Tag Archive for: machine learning

ChatGPT, GPT-3, and Your Conversational AI Solution

By Chris Ezekiel, Founder & CEO

Since the official announcement in November 2022, there has been an enormous amount of buzz and excitement about OpenAI’s ChatGPT. Industry experts are publishing articles about it, social networks are filled with comments about it, and local, national, and global news organisations are reporting stories about it. From students using ChatGPT to complete assignments for class to me getting a little help from ChatGPT to write my latest ‘Virtual Viewpoint’ column, it certainly seems like everyone is testing it out.

As a specialist within the conversational AI space, Creative Virtual is excited about what ChatGPT and the technology behind it bring to our industry. We’ve been having lots of discussions with our customers and partners, as well as internally, about how this can deliver value to businesses using our V-Person™ solutions.

ChatGPT is an extremely powerful language model that is changing quickly and will continue to get more sophisticated. However, like any deep neural network, it is a black box which is hard – if not impossible – to control. Using it as a generative tool means you can’t steer in detail what it’s going to say.  You can’t deliver reliable, accurate self-service tools if you can never be certain what response might be given.

These limitations don’t mean you should write off ChatGPT or GPT-3 (and future versions) as completely ineffective in the realm of customer service and employee support. In some cases, one might be willing to accept a certain risk in exchange for very efficiently making large chunks of information available to a chatbot. Also there are ways to use the language power of GPT in a non-generative way, as we’ll explore in this post.

In any case, ChatGPT can only ever be used as just one piece of the puzzle, like content management, integration, user interface, and quality assurance. ChatGPT alone cannot replace all of that.

One of the design features of Creative Virtual’s conversational AI platform is the flexibility to integrate with other systems and technologies, including multiple AI engines such as transformer models like GPT-3. We are currently exploring the best way to interface with this model and use it to deliver value to our customers and partners.

Let’s take a closer look at ChatGPT, how it works, and the ways it can be used to deliver customer service and employee support.

 

What kind of AI is ChatGPT and how is that different from how V-Person works?

ChatGPT is a transformer model, a neural network, and is trained to predict text continuation. It uses a variation of GPT-3 which is OpenAI’s large language model (LLM) trained on a wide range of selected texts and codes. It is extremely powerful with respect to language understanding and common world knowledge. However its knowledge is not limitless and so on its own it will not have large parts of the information needed for specific chatbot use cases. Also its world knowledge is frozen at the time it was trained – currently it doesn’t know anything about events after 2021.

V-Person uses a hybrid approach to AI using machine learning, deep neural networks, and a rule-based approach to natural language processing (NLP). The machine learning component is integrated with workflow functionality within our V-Portal™ platform so enterprises can decide the best configuration for their conversational AI tool to improve in a controlled and reliable way. At the same time, natural language rules can be used as an ‘override’ to the machine learning part to ensure accuracy, resolve content clashes, and deliver very precise responses when needed.

We developed this approach to give our customers control over the AI to create accurate, reliable chatbot and virtual agent deployments. The use of natural language rules as a fallback option to fix occasional issues and finetune responses is much more efficient than trying to tweak training data.

 

Can businesses use ChatGPT to directly answer questions from customers and employees?

At the time of writing, ChatGPT is still in a research preview stage and highly unstable with no clean API available, so it’s not possible yet for businesses to use it in this way. However with its predecessor, InstructGPT, it is. It’s also worth noting that GPT-3 is high quality only in English and a few other languages which is another potential limitation for global use.

The biggest issue with using ChatGPT to directly answer questions from customers and employees is that it does not give you control over how it will respond. It could give factually incorrect answers, give answers that don’t align with your business, or respond to topics you’d prefer to avoid within your chatbot. This could easily create legal, ethical, or branding problems for your company.

 

What about simply using ChatGPT for intent matching?

There are two ways in which GPT-3 could be used for intent matching.

The first way just uses GPT-3 embeddings and trains a fairly simple neural network for the classification task on top of that. The second option also uses GPT-3 embeddings and a simple nearest neighbour search on top of that. We are currently exploring this last option and expect to get some quality gains from that approach.

 

Can I just provide a few documents and let ChatGPT answer questions by ‘looking’ at those?

