Google Dialogflow has been used for building conversational assistants, mainly chatbots. I worked on building whiteboards as well, but primarily building conversational AI assistance for various use cases.
What is our primary use case?
What is most valuable?
The features I appreciate about Google Dialogflow include the flow editor, where users can drag and drop and build conversation assistance. The low code approach is a valuable feature of Google Dialogflow. The AI, supported by Google, does an excellent job of understanding customer intent. It is also integrated with Gen AI powered by Google Gemini. Additionally, the extensibility is notable as it can integrate with APIs and other systems, all within one single Google ecosystem.
Language support in Google Dialogflow is exceptional. One of the best features is the ability to build one flow and easily change the language. While direct translations in Google Dialogflow require customization within the platform, we can easily support multiple languages, which has been working effectively for us.
Prebuilt agents in Google Dialogflow have helped us experiment and test various capabilities. The travel prebuilt agent serves as a good example for demonstrating the platform's value quickly. These agents have been a starting point for building conversational assistants. For production deployments, most bots have expanded beyond prebuilt use cases, as they are standard templates, and production requires custom use cases. Prebuilt agents remain helpful in demonstrating capabilities and serving as a starting point for newcomers to the platform or particular industry.
Intent recognition in Google Dialogflow has been one of the early features and performs exceptionally. We have built our own set of tools around intent recognition. For example, in intent recognition, sample sentences are required for functionality. We have been building custom features as an add-on to intent recognition to determine exactly what sentences we need to add to improve the capability within Google Dialogflow.
We don't use sentiment analysis in Google Dialogflow because when examining use cases, sentiment analysis has not been a helpful feature.
Some of the out-of-the-box standard integrations in Google Dialogflow could be improved. When developing a bot that cannot answer a particular question, transferring to the contact center is recommended. The system should transfer the chat to a human agent who could address the customer's query. While manual integration is possible, Google Dialogflow does not offer out-of-the-box capabilities for these functions. Some competitors have these features available out of the box.
Entity extraction in Google Dialogflow is used in specific cases. For example, in flows where information collection is needed, collecting email addresses demonstrates where entity recognition would be utilized. Bots with that particular scenario utilize entity extraction.
What needs improvement?
The availability of Google Dialogflow needs improvement. It is currently available in markets such as Europe and India. However, some newer Google Cloud regions, such as the Middle East and KSA region, are still not available. Having Google Dialogflow available in those regions would allow countries with geo-specific requirements to deploy and manage their bots within one single region.
From a developer experience and user standpoint, they can build these quickly because these elements are defined and standardized within the platform. The end customer experience allows use cases to be up and running rapidly. Creating an engine, creating a simple flow, and making it available can be accomplished efficiently. Custom training is unnecessary, and intent training does not take hours. While creating a new intent involves a real training process manually, the platform handles most processes effectively in the back end.
For how long have I used the solution?
I have been working with Google Dialogflow for almost five years.
What do I think about the stability of the solution?
I have never experienced any issues with Google Dialogflow. It has maintained consistent stability.
What do I think about the scalability of the solution?
It is definitely scalable. Google handles all processes in the background, eliminating the need to manage those aspects.
How are customer service and support?
Tech support for Google Dialogflow is exceptional. They have been able to reach out and resolve minor issues whenever encountered. I have good access to their engineering team for any needs. We have closely collaborated with them for the last five years, and their support has been outstanding. All product documentation is comprehensive. They continuously update the product and documentation, making it an ongoing journey to learn new features.
I would rate the technical support for Google Dialogflow as a nine or ten. Their support has been exceptional compared to other vendors, specifically for Google Dialogflow.
How would you rate customer service and support?
Positive
Which solution did I use previously and why did I switch?
I have been working with Microsoft Bot Framework for quite some time. We have also experimented with platforms such as Code.ai and Cognigy, along with several other platforms.
How was the initial setup?
The initial setup process of Google Dialogflow is straightforward. It takes less than thirty seconds to set up the project and select the region for deployment. Users can define whether to use a prebuilt agent or build from scratch. Whether new to Google Dialogflow or experienced, starting from scratch or using a prebuilt agent is always possible as needed.
What's my experience with pricing, setup cost, and licensing?
Google Dialogflow operates on a per-session cost model. When deploying this solution to an end customer, additional costs must be considered, such as integration costs and hosting UI components. Having experienced this across multiple customers, understanding that pricing involves more than just the per-session cost is important for newcomers to Google Dialogflow.
It is a cost-effective solution. Bots are intended to deflect calls or chat in a contact center operation. For example, with a hundred agents in the contact center, deploying a bot solution would reduce twenty to thirty percent of the workload through simple automation, resulting in significant production cost savings. Google Dialogflow has demonstrated its ability to reduce operational costs for businesses. Regarding general AI capabilities, improvements can be made in overall cost prediction, as generic costs are based on tokens, knowledge size, and various influencing factors. Overall, it remains cost-effective by automating tasks traditionally performed by humans in regular operations.
Which other solutions did I evaluate?
Each platform has unique capabilities. Microsoft demonstrated this by being one of the few vendors to offer a code interface to build a bot five to six years ago. While Google Dialogflow maintained a low code interface, Microsoft now also offers this feature. Google has pioneered early technology and demonstrates understanding of contact center operations. Platforms such as Cognigy and Code.ai offer multiple options, including integration with Google Gemini or OpenAI and deployment to other clouds. These providers offer additional flexibility. Platforms such as Code.ai include out-of-the-box integrations with Salesforce and other leading platforms, providing integration capabilities not available standard with Google Dialogflow.
What other advice do I have?
In refund scenarios, when a bot responds with policy information denying a refund and offers limited options or transfer to an agent, the sentiment tends to be negative. In customer support scenarios, most situations involve negative sentiment, making it difficult to utilize sentiment analysis data for customization. This feature has not provided significant value.
Everything required from a standard chatbot platform is available. While exciting features are in the roadmap, all necessary table stakes are present within Google Dialogflow platform.
The main benefit of Google Dialogflow is its Google integration capabilities. When building an end-to-end Google solution with custom widgets and integrations, Google Dialogflow provides extensibility while maintaining a low code interface for bot management. New users can build simple bots while experienced users can create complex scenarios. It functions as an end-to-end platform, integrating with other Google products such as agent assist, Vertex AI, insights, BigQuery, offering benefits in reporting and end-to-end automation.
Organizations considering Google Dialogflow should examine its capabilities comprehensively. While all bot vendors offer similar basic features, it's important to consider beyond simple chatbot use cases. Future needs might include extending bot capabilities to agent assist, quality AI integration, and operational optimization reporting. Google Dialogflow's adaptability for various use cases makes it notable.
I rate Google Dialogflow an 8 out of 10.
