What is our primary use case?
I had a hackathon integrated with Capgemini, through which I gained access to Glean Platform and became familiar with the agents, architecture, and workflow, which was straightforward to understand. Since then, I use Glean daily for my queries and various solutions more than ChatGPT. It provides me with excellent assistance, deep analysis, and research answers that are unique to this platform and straightforward to build any agent that we require.
The hackathon project involved building a healthcare agent, but I am currently working on AI agent orchestration, which keeps my system in sync.
I have used all the major cloud providers including AWS, Azure, and GCP. I prefer to stick with AWS and GCP because they are thriving and the majority of AI implementations are choosing to be part of these cloud ecosystems. This is where AI expands across territories. There are already numerous implementations in these cloud providers. In AWS, I have used Bedrock, SageMaker, Lambda, and S3 buckets, which are already implemented and help when entering the world of AI. With GCP, I have been in a project and used the Agent Development Kit (ADK), which is very similar to Glean Platform and helps build AI with the agent kit. I have also used Vertex AI, which is impressive through their workflow.
How has it helped my organization?
Orchestration primarily decides which agent handles which part of a task and manages errors, retries, and complete state while maintaining the workflow effectively.
On the orchestration side, branching, looping, and human approvals support non-linear workflows in production environments with real-time access. Through refined and improved prompts, this helps in closing the loop with actual users.
One unique feature that stood out is that every step in the workflow can pause or recover from failures. Instead of finishing the entire agent in one go, this approach is interesting because if there is a minor change in a previously completed step regarding the state, I can re-edit and restart from the beginning. This is truly impressive.
What is most valuable?
I am currently working on an important HSBC use case, which primarily focuses on orchestration. Glean Platform helps in handling daily tasks with NLP and organized workflow, which branches out through the architecture. It fills up missing details, keeps track of context, automates the workflow, and assists with orchestration. Glean Platform places emphasis on security features, has full visibility and debugging support, integrates easily with APIs, and efficiently handles agents through orchestration.
What needs improvement?
Glean Platform can implement more models. The NLP creation could be a little faster, and while debugging is amazing, the routing also