Hi,

We have been using Intercom for our customer support for several months now and we are exploring AI automation solutions. Initially, we built workflows to handle basic support tickets, but we quickly found this approach to be unsustainable, as it became burdensome to maintain and scale.

We also tried Intercom Fin, which we found to be an improvement over workflows. However, it is quite expensive, occasionally provides incorrect answers, and struggles with more complex tickets. With Fin charging $1 per resolution and our support agents resolving tickets at about $0.66 each, we decided to explore alternative solutions.

I was tasked with identifying a viable option, so I’ve reviewed nearly all the apps listed under the Automation category in the Intercom App Store.

There are approximately 140 apps in the Automation category, and we focused on identifying those that can automatically respond to customer messages. From this list, we found over 20 apps that either offer or claim to offer this capability. Below is my research summarization, with a focus on two key factors: affordability and functionality.

Among the 20+ apps, “Foundkit AI Assistant,” “Fedra GPT,” and “Zane” appear to be discontinued, as their websites are no longer accessible. Additionally, "Gus" is limited to Spanish in its console, so we excluded it from consideration as well.

Our boss is cheap, so we’re looking for affordable solutions that offer a free trial, allowing us to thoroughly evaluate them before making a full commitment. Some apps lack transparent pricing or require a credit card to proceed, which we find less convenient. We value clear and upfront pricing, as the absence of public pricing often indicates higher costs. As a result, we decided to skip the following apps.

Among the remaining apps, some rely on rule-based systems combined with predefined templates to answer user questions, while others use workflow-based solutions that require defining and managing workflows to set up automation. We find these approaches outdated, as managing workflows can be cumbersome, and they often fail to handle complex cases effectively. Additionally, users frequently follow the flow only to receive answers that don’t meet their expectations. As a result, we decided to eliminate these options after initial testing.

Finally, we identified four apps that meet our requirements. Among them, Quickchat, Aidan, and Airbot are chatbots capable of automatically responding to customer messages. Aissist.io, on the other hand, offers two distinct modes: auto-draft and auto-reply. This approach seems more practical, as it allows us to start with auto-draft and transition to auto-reply once the AI is fully prepared and optimized.

One thing worth noting is that most apps may have different user interfaces, but they generally function in the same way. You provide instructions (often referred to as AI guidelines, which are essentially prompts for GPT) along with a knowledge base (such as a website, PDF, etc.). When a user asks a question, the AI searches the knowledge base and combines the results with the provided instructions to generate an answer.

However, based on my findings, this approach falls short for more complex questions. Additionally, when the knowledge base is too large, the AI can become overwhelmed with excessive information, leading to confusion and inaccurate answers. Feeding the AI concise and relevant information is crucial for accuracy.

Among all the options I reviewed, only Aissist.io addresses this issue effectively. Beyond the standard instructions and knowledge base, it offers a smart tag feature, which allows for specific contextual instructions tailored to particular scenarios. This fine-grained approach significantly improves the accuracy of AI responses.

This is just my initial research. Does anyone have experience with these apps? If so, I’d appreciate it if you could share your thoughts, including the pros and cons of these solutions.