Chatbots and automated assistants can take on many use cases and be deployed into a number of different platforms to support your customers.
They can be integrated into any platforms such as website, Facebook Messenger, WhatsApp, Slack, Microsoft Teams.
The Rapid Response Virtual Agent program includes open source templates for companies to add coronavirus content to their own chatbots.
Artificial intelligence and machine learning are continuing to take a front-row seat in fighting COVID-19, with Google Cloud launching an AI chatbot on Wednesday. The chatbot, which it calls the Rapid Response Virtual Agent program, will provide information to battle the COVID-19 pandemic, as announced in a Google blog.
The program will Google Cloud customers to respond more quickly to questions from their own customers about the coronavirus. It’s designed for organizations who need to be able to provide information related to the COVID-19 pandemic to their customers, such as government agencies, healthcare and public health organizations, as well as travel, financial services and retail industries.
Google also offers Contact Center AI for 24/7 self-service support on COVID-19 questions via a chatbot or over the phone. Google also allows for businesses to add COVID-19 content to their own virtual agents with the ability to integrate open-source templates from organizations that have already launched similar initiatives. For instance, Verily partnered with Google Cloud to launch the Pathfinder virtual agent template for health systems and hospitals. It enables customers to create chat or voice bots that answer questions about COVID-19 symptoms and provide guidance from public health authorities such as the Centers for Disease Control and Prevention and World Health Organization (WHO), according to the Google blog.
The Contact Center AI’s Rapid Response Virtual Agent program is available in any of the 23 languages supported by Dialogflow.
We’ve been looking in more detail at this template and created our own chatbot. This is a work in progress and will be something which we are updating and improving daily. You can interact with this chatbot in the bottom right of this page.
We are often asked this question by clients looking to start their first chatbot project. Like most software projects the price really depends on the scale and complexity of the project. This will govern the effort involved in building a chatbot.
So, let’s take a look at the key drivers impacting the price for custom chatbot development.
The first one is the channel in which the bot should function. By this, I mean where the bot will be used. It could be as a website widget implemented as a popup interface or webpage (You can learn more about webpage chatbots here) or deployed into existing messaging platforms. Facebook Messenger, WhatsApp (read more about WhatsApp chatbots here), Microsoft Teams, Slack, Telegram, Viber are just a few examples with some providing more complex UI elements which can be utilised in your bot. Other obvious channels to deploy a chatbot are voice; Google Assistant, Alexa or even IVR systems can often use the same conversational engine as your text-based chatbots.
For our custom chatbot integrations, we normally provide one channel with the project and then charge per extra channels as required. We’ve found that our clients will often want a web-based chatbot first and role out the chatbot onto different channels. It really does depend on the type of project. Extra channel costs can be in the region of £1000-2500.
There are a number of really good chatbot building platforms for creating simple UI chatbot interfaces such as Botstar, Chatfuel and Manychat; these offer some free templates and the ability to create chatbots which can be deployed to specific channels, for example, Facebook Messenger (our Carly chatbot is a great example of a Chatfuel bot). These Bot building platforms offer some good features: Visual Flow Editors, APIs & Webhook Integrations, Templates, Broadcasting etc. However, in our experience, we have found that even with free templates there is always some costs due to setup and minor customisation to attain the maximum chatbot behaviour. Our minimum price is around £1500 with some monthly costs for hosting and chatbot management and training.
Depending on which channel the bot will work in there is also scope to provide other functionality such as voice capability for web chatbots or enhanced chatbot interface features. Again costs depend on the amount of complexity and effort involved in building each feature. As an example adding voice interaction capability to a Web chatbot would be £1000. Human-agent handover or WhatsApp integration is also a popular feature.
Natural Language Ability
If a chatbot is required to support more complex natural language processing(NLP- you can read more about some of the tech terms here) features and not just UI elements such as buttons then this will mean that additional effort is needed to train the bot and design and implement a more complex conversational flow. These sort of bots start at around £2500. We utilise the best of breed cloud NLP solutions for these types of projects. In particular, we use Google Dialogflow as we are Google Tech partners and experts in Dialogflow.