Yes, this is absolutely possible. In fact, we have offered this functionality with V-Person for several years without needing GPT but none of our clients have been interested. GPT-3 improves the quality of this in most cases, but also comes with a higher risk of being very wrong. If an organisation is interested in using GPT-3 in this way, we can support it within our platform but what we currently offer already enables us to deliver document-based question answering.

It’s important to keep in mind that using ChatGPT to answer questions from documents is only addressing one aspect of the support expected from a virtual agent. For example, no transaction triggering API will ever be called by GPT looking at a document.

 

Is it possible to give GPT-3 a few chat transcripts as examples and let it work from them?

You can provide GPT-3 with sample transcripts and tell it to mimic that chat behaviour. But unless you want a chatbot with a very narrow scope, a few transcripts won’t be enough. If there are complex dialogue flows that need to be followed, you’ll need to provide at the very least one example of each possible path – most likely you’ll need more.

This raises some difficult questions. How do you maintain those if something changes? If you try to use only real agent transcripts, how do you ensure that you have complete coverage? How do you deal with personalised conversations and performing transactions that require backend integration? It may not be too difficult to train the model to say ‘I have cancelled that order for you’ at the right time, but that doesn’t mean GPT will have actually triggered the necessary action to cancel the order.

When you really examine this approach it becomes clear that this is not an efficient way to build and maintain an enterprise-level chatbot or virtual agent. It also doesn’t address the need to have integration with backend systems to perform specific tasks. Today our customers achieve the best ROI through these integrations and personalisation.

 

What other key limitations exist with using ChatGPT to deliver customer service or employee support?

Using a generative ChatGPT-only approach to your chatbot does not give you the opportunity to create a seamless, omnichannel experience. To do that, you need to be able to integrate with other systems and technologies, such as knowledge management platforms, ticketing systems, live chat solutions, contact centre platforms, voice systems, real-time information feeds, multiple intent engines, CRMS, and messaging platforms. These integrations are what enable a connected and personalised conversational AI implementation.

With ChatGPT there is no good way to create reliable and customised conversation flows. These flows are regularly used within sophisticated conversational AI tools to guide users step-by-step through very specific processes, such as setting up a bank account. This goes a step further than just creating a conversational engagement to employing slot-filling functionality, entity extraction, and secure integrations.

You also won’t have the ability to optimise the chatbot for the channels and devices on which it will be used. This includes using rich media – such as diagrams, images, videos, hyperlinks – within answers. For example, you can’t include an image carousel to display within a messenger platform. You won’t be able to show photos or drawings to help with a new product set-up. You don’t have the ability to display clickable buttons with options for the user.

 

As ChatGPT continues to change and moves out of the research preview stage, our expert team at Creative Virtual will stay on top of new developments and opportunities this technology offers. Our mission is always to innovate in a way that will help companies tackle their real challenges and deliver real business results – and our approach to this language model is no different.

If you’re interested in discussing more about how ChatGPT and V-Person might fit with your conversational AI strategy, get in touch with our expert team here.

Will AI be 2023’s Co-worker of the Year?

By Mandy Reed, Global Head of Marketing

It’s that time of year when business predictions from analysts, experts, and industry insiders start to make an appearance. Not surprisingly, artificial intelligence (AI) is featuring prevalently in predictions for 2023.

For example, the analysts at Forrester are predicting that AI will become an indispensable and trusted enterprise co-worker next year:

“Rapid progress in areas of fundamental AI research, novel applications of existing models, the adoption of AI governance and ethics frameworks and reporting, and many more developments will make AI an intrinsic part of what makes a successful enterprise.”

This prediction is not shocking or out of left field. A Harvard Business Review article published in 2017 referred to AI, particularly machine learning, as ‘the most important general-purpose technology of our era’. Today a majority of enterprises have already made significant investments in AI and are seeing positive results. These successes are laying the groundwork for further investment and expansion across industries, departments, and use cases.

I particularly like the use of the word ‘co-worker’ in this prediction. Just like any other employee, these AI applications will be a part of the team and require human collaboration to be successful. AI is not poised to take over the enterprise. Instead, it is being used to support the organisation’s goals and help its human colleagues perform their jobs better.

