones that are considered safe bets because they utilise established technologies that have documented business benefits.

Conversational AI is one technology that is regularly described with words like ‘innovative’ and ‘cutting-edge’. Simply having ‘AI’ in the name makes some people think of it as being futuristic or only for companies with the resources to implement it for the cool factor. It can be easy for business leaders to associate conversational AI with being a high-risk investment.

Deploying conversational AI solutions like chatbots and virtual agents can be risky but doesn’t have to be. Your organisation doesn’t need to be an early adopter of new innovations to benefit from this technology. Chatbot and virtual agent technology has been used by businesses for over two decades as part of their customer experience and employee engagement strategies, and you can take advantage of those learnings to leverage conversational AI within your organisation.

Over the course of this three-part blog series, I’ll outline three steps for minimising risk and maximising benefits of conversational AI projects. Let’s get started with the first and most important step:

Step 1: Be selective when deciding on a vendor and technology.

The conversational AI market is oversaturated with new, inexperienced start-ups and technologies that haven’t been well-tested in the real world. The first step to reducing your risk is to choose a vendor that is established in the industry and provides a technology that has proven results. Both criteria are important when it comes to risk level.

Vendor experience is critically important because the more knowledge your selected provider brings to your project, the more confident you can be in their advice and guidance. You want a vendor that will become an extension of your own team and knows what they are doing because they’ve done it all before. Working with experts means you benefit from their many years of experience, thereby making your investment less risky even if your company is new to this type of technology.

When evaluating a vendor’s experience, ask specifically about how many years the company has provided conversational AI technologies, as these solutions may be an offering added recently even though the company has been in business for decades. Also ask about the experience of their individual team members and staff turnover rates. If they have high staff turnover and are constantly training replacements for departing employees, then you will likely miss out on the risk-reducing personal expertise you want the vendor to bring to the collaboration.

Just as critical as the vendor experience is having proof of their technology delivering positive results in real world applications. Don’t assume that just because a provider isn’t a brand-new start-up that they have a well-performing conversational AI technology. If the company has been in business for four or five years and only has one customer, you should question why more companies aren’t using their technology and if working with them is a risky option.

To reduce risk, ask about how the vendor has deployed their technology within your industry and what documented business benefits those solutions are providing. Saying they have the ability to deploy important features and functionality is great, but you want to see the technology in action in live installs. Also ask them about the length of their customer relationships as long-term engagements indicate that existing customers are happy with the technology, their results, and the collaboration. The vendor should be able to provide you with customer references so you can get first-hand feedback on their conversational AI projects.

Keep in mind that even if your company is minimising risk by selecting a proven solution with reliable results, you still want to partner with a vendor that is consistently innovating. You may not be the organisation trying out those new innovations first, but you don’t want to invest in a solution that’s not going to improve as those advancements become well-tested and are shown to deliver business benefits.

In my next post, we’ll explore building a realistic business case as part of Step 2 for reducing risk. In the meantime, check out the Guide to Selecting a Virtual Agent or Chatbot Vendor: Forget the Technology & Focus on Experience. It explains in more detail the most important questions to ask a vendor about their experience during your procurement process.

Conversational AI Data and Powerful Customer Insights

By Mandy Reed, Global Head of Marketing

These days it can be easy to feel like we’re in a constant state of information overload. The 24-hour news cycle combined with social media and the ability to Google anything always at our fingertips makes it difficult to escape the relentless barrage of information. This can also make it difficult to discern what information is genuinely important and worthy of our time.

The same is true when it comes to your organisation’s customer experience (CX). There is information floating around everywhere on customer behaviours, preferences, and sentiments. However, it can be a struggle to gather and identify the customer insights that are meaningful and most important for your CX strategy.

In fact, Gartner analyst Augie Ray identified ‘looking in the wrong places for customer insight’ as one of the key mistakes that kill CX programmes in their infancy. He points out that you can’t expect customer-centric results from a strategy that is not based on real customer insight and understanding. Likewise, you won’t be able to measure the impact of your programme if you aren’t gathering and analysing customer feedback, sentiment, and experiences.

Conversational AI is growing in popularity among businesses and their customers because of the positive self-service experience high-quality virtual agents and chatbots are delivering. However, organisations should never overlook the added benefits they gain by analysing the conversational data these tools can collect.