The complexity, scope and volume of the required conversational ability effects cost. This relates to specifics such as:
- The number of branches in the conversation tree.
- Quantity of questions that have to be handled by the chatbot which can often be in the 1000s.
- The numbers of training phrases needed; this can improve the accuracy of the chatbot, (you can read more about training data here)
- The complexity of conversational ability ie. support for complex user enquires.
So chatbots which combine richer elements with complex conversational ability can be in the region of £2500 – £5000
Chatbots are capable of supporting different languages, as long as these are supported by the NLP engine it’s possible to add different language permutations fairly easily. However, each language will need its own testing so costs can be in the region of £1500-3500 for each language depending on the size and complexity of the conversational ability.
Connect to existing systems, APIs, RPA, and knowledge bases
The other area which will impact cost is dependent on the planned role of the chatbot: what the chatbot will need to do carry out its role? Will the chatbot need to integrate with current systems to provide its responses? Will it need to hand over to live agents? Connect with CRM and ticketing solutions? Some chatbots may need to carry out complex interactions to provide answers to customer queries. Chatbots can leverage other AI systems to provide relevant information to govern conversational flow. For example, sentiment analysis.
With so many possibilities for chat bot features It’s hard to estimate the price here, but as a rule, these types of chatbots cost from £4000. For each integration, we would suggest a cost of £1000-2500 it really depends on the amount of development effort required for each one.
Often security demands for a chatbot project need specific features, for example, Hippa compliance. In these cases SSO, RBAC, and on-prem or private cloud deployment can be used to ensure compliance with company security policies. These can have an impact on overall project costs and again, costs are based on the demands of a specific project.
We offer our chatbot solutions based on a SAAS model. Costs incurred tend to be based on a yearly subscription and again depend a lot on the scale and complexity of the chatbot. These monthly costs will cover some of the following:
- Daily supervised learning and improvements.
- Monitoring conversations and confirming qualified intents as well as checking for unmatched intents and fixing them as needed.
- Post-development support.
- Third-party and integration maintenance. Making sure your bot is performing well and healthy!
- NLP costs(if applicable).
- Hosting and data storage.
- Chatbot reporting interface.
As a rule, maintenance subscription costs tend to be in the region of 10-15% of initial implementation costs per month.
As you can see, the cost of a chatbot project can vary widely depending on the features required. Each chatbot project is different.
It’s worth keeping in mind that the cost of a capable chatbot does not have to be prohibitive and it’s often a good idea to start small and add features as business needs require them.
At The Bot Forge, whatever the cost we propose for your chatbot project our aim is to create value for your business and hopefully form a long-lasting relationship.
The Bot Forge are always available to discuss further improvements and functionality to add to your chatbot or just to talk to us about your next great idea.
Training data for chatbots.
I’m going to look at the challenges in creating a chatbot which can answer questions about its specific domain effectively. In particular, I’m going to look at the challenges and possible solutions in creating a chatbot with a reasonable conversational ability at their initial implementation. Every chatbot project is different but often clients come to us with a large knowledge base which they want a chatbot to support from its release but with very little training data.
We are going to concentrate on a Dialogflow project to look at some examples however the challenges and solution are similar for all the most well know NLP engines, Watson, Rasa, Luis etc.
One of the key problems with modern chatbot generation is that they need large amounts of chatbot training data.
If you want your chatbot to understand a specific intention, you need to provide it with a large number of phrases that convey that intention. In a Dialogflow agent, these training phrases are called utterances and Dialogflow stipulate at least 10 training phrases to each intent.
Depending on the field of application for the chatbot, thousands of inquiries in a specific subject area can be required to make it ready for use with each one of these lines of enquiry needing multiple training phrases.
The training process of an ai powered chatbot means that chatbots learn from each new inquiry. The more requests a chatbot has processed, the better trained it is. The NLU(Natural Language Understanding) is continually improved, and the bot’s detection patterns are refined. Unfortunately, a large number of additional queries are necessary to optimize the bot, working towards the goal of reaching a recognition rate approaching 90-100% often means a long bedding in process of several months.