Any AI or machine learning tool can only be successful if it has the right human co-workers. Humans are needed to create the application, train the system, monitor the performance, and perform necessary maintenance. Humans are needed to identify which tasks should be automated with this technology and which are better performed by a real person. Humans are needed to make decisions about when the system should be able to ‘learn’ automatically and when it needs a human-in-the-loop to make that judgement.

In return, the human co-workers benefit from having mundane tasks and processes automated so they can focus on more complex work. Human contact centre agents benefit from easy access to information so they can focus on providing compassionate, emotionally intelligent engagements. Human employees have instant access to IT and HR support online so they can easily get help regardless of when or where they are working.

The advancements in AI over the past several years have contributed to a growing list of practical and beneficial use cases. Enterprises are seeing success with AI-backed customer service, employee training, customer onboarding, personalised sales, advertising, content generation, code writing, product performance tracking – the list goes on. And they are seeing success because of the humans involved with creating, optimising, and using these tools.

Are you making AI an indispensable part of your 2023 plans? Will AI become a trusted co-worker for members of your team in the coming year? As with any prediction, it will be interesting to see how this one plays out within organisations next year.

If you’re looking at adding conversational AI to your 2023 strategy, the team at Creative Virtual can help. Our V-Person™ technology puts you in control of the AI so you can better care for your human employees, contact centre agents, and customers with strategically designed automated support. Request your personalised demo with an expert member of our team to learn more.

[Meet the Team] Eileen Stall: Looking Past the Machine Learning Buzz & Chatbot Hype

By Mandy Reed, Marketing Manager (Global)

Creative Virtual is one of only four companies in London to be recognised by The Queen’s Awards for Enterprise in the category of Innovation in 2017. The company was selected for this honour not only for our innovative technology, but also because of the hard work and dedication of our global team. Creative Virtual’s experienced, expert team is what sets the organization apart from others in the industry. As part of the five-year celebration of our Queen’s Awards for Enterprise, we’re talking with some members of the Creative Virtual team about the technology, their involvement in the industry and what winning this award means to them.

Today we introduce Eileen Stall, Knowledgebase Manager with Creative Virtual USA. Having prior management and customer care experience and a degree in art history, Eileen came across a Creative Virtual job posting in 2015 while looking for a new role with a better work-life balance. Now a key member of our operations team, she is responsible for the building and upkeep of virtual agent knowledgebases for a number of large enterprise clients, including Cox, The New York Times and IHG. Eileen works closely with these organizations to enhance their content, utilize analytics to improve answer quality and coordinate new expansions of their virtual agents. She shared with me how the media buzz around machine learning has created industry challenges and why Creative Virtual’s diverse team is key to the company’s success.

What have you found to be major challenges facing the chatbot and virtual agent market today?

Chatbots have gone from being an obscure technology to one that everyone knows about and have likely had experience with. It’s nice that now chatbots have some recognition value when I explain my job! But “machine learning” becoming a buzz word has created a big challenge for the market. Companies don’t understand the huge amount of data that you actually need to implement a solution that only relies on machine learning. The idea that you can just deploy a chatbot and never have to work on it again is simply not possible. I find that a divorcing of technologies from their original contexts has created a lot of misplaced hype in the tech world.

Creative Virtual USANow the market is seeing a number of chatbot pilots which attempted to implement machine learning to inappropriately small datasets ending with some messy results. In several cases, we as Creative Virtual have had to go through a re-education process about the realities of machine learning after potential clients have heard some fantastical pitches from competitors. We’ve opted not to rely solely on machine learning in spite of the buzz. Rather, we leverage human curated content and look at user intents to determine the best responses. This may take a little longer than what’s promised with some other solutions, but that extra time is better than a Twitter bot disaster!

This is what sets us apart from those other vendors. We have a solid understanding of what we’re doing and why we’re doing it, not just making technology decisions based on the buzz. Even when the market started going crazy about deploying fully self-learning virtual agents, we stood firm in our combination of machine learning and human curation of content because we knew that was the best approach to accurate and reliable solutions.

Why is it important to work with an experienced chatbot and virtual agent vendor like Creative Virtual?