In a chatbot vendor selection guide, conversational AI expert Ridhi Mathur explains:

“In this era of data, a sophisticated virtual agent can play a very significant role for many businesses. The conversational data collected is very rich. It can be analysed and mined to understand your customers’ views, identify developing trends and make informed predictions, which can in turn contribute to areas such as product or service innovations and marketing.”

As you are developing your CX programme and selecting tools, be sure to keep this in mind. You want to implement a conversational AI solution that is not only built based on a deep understanding of your customers but will enable you to gather valuable customer data as well. You also want to ensure that you are properly analysing that data to assist with ongoing improvements to your CX and inform other areas of your business, such as product development and marketing programmes.

The customer insights that can be provided by chatbots and virtual agents are too beneficial to overlook when you are selecting a technology to implement. Consider this as you are reviewing conversational AI pricing options. Purchasing a chatbot without the right tools for data collection and reporting functionality never pays off, no matter how much money you might think you are saving. It is also key to work with a provider that has the expertise to assist you with mining and analysing the conversational data to maximise your benefit.

If you need some help cutting through the information overload when it comes to selecting a conversational AI vendor, check out the ISG Provider Lens™ Intelligent Automation – Solutions & Services. This report provides an independent, market-driven evaluation of the industry and comparison of 19 conversational AI providers.

Combining Chatbots and Voice for Omnichannel Experiences

By Liam Ryan, Sales Director

Last week Creative Virtual joined our partner Spitch as co-sponsors of The European Chatbot & Conversational AI Summit. The event was 100% virtual this year and featured two afternoons of various presentations, workshops, and panel debates. While being virtual is never quite the same as talking with someone face-to-face, it was great to see so many thought-provoking exchanges on the Discussion board and interesting questions asked by attendees during the sessions.

I teamed up with Gary Williams from Spitch to present on Day 1 of the Summit. Our session, The Omnichannel Solution: Chatbots + Voice, explored combining natural language chatbots with speech recognition capabilities to create powerful voicebot solutions. We shared some industry research from ContactBabel that showed 85% of CX professionals identified creating omnichannel/connected journeys as very or somewhat important to their strategies in the next two years (download the full report here). It’s no secret that customers want and expect an omnichannel support experience.

The tight integration of chatbots and voice creates a seamless journey as users switch between channels to help you deliver that connected experience. Gary and I shared two example voicebot use cases, one for customer support and one for employee support, that showed how the user could start a process on one channel and complete it on another in a smooth, seamless way.

For those interested in getting started with their own voicebot project, we ended our session with three important tips:

  1. Work with experienced vendors – Today’s market is crowded with new start-ups and inexperienced providers. You want to work with vendors that already have proven experience with both deploying and maintaining these solutions in your industry or sector. By partnering with experts, you immediately benefit from their experience. They can help you avoid common pitfalls, guide you on best practices, and ensure compliance with industry requirements and regulations.
  2. Select reliable technologies – This is why Creative Virtual and Spitch have partnered on voicebot solutions. We both bring years of expertise and documented results for each of our respective technologies. Be cautious about vendors that have attempted to tack on their own poorly developed chatbot or voice technologies to their main solution just so they can shout ‘Me too!’ You want technologies that are secure, can scale to current and future requirements, offer the hosting options you need, and will give you reliable results.
  3. Understand the integration options – When it comes to integrations, you need to first make sure there is a deep, seamless integration of the chatbot and speech technology to have a successful voicebot. Then consider what other integrations you are going to need to create a personalised user experience. You want a solution that can easily integrate with any existing content sources, backend systems, CRMs, other communication channels, etc. so that you can create a custom experience and connected journey.

If you’re interested in learning more about voicebots, schedule a demo to see these solutions in action and discuss possible use cases. You can also read more about the Creative Virtual and Spitch collaboration in our integration overview.

Thanks to Gary and Spitch for their event partnership and to The European Chatbot & Conversational AI Summit organisers and attendees for two days of great virtual content!

How Much Does it Cost to Make a Chatbot that Actually Works?

