NLX

I used Conversations by NLX  once last year. The main use case for Conversations by NLX  when I used it was to build a chatbot conversation bot for a BFSI client.

I built a chatbot wherein users, when called or using chat-based tools, get authenticated and raise queries such as raising a dispute regarding their balances, paying their account balances, or checking the account balances. The chatbot is an intelligent bot that gives users answers to their account balances after verification and asks them to pay the account balances immediately. If the balance is between $50 to $100, it provides one payment installment and a 10% instant discount. For $100 to $500, it offers four payment installments and a 12% discount, and it scales up to $10,000.

Conversations by NLX was pretty easy to use. The GUI was straightforward and anyone can use it without any prior knowledge. I am happy with it.

The best features Conversations by NLX offers are the NLP engine and the bot reading. What I appreciate most about the NLP and bot reading features is the bot training capabilities. I cannot compare it with any other tool currently, but it is straightforward and very easy to use, so I would rate it five out of five.

Conversations by NLX positively impacts my organization by helping in gathering customer inputs and providing resolution for their day-to-day conversations based on disputes, checking account balances, or paying account balances, thus reducing the average handling time. The average handling time was earlier around seven to nine minutes and has been reduced to five minutes.

One thing I would share regarding Conversations by NLX is the need to refine the self-service flows and integration capabilities for multi-modal usage and leveraging the analytics for personalization.

Regarding Conversations by NLX's AI capabilities, it has Intent automation that can trigger autonomous API calls. It can guarantee zero hallucination, prevent mixing topics, and wrap the conversation in a structured way.

For governance, I know it has RBAC (role-based access) and a virtual workspace. For security, it is compliant with data protection guardrails and enterprise-ready hosting, so it has all the security compatibilities required nowadays.

The accuracy and reliability of Conversations by NLX's output has core capabilities such as voice and multimodal AI, generative script workflow, and global reach. The output of the AI capabilities includes intent detection.

Conversations by NLX's intent detection and global reach has actionable outputs that can generate context-aware, cited references and can highlight information from large document sets, such as manuals or technical documents, directly to users.

The areas of capability improvement for Conversations by NLX are agentic AI and task automation, multi-modal UI, and intent classifications. These areas of improvement would enhance the voice and technology, expand the platform's ability to handle complex multi-step native calls, and expand support beyond native languages to add more languages.

Conversations by NLX is stable.

Conversations by NLX's scalability is pretty much scalable. It has a lot of speed to scale and can be scaled based on the number of chats or conversations it can handle. It has a seamless integration plugin to enterprise architecture such as Lex, Bedrock, or Contact Lens without disrupting existing backend solutions. This demonstrates the scalability of Conversations by NLX and it is beneficial that it can scale on its own based on requirements.

I do not require customer support for Conversations by NLX currently, but if I do require it or find out about the customer support in the future, I will definitely provide feedback.

I previously used Unifor for some clients and the Google Dialog bot with integration with Genesis. The choice depends on client requirements. All solutions are unique in their own ways and the selection would depend entirely on what the client's requirements are.

My experience with pricing, setup cost, and licensing for Conversations by NLX is that there is no upfront cost or setup cost. It has a feature-based pricing structure that begins with a free tier and scales depending on usage, whether they are using standard chat or multi-modal conversation. Depending on the capability being used, such as voice-plus technology, low code, omnichannel, or multi-language support, this is the best part of using Conversations by NLX and that is impressive.

AWS  is used for deploying Conversations by NLX.

I have not seen a return on investment regarding Conversations by NLX yet. The client has not come back to me regarding the return on investment. However, it has definitely reduced the client's need to increase the headcount or the agent count after building the conversation AI tool or bot using Conversations by NLX analytics.

My experience with pricing, setup cost, and licensing for Conversations by NLX is that there is no upfront cost or setup cost. It has a feature-based pricing structure that begins with a free tier and scales depending on usage, whether they are using standard chat or multi-modal conversation. Depending on the capability being used, such as voice-plus technology, low code, omnichannel, or multi-language support, this is the best part of using Conversations by NLX and that is impressive.

I did not evaluate other options before choosing Conversations by NLX.

Conversations by NLX is designed for multimodal voice and digital self-service and can handle multilingual conversations. If they have similar usage, they can definitely consider Conversations by NLX because it can also leverage AI applications and you can bring your own LLM models to be used with Conversations by NLX. It can handle multilingual conversations in one frame of time and also has seamless channel deployment which can be tested in real time without wasting time or utilizing a native testing environment before going live. These are the tips and suggestions I would give to somebody considering using Conversations by NLX. I would rate this product a 7.5 out of 10.