Creative Virtual has what I like to call a concierge approach. We offer a packaged deal of innovative software and years of accumulated experience in the industry. Clients can leverage our experience to create a successful virtual agent. Companies don’t have a “bot division” so the more we can provide a concierge type of experience, the better the result and the less overwhelmed they feel. How can you expect someone with no experience to build out the bot, to know what content to put in, to identify when a user needs to be escalated to live chat? That doesn’t make sense. We don’t burden our customers with requests to provide huge amounts of content or put pressure on their copywriter. We have the experience to guide and assist through the entire process. We can be ghostwriters, creating content in the brand’s voice, but tailored to the chatbot platform, such that all that is needed is the client sign-off.

We customize our virtual agents on a case-by-case basis but have developed expertise in many industries. For example, I’ve worked a lot with telecommunication companies so I’m very familiar with the type of questions their customers are asking most often. The 80/20 rule has definitely applied to deployments in this sector wherein nearly 80% of the inputs are asked about the top 20% of content. Deploying a slimmer knowledgebase with the plan to build out is a great approach for that situation. You start with the most common questions being asked repeatedly – questions about a bill, how to cancel or change service, etc. – and then build the other content organically based on actual user conversations and the customer’s voice. That way the knowledgebase is being driven by what customers are really asking about. The more unnecessary content you have in a chatbot knowledgebase, the harder it is to achieve a high level of accuracy. Of course, the virtual agent won’t be able to answer 100% of questions at first, so it’s important to make sure your secondary support system, such as live chat or a help page, is robust. That way you have a back-up for users and can still provide a good customer experience. This approach is successful when you’re working with an experienced vendor like Creative Virtual because we have the expertise to build out an effective knowledgebase quickly and efficiently.

How has the explosion of new contact channels affected the deployment of virtual agents?

A lot has changed in the world of virtual agents over the past few years. Now companies are coming into new deployments with a plan for their bot. They have a space in their business where they need to provide help and know they don’t need a human to do that. They also know enough about the technology to understand the benefits of working with an experienced vendor who can help them finetune and implement their plan to achieve the best possible results. They have confidence in our expertise and take advantage of our large library of existing integrations and knowledge about rolling out chatbots to additional channels.

Creative Virtual USAIt can be like the wild west with new technologies though, with people and companies suffering from tech hype. There’s a growing confidence in bots and so you don’t need to convince people to use them. However, before rolling out a chatbot to another channel, companies need to keep in mind the context of their customer engagement rather than making a decision simply on that the fact that something is new and creating a lot of buzz.

For example, enterprises might already have some stats and live chat transcripts from a channel like Facebook Messenger that they can use when considering adding their chatbot to that type of channel. Similarly, adding a bot to their app can be informed by what customers are already doing on mobile. When faced with a completely new channel where there is no history of experience with customers there, that’s when it’s hard to determine whether that channel can provide value. Voice assistants like Amazon’s Alexa and Google Home are getting a lot of media coverage now, and they can be a great channel for chatbot deployment for some use cases. This is where context is important to provide a positive customer experience. Does your content lend itself well to a voice conversation? Or do your customers need a visual, such as a diagram or chart, to really benefit from the virtual agent’s answer? That’s another reason to work with an experienced provider like Creative Virtual. We can help guide you through the maze of new contact channels to develop the best experience for your customers.

I love that there’s always something new in this industry, so my work never gets stale! There’s always something to learn and new ways to use these developments to help our clients provide better customer support.

What does Creative Virtual winning The Queen’s Awards for Enterprise: Innovation 2017 mean to you?

Creative Virtual USABeing American, I had never heard of the Queen’s Awards, so when Creative Virtual was announced as a winner I did a little research and was happy to discover that the award went beyond simply recognizing companies for their financial success. The award also recognized the people behind that success who bring together our variety of experiences every day to create and deliver our innovative solutions. This really resonated with me as I’ve always felt that the diversity of our team in the US is what makes us a good vendor and partner. Everyone is coming together from different backgrounds with different perspectives to become experts in the field.