By Paulo Barrett, Chief Operating Officer

Ask any seller of a highly complex and customizable chatbot or virtual agent system about cost and you’re likely to get an evasive answer. ‘There’s no one-size fits all.’ ‘I’d need to talk to you on the phone to give you an accurate quote.’ Increasingly, in this ever-saturating market, it’s easy to find elements of chatbot pricing (i.e., API request fees) or flat monthly subscription costs for low-end systems, but who is giving the educated bot buyer a clear, top to bottom view of what it costs to build a system that will really work?

By ‘really work,’ I mean one that will materially contribute to cost savings, improve customer satisfaction, and maybe even generate new revenue. In other words, how much is it realistically going to cost to build a bot your customers will actually want to use.

The truth is, building a successful chatbot is not purely a question of technology. Whether you are buying a platform to BYOB, getting something cheap and cheerful off the shelf, or looking for a bot consultancy to support your internal efforts, your work is really just beginning once you have the system configured and deployed. The ongoing work to improve the chatbot’s performance and to get the best out of self-service in your unique deployment is what makes the difference.

It can be difficult to predict exactly what actions your customers will want to take in the beginning. That means being able to take an informational system and swiftly evolve it as desired customer outcomes become clear is key for success. This is enhanced by using great technology, of course, but ultimately, you need the right experts (internal or external) to separate your deployment from the crowd of others (often failures) which are flooding the support world.

Think of it like buying an instrument. No matter how expensive or special it is, either you learn to play it, or you get someone else to play it for you (alternatively, it ends up gathering dust and doing absolutely nothing for anyone). One way or another, this expert training costs time and money. You have to weigh this investment against the return.

Now to the million-dollar question … pun intended. What will a chatbot that your customers actually want to use cost for a large enterprise? While it’s true that most deployments are unique to every client and require some customization, there are some standard pricing building blocks you can expect to see.

The first cost to nail down is the pilot fee. How much am I gonna spend to test this thing out and see if it works for me? The financial risk associated with a pilot should be shared by the customer and the vendor and typically runs around $50,000 (USD). While the client assumes some risk via the initial cost, you should expect this to be credited against the cost of the full production-level deployment if you choose to move forward. This fee will cover all hosting, software deployment, content development, technical consultancy, and transactional fees for the agreed period. Typical pilots run 30-60 days from go live to give you enough time to see material results and make a decision about the ongoing plan.

Once you convert from pilot to full system (we pride ourselves on a +90% conversion rate!), you have some choices to make about how you pay for the tools and ongoing consultancy. Some customers wish to be purely pay-for-performance. Often, they go with a tiered model based on volume with session costs starting at a dollar (i.e., a single interaction with a user, with unlimited question/integration calls in that session). This per session cost may fall based on meeting certain volume thresholds. With any variable pricing model there are pros and risks for both the customer and the vendor.

If you prefer not to have a variable rate in your forecasting, you can purchase a more traditional software/services package. A standard production system should include integration with a live chat platform as well as your CRM. This will ensure your customers get personalized answers to their questions and are able to complete transactions with the bot online, rather than just receiving flat, informational content. The cost to provide the software and ongoing consultancy, along with an adequate knowledgebase of 100-200 solutions, will generally cost somewhere between $150,000 and $250,000 (USD) per annum, depending on the number of sessions.

While this may seem like quite the investment, you have to ask yourself: What is the cost of deploying a support tool that my customers don’t want to use and delivers a negative, frustrating experience?

Ready to learn more? Our team is always on hand to arrange a personalized demo with you and answer any questions you may have about getting started with your pilot.

Customer Service Week Musings: How does a machine know if it’s wrong?

By Laura Ludmany, Knowledgebase Engineer

There are many comparisons dealing with the main differences between humans and machines. One of the recurring points is while humans have consciousness and morals, machines can only know what they are programmed to, hence they are not able to distinguish right from wrong unless they are provided data to make decisions based on that information. There have been many discussions on the self-awareness of robots, which is a topic as old as Artificial Intelligence, starting from Isaac Asimov’s three laws of robotics, continuing to the Turing test and nowadays AI ethics organisations.

One thing is commonly agreed – bots need to be ‘taught’ morals, and to achieve this there could be two approaches, both having their advantages and disadvantages. The first one contains a loose set of rules, but plenty of space for flexibility; this system could always reply to questions. However, it could also result in many false positives cases and could go wrong on many levels. The other would mean more rules and a narrower approach. The system could answer a limited number of queries, however, with very few or non-false judgements.

