Hi Everyone,
I was trying to explore Appian integration with Gemini to summarize document, email and chatbot interface. If anyone did this either with or without plugin, let me know the details in brief how you achieved it. Also, did you faced any limitations?
(PS: Specifically needed to integrate with Gemini)
Discussion posts and replies are publicly visible
Hi Anvesh, do you have other model families available on the Google Vertex platform available to use, such as Anthropic Claude, or does you have to leverage a Gemini model? If it must be a Gemini model, why?Appian is currently evaluating direct support for more cloud vendors to power Appian's OOTB generative AI features, so this feedback would be helpful.
Not the OP, but been exploring the latest AI features recently and here are some thoughts.
Many organizations have pre-allocated cloud budgets on AWS, Google Cloud, etc., with credits tied to specific resources or LLM pricing models. Given this, why would they use Appian's OOTB generative AI features (which create vendor lock-in) instead of directly leveraging APIs like Gemini? Similarly, since Appian's storage cost an arm and a leg at large scales, many businesses prefer offloading documents to Google Cloud Storage or S3 and running LLMs directly against those files. This raises the question: what incentive do Appian customers have to store documents locally and lock themselves into Appian's generative AI ecosystem when they have allowances on other cloud platforms? Only reason I can think of is easy of use and setup, which business aren't very interested in.
We've found that many customers appreciate the ease of setup and rapid value that comes from a unified platform. If you've deployed AI features using gen AI you already know there are significant risks and complexities beyond simply calling the LLM API. With Appian we handle those at the platform level so that you can get right to the value and drive impact sooner. That's how Appian customers are outperforming the general market when it comes to delivering real value with AI.
We also recognize that some customers will have preferences for specific vendors or models. We want to deliver the rapid value of a unified AI approach while also providing choice and governance controls through the AI provider. It's wonderful to have community feedback to help drive the most impactful options!
When we go with Appian AI capabilities to handle manual, unorganized business processes, we can certainly propose what Appian’s out-of-the-box GenAI features can do. However, sometimes the business asks: our current employees perform this activity using xyz AI (e.g., Gemini); can we do the same with Appian? In such cases, we want to provide some promising assurance, even though we may not use it during the actual implementation, without positioning it as a platform imitation. Keeping all these things in mind, as we engage with different business teams and propose our use cases, we are exploring and implementing certain POCs to evaluate Appian’s capability to integrate with a few major AI models.