This award is fitting because on top of delivering innovation to our customers, we also prioritize developing our staff. Cultivating positive interpersonal relationships in the office and across customer and partner companies is an important part of the Creative Virtual culture. Our diversity has been key to our success, particularly on difficult projects. We’re not your stereotypical bunch of old white men sitting in a boardroom; we’re a diverse group of people combining our unique outlooks and experiences to deliver the best solutions possible to our customers.

Cutting Through the AI Hype

By Liam Ryan, Sales Director

Last year when I attended the Social Robotics & AI conference I had the pleasure of speaking with Professor Noel Sharkey which brought back memories of being a contestant on Robot Wars. This month I attended the event, now called AI & Robotics THE MAIN EVENT, for a second year and again got to hear Professor Sharkey speak. This time I started reminiscing about my appearance on Techno Games where he was also a judge. I was also very aware that the robot my team entered into that competition back in 2002 seemed like a child’s school project when compared with the robots at this event!

This year’s AI & Robotics conference was held in London and aimed to investigate emerging technologies and their real world impacts on business and work. It cut through all the hype to focus on the reality of AI and automated technology in the corporate world. The event agenda featured a variety of academic and industry speakers sharing their expertise, including our Founder & CEO, Chris Ezekiel.

chatbots and your customersIn his presentation, Chatbots & Your Customers: A Realistic Look at Using AI & Machine Learning, Chris drew from his extensive experience working with companies around the world to give insights into the world of enterprise chatbots. He explored the current realities of machine learning and AI and the role they should play in the customer experience space. Customer service chatbots need to use a combination of machine learning and human curation of content in order for companies to ensure a predictable, reliable and accurate customer experience.

In addition to watching the live demonstrations Chris gave during his presentation, many attendees also stopped by our stand during the ‘AI in Action’ sessions to see more examples of chatbots and virtual agents we currently have deployed for our customers. While speaking with them, it became clear that many of them were actively looking to add this technology to their customer support strategies. Some mentioned that they had been looking into these types of tools for a couple of years and were now very aware that the time is right to push the project forward or risk missing the boat. These sentiments didn’t seem to be specific to any one industry either. I heard the same thing being echoed across the board.

In my opinion, this year’s event was just as good as last year, if not better – from the list of speakers to the lunchtime spread! Our thanks to the event organisers for inviting Creative Virtual to sponsor the event again.

Please take a look at our Products page to learn more about Creative Virtual’s suite of Smart Help solutions and then request a personalised demo to see how chatbots can help you build relationships with your customers.

I’m sure you’re itching to see my appearance on Techno Games, so check out the video below starting at about the 11-minute mark.

“Keeping it real with chatbots” – Impressions from the Chatbot Summit

By Björn Gülsdorff, Head of Business Development

“Your talk was the most down to earth”, an attendee wrote me via the event’s messaging app after I had delivered my keynote “A Hybrid Approach to Machine Learning” at the Chatbot Summit in Berlin.

Now, is that good or bad for a keynote? Shouldn’t I have been breath-taking or ground-breaking or both – instead of down to earth?

Gaging by the talks we had later at our booth, it was a compliment. Creators of chatbots, as well as potential buyers, were happy to see a way out of the current AI craze where unsupervised machine learning seems to be the one and only solution and the power of the human brain is overlooked – and which requires enormous amounts of groomed, relevant data.

Enterprise ChatbotsIn fact, the chatbot community struggles to add learning, conversation and personalisation to their systems, which are very often highly scripted things. I spoke to a chatbot builder who was forced to use algorithms to create training data. Puts a new spin on the “artificial” in Artificial Intelligence.

There was also a lot of uncertainty about how to move chatbots into the enterprise world. A world of timelines, budget, ROI and predictability. Again, what it takes is a robust, proven, down to earth approach to each project, combining the virtual (artificial) intelligence with the real and unlocking existing silos of information. Not everyone from the chatbot crowd was happy to hear that though.

Nonetheless, it was a fantastic, energising event! I was pleased to join this buzzing chatbot community (we had our first discussions with other attendees while setting up our booth; it wasn’t even 8am!) and to learn that the number of bots is currently growing faster than the number of apps grew when the age of the apps began (and it is in decline again already). It is even fair to say I was touched to see a technology lift off that Creative Virtual has fostered and helped grow for over a decade – in fact, some of our senior management team started this in the last millennium.