What does this mean from the customer service and customer experience (CX) view and for virtual agents answering real time customer queries? If we narrow down our conditions, bots would deliver the right answers at most times. However, they could not recognise many simple questions, making users frustrated. The same can happen with the loose set of conditions: the assistant would easily deliver answers but could misinterpret inputs, resulting again in annoyance.

To solve this problem, we must use a hybrid approach: an AI tool can only be trained appropriately with real-life user inputs. While we can add our well-established set of rules based on previous data and set a vague network of conditions, the bot will learn day-by-day by discovering new ways of referring to the same products or queries through user interactions. Half of a virtual assistant’s strength is its database, containing these sets of rules. The other half lies within its analytics, which is an often-overlooked feature. What else could be better training for a CX tool than the customers leaving feedback at the very time an answer was delivered? Conversation surveys are not only important to measure the performance of the tool. They are also crucial for our virtual assistants to be able to learn what is wrong and what is right.

Our approach at Creative Virtual to reporting is to follow the trends of ever-changing user behaviour. We offer traditional surveys, which measure if a specific answer was classified as helpful or not by the user and if it saved a call. Sometimes, the specific required action or transaction cannot be performed through self-service options and the customer must make a call, or else, the answer has been slightly overlooked and needs to be updated – for these cases there is a designated comment section, so users can express themselves freely.

We all know from personal experience, that we can’t always be bothered to fill out long or detailed surveys – we are on the go and just want to find the information we were looking for without spending extra time to leave feedback. This is typical user behaviour, and for this we came up with different options for our clients such as star ratings and thumbs up and down, keeping the free text box, to make the rating simpler for users. The solutions deployed are always dependent on the requirements and preferences of our clients, which are in line with the nature of their business and their website design. For example, financial organisations usually go with the traditional options for their customer-facing self-service tools, but internal deployments often have more creative user feedback options.

What if, during a conversation, a virtual assistant delivered the correct answer to five questions, but two answers advised the user to call the customer contact centre and one answer was slightly outdated? Does this rate as an unsuccessful conversation, due to three unhelpful answers? To solve this dilemma, we have End of Conversation Surveys, which ask customers to rate the whole conversation, on a scale to 1-10 and choose what they would have had done without the virtual assistant. As always, there is a free text box for further communication from the customer to the organisation. These surveys show high satisfaction levels as they measure the overall success of the conversation, which can have some flaws (just as in human-to-human interactions), but still can be rated pleasant and helpful.

Let’s take a step further – how can the virtual assistant learn if it was right or wrong if none of these surveys have been taken up by the user? Is this valuable data lost? Our Creative (Virtual) analytics team have levelled up their game and came up with a solution! During voice interactions, such as incoming calls to customer contact centres, there is a straightforward way to understand if the conversation wasn’t successful, even if it wasn’t stated explicitly, as the tone might change or the same questions might be repeated. But how can we rate a written communication with our customer? There has been a specific platform developed, which sits on the top of our previously described survey layers. This platform classifies the whole conversation, with a carefully weighed several-factor-system, which can be tailored to our client’s needs, containing factors such as if there has been more than one transaction, whether the last customer input was recognised by the virtual assistant, if there have been negative user responses recorded, etc.. The primary ‘hard’ indicators remain the user-filled surveys, so this is just a nice icing on the cake, as our mature deployments show over 80% of successful conversation rates.

With our proactive approach and multi-layer analytics tool sets, we can be sure that our virtual assistants will learn more and more about what is right and wrong, to increase the customer satisfaction level continuously. However, I think no machine will ever be able to answer all questions correctly, as this would mean that deployments have stopped being fed up-to-date real-life data. Our world is changing rapidly as are our user queries. These cannot be fully predicted ahead, just analysed and reacted to appropriately. As long as AI tools serve customer queries, they will always face unknown questions, hence they will never stop learning and rewriting their existing set of rules.

As we celebrate Customer Service Week this year, we need to recognise the role customers play in helping to teach our AI-powered chatbots and virtual assistants right from wrong and the experts that know how to gather, analyse and incorporate that data to help train those tools. Check out our special buyer’s guide that explains why experience matters for using this hybrid approach to create reliable and always learning bots.