One of the main issues in today’s chatbots generation is that large amounts of training information are required to match the challenges described previously. You have to give it a large number of phrases that convey your purpose if you want your chatbot to understand a specific intention.
To date, these large training corpus had to be manually generated. This can be a time-consuming job with an associated increase in the cost of the project. One of the main issues we have faced is that often clients want to see quick results in a chatbot implementation. These types of chatbot projects are often use cases which are providing information regarding a wide-ranging domain and may not necessarily have a lot of chat transcripts or emails to work with to create the initial training model. In these cases there is often not enough training data and so it takes time to get decent and accurate match rates.
THE BOT FORGE PROVIDES
CHATBOT TRAINING DATA
The Bot Forge offers an artificial training data service to automate training phrase creation for your specific domain or chatbot use-case. Our process will automatically generate intent variation datasets that cover all of the different ways that users from different demographic groups might call the same intent which can be used as the base training for your chatbot.
Multi NLP platform support
Our training data is not restricted solely to Dialogflow agents, the output data can be formatted for the following agent types:
- rasa: Rasa JSON format
- luis: LUIS JSON format
- witai: Wit.ai JSON format
- watson: Watson JSON format
- lex: Lex JSON format
- dialogflow: Dialogflow JSON format
We provide training datasets in 100+ languages
We offer our synthetic training data creation services to our chatbot clients. However, if you already have your own chatbot project and just want to boost its conversational ability we can provide synthetic training data to meet your needs.
Testing the Solution
We wanted to test the effectiveness of using our synthetic training data in a Dialogflow chatbot agent by varying the number of utterances per intent using our own synthetic training data.
Dialogflow test agents
We carried out three different tests (A B and C) with 3 separate Dialogflow agents. Each agent had identical agent settings. The agents had 3 identical intents to provide information about the topic of angel investors:
In the first test (A) the chatbot was trained with 2 hand-tagged training phrases (utterances) per intent. Test (B) had 10 training phrases from our own synthetic training data per intent and test (C) had between 25 and 60 training phrases per intent.
We tested each agent with 12 separate questions similar to but distinct from the ones in the training sets.
We didn’t carry out any training during testing once the chatbots were created.
We recorded the % of queries matched to the correct intent, the incorrect intent or no match and also the intent detection confidence 0.0 (completely uncertain) to 1.0 (completely certain) from the agent response.
Overall test results
% correct match rate
|% incorrect match|
Average Intent Detection Confidence
|Test A (2x utterances)||50%||42%||8%||0.6437837225|
|Test B (10x utterances)||91%||9%||0%||0.7590197883|
|Test C (25-60x utterances)||100%||0%||0%||0.856748325|
Test A provided a 50% match rate. We observed a significant improvement in test B with the introduction of some of our synthetic training data to the agent. We were able to improve the match rate from 41% to 91% whilst TestC with 25-60 training phrases enabled a match rate of 100%. The average intent detection confidence also grew
In summary, chatbots need a decent amount of training data to provide accurate results. If there is not enough training data then a chatbots accuracy is affected and it can take some time to train it whilst being used to reach acceptable performance levels. At the same time, it can be costly and time-consuming to create training data for a chatbot needing to handle large numbers of intents.
Our synthetic training data creation service allows us to create big training sets with no effort thus reducing initial costs in chatbot creation and improving the usability of a chatbot from the initial release stages. If you only have a limited number of training phrases per intent and have large numbers of intents, our service is able to generate the rest of variants needed to go from really poor results to a chatbot with greater levels of accuracy in providing responses. We have carried out these tests with Dialogflow, but our conclusions are relevant for ML-based bot platforms in general. We can conclude that our Artificial Training Data service is able to drastically improve the results of chatbot platforms that are highly dependent on training data
Chatbot Training Never Ends!
I’ve looked at the benefits of using our training data at the early stages of a chatbot project. However, it’s important to note that the key to success, in the long run, is to constantly monitor your chatbot and continue training to get smarter. Either by doing constant training with human effort or by scheduling regular training cycles, incorporating new utterances and conversations from real users.
If you want to know more about our chatbot training data creation services get in touch