I have not used Conversations by NLX  in the production environment, but I have used it for proof of concept projects for three different customers in the UAT environment for the last six months. The main use case is for conversational AI where the call comes to our call routing solution, and from there, it gets transferred to Conversations by NLX  to provide intelligent IVR treatment for customers. Conversations by NLX performs intent classification based on what the customer is saying, automatic speech recognition, and text-to-speech. It also conducts account lookup based on customer-provided information. A small amount of LLM or AI usage helps with better intent classification and FAQs. Once Conversations by NLX completes its processing, if it cannot solve the customer's issues, the call comes back to our call routing solution. Conversations by NLX also provides the most appropriate queue to which the call should be routed based on the current status of the account, and using that information, our call routing solution sends the call to that agent or queue.

What I like the most about Conversations by NLX is that the contact center industry, especially the legacy contact center industry, is very much accustomed to and comfortable using drag-and-drop blocks, connecting them, and writing the required business or application logic to serve business requirements. Conversations by NLX provides the same kind of web-based studio where you drag and drop blocks, connect them, and wherever required, you write business logic primarily using JavaScript. The studio is not only a place to develop your solution but also acts as a graphical representation of the overall solution. This is the first thing I appreciate about it, and it is what our customers appreciate about it, especially when compared with Amazon Lex , where they still have a graphical representation of things but not as comprehensive or anywhere close to Conversations by NLX.

The cost per call or per minute is relatively less. I cannot definitively compare it with the other two products that I have closely worked with, but the cost is relatively less compared to a couple of other competitors of Conversations by NLX that I worked with. The amount of coding required is less compared to other solutions, and while it still requires some amount of coding, it is relatively less when compared with other competitors. Integrating or building AI solutions or AI agents through Conversations by NLX is also easier compared to other competitors. These are the top four things that I appreciate or my customers appreciate about it.

I was not able to get approval to implement Conversations by NLX to more customers and business units because the cost is still high, especially when LLMs or AI get involved. That is one challenge regarding cost. Secondly, there are over half a dozen conversational AI providers who are doing exceptionally well in the US market as well as the Europe market. It is very competitive, so Conversations by NLX has to do a lot better, especially as I have difficulties promoting Conversations by NLX to some parts of European countries like Greece, Poland, and Austria. When it comes to European languages, they have to do a lot better. They are doing great with US English and Spanish, but compared with that part of Europe, it is a problem. Thirdly, it is hard for customers to open an account on the nlx.ai website and start building use cases. When I take Amazon or Google, it is relatively easy to create a contact center instance or conversational AI instance and start building use cases. Those kinds of things are not there in Conversations by NLX. I am also unable to reach NLX support to find answers for those things. Fourthly, since it has been acquired by Amazon, access to any new customers is completely blocked, and as of this month, it is simply not available since for the last several weeks. That is another big disadvantage.

I have been using Conversations by NLX for over six months.

Regarding stability, I would rate it ten out of ten.

For scalability, I would also rate it ten out of ten.

If you are a registered customer and need technical support regarding Conversations by NLX, it is easy. It is just that if someone is a random person trying to access nlx.ai to look around for training or access, technical support is not good. For registered customers, getting technical support is very easy. I would give a score of nine to the support on a scale from one to ten.

When I first started using Conversations by NLX, it was relatively easy for me to set up because I have worked on similar products such as Amazon Lex  and Amelia . However, for any new customer, it is still difficult without proper training. Not just Conversations by NLX, if the customer is totally new to the conversational AI world, it is not easy. Much more training and access to build things for a reasonable cost is required.

Compared to other conversational AI solutions, Conversations by NLX requires very little maintenance. Once you are done building it, things like automatic speech recognition and text-to-speech going down or any latency issues are extremely rare. With that said, maintenance is relatively very little. I would give this review a rating of nine out of ten overall.

I have been using Conversations by NLX  professionally for four years.

The best features Conversations by NLX  offers include flexibility and the ability to loop back to the main question. I appreciate that it asks, "Is there anything else I can do for you?" so I do not need to initiate the same conversation again and go through the same flow because, as a user, I do not know the flow. If it helps me go back to previous stages of the conversation, that helps significantly, whereas a one-directional path makes it really hard for users to navigate.

Conversations by NLX positively impacts my organization by using an LLM to help us develop this type of tooling. However, I feel that it is also pushing the problem forward. Implementation is cheap, but review and quality are expensive. Thus, it pushes the problem to a further stage of the standard development cycle.

Improvements for Conversations by NLX can be made by focusing on context. The more data you push, the more personalized and problem-solving it becomes. For instance, using the phone assistant starts with a zero stage, knowing only my phone number, but it should also save some data. When starting the conversation, I expect this assistant to already know things about me. Data protection rules exist, but I want to feed it as much information as possible so the conversation flows better. It should at least know my job, name, age, and perhaps previous history of conversations, so I do not have to explain the whole history repeatedly.

For the fintech, I have been working for six months, and for the healthcare, it was around four years.

In my experience, Conversations by NLX is not stable, and I have not used it.

I have not switched from a different solution before this; I am still using the old one.

Regarding outcomes from Conversations by NLX, the implementation is probably cheaper, around fifty to sixty percent for many companies using this type of practice. This allows us to bring new tooling to the market faster and experiment quicker. In one day, during a hackathon, we can do a lot and showcase it. However, while this implementation saves time, it has pushed the quality issue forward. Making it production-ready becomes harder because dumping a lot of code does not necessarily speed up the job. Many people can work on languages and systems they do not know, allowing them to learn and implement faster, but I would still be cautious about the end stage, which is quality.

I have not seen a return on investment with similar systems since I found that usually the application and user experience are better rather than saving time or money.

Before choosing Conversations by NLX or similar systems, I did evaluate options, including a German company.

My main use case for Conversations by NLX is building one, and I am also a user of the same application, which is called the phone assistant by Doctorlib.

I use Conversations by NLX to communicate with my doctor's practice by calling the practice, where an AI phone assistant picks it up and asks what I need. I explain my needs, and then it goes into one of the flows such as appointment booking or prescription, and I can automatically book an appointment on the phone without using anything else.

As a user, it feels as though Conversations by NLX is trying to be smart, but it is sometimes hard, and I might need human interaction for this type of thing, especially for healthcare. When it involves something personal, I would prefer to have a personal human touch. AI systems are getting better and more personal, but some users still do not prefer talking to an AI system at all.

My application was a phone one, but Conversations by NLX can also be adapted to a web application. Usually, it is a chat interface to talk through, but I think it needs a revamp of how we interact with Conversations by NLX systems. I prefer it to be all the way through a phone call, not a chat interface because I do not want to write; I just want to talk or visualize it rather than reading the whole text.

Concerning Conversations by NLX's AI capabilities, governance and security are significant topics. Data can be shared, but it all goes back to security. Securing the models and having guardrails is really important because input and security posture are crucial regarding the attack surface. The attack surface is extensive, as it only accepts text and can do anything with it. Therefore, you need to control what you are feeding it and limit its access and manipulations.

My advice for others looking into using Conversations by NLX is simple: try it. It is not complicated. I would rate this review nine out of ten.

We used Conversations by NLX  during one of our POC initiations to automate our tier-one customer support, handling everything from basic troubleshooting to order status inquiries. It integrated smoothly with our backend systems, so the automated flows have access to real-time data.

A great example of how Conversations by NLX  automation helped our Taiwan customer support was during a major regional product promotion in Taiwan last quarter. Our local support queue usually gets completely slammed with routine inquiries about shipping timelines and promotional code glitches. Before Conversations by NLX, our small, Taipei-based team would get buried under a mountain of identical tickets. We deployed Conversations by NLX with a localized automated flow in Traditional Chinese. It handled 65% of those incoming tier-one inquiries right out of the gate, verified the customer IDs, and had the real-time tracking data from our logistics backend.

We noticed a drop in average wait times for customers, as the automated flows handle a large portion of routine questions immediately. It also relieved a lot of pressure on our support team, allowing them to dedicate more bandwidth to complex, nuanced issues. Overall, it has really helped us scale our operations efficiently.

Conversations by NLX fits seamlessly into our existing CRM . When the bot is not able to resolve an issue, it performs a warm handoff to a live agent, passing the full context of the conversation so that the customer never has to repeat themselves. Combined with their analytics dashboard, we are constantly refining our flows based on real-time data, making our workflow more effective every single day.

The ability to handle complex and multimodal conversations seamlessly is one of the key things that Conversations by NLX offers. The interaction feels natural, whether it is through voice, text, or visual elements. It has a good analytics dashboard that gives real-time insights into customer behavior and helps pinpoint exactly where people might be stuck in the flow.

The analytics dashboard of Conversations by NLX is the true daily driver. While the multimodal conversation handles what makes the magic happen for the customer, the dashboard is where we do the heavy lifting to see how well it is actually performing. It also allows us to spot trends, see exactly where users might be dropping off, and make real-time adjustments to improve the flow. This is essential for keeping our support as effective as possible.

One area for improvement in Conversations by NLX would be the initial learning curve for setting up some of the more advanced integrations with our older legacy systems. We occasionally encounter small hiccups where the bot has trouble maintaining context across a very long, multi-turn conversation. Having even more robust, out-of-the-box reporting features would also be a fantastic addition for deeper analysis.

More comprehensive documentation for Conversations by NLX with more detailed code examples and a more centralized developer hub would make the initial development and debugging process much smoother.

The main things holding Conversations by NLX back from a higher score of 9 or 10 are the hurdles we faced during the initial setup with our old legacy systems and the need for more in-depth documentation to support developers. Those context gaps during longer conversations also mean it is not quite at a perfect level yet.

I have been using Conversations by NLX for about a year and a half.

Conversations by NLX has been very stable overall. We rarely experience any downtime that impacts our users, and it handles traffic spikes quite well. The only minor hiccups we have seen are those occasional context retention issues.

Conversations by NLX handles scalability exceptionally well, which was one of the main reasons we chose it. Since it runs on public cloud infrastructure, it automatically adjusts to handle traffic spikes without any noticeable lag, ensuring a smooth experience for our users, no matter how many people are interacting with it at once. It also allows us to easily deploy new conversational flows.

The accuracy of Conversations by NLX is quite strong, particularly when it comes to understanding customer intent and providing relevant information. It is highly reliable for handling those repetitive tasks without drifting off-topic. However, there are occasional moments where it might struggle slightly to maintain context through a very extended, complex discussion, which is why we do not consider its reliability to be absolutely perfect, and sometimes it hallucinates.

From a security and governance perspective, Conversations by NLX holds up really well for enterprise use. Its compliance with major industry standards gives us peace of mind, and the management features allow for clear role-based access control.

We have generally seen wait times with Conversations by NLX cut by at least 17% to 20%. Our support team is handling roughly half as many routine tier-one tickets as they were before it. It really freed up a significant amount of their day.

I am not certain about the exact metrics regarding return on investment with Conversations by NLX because I am a developer and do not handle the return on investment or the payment levels.

Conversations by NLX is doing a great job and having a decent amount for the subscription. I think it is good. You can also have multiple offers, but I think it is a good and reasonable way.

Before deciding on Conversations by NLX, we did evaluate a few other options. We used Google Dialogflow  and Amazon Lex  to see how they compared in terms of multimodal capabilities and integration with our cloud infrastructure.

We were a startup company and wanted to try out a few tools that are available before switching to Conversations by NLX to get those advanced features, especially to handle more complex, context-aware interactions and to streamline our agent handoff process. We have not productionized it completely, but it is still in the POC level. We are seeing a good performance in Conversations by NLX. I have covered most of the features of Conversations by NLX and everything regarding other improvements Conversations by NLX needs. My overall review rating for Conversations by NLX is 8 out of 10.

I used Conversations by NLX  just once as my main use case. I wanted to learn how to use Conversations by NLX  for conversation AI, such as practicing how to talk to someone. I practiced simulated conversations with Conversations by NLX, including talking to my boss and an angry peer, and I discussed these scenarios.

It stood out to me that Conversations by NLX can mimic human-like conversations. Even though I only used Conversations by NLX once, the language positively impacted my organization.

I was better prepared for conversations with my boss and peers. From now on, I am better prepared for questions that will be thrown at me by my boss and peers thanks to Conversations by NLX, and I am well-prepared now.

Conversations by NLX can mimic the tone and emotion, language, and the way a human talks, along with the necessary knowledge a human has for practicing conversations without actually doing it. This is more effective than a text conversation.

Conversations by NLX is very accurate.

I think improvements are needed for Conversations by NLX through the learning platform, even small ones.

I have been working for the last seven years using Conversations by NLX.

I have not experienced stability issues. I appreciate Conversations by NLX as it is very good and very helpful.

At the end, I received feedback from Conversations by NLX.

I think Conversations by NLX was deployed in my organization through a LinkedIn learning role-play or a Udemy role-play.

I would recommend Conversations by NLX to others looking into using it. I will get Conversations by NLX after the conversation, so I have additional thoughts about it before we wrap up. I gave this review a rating of ten out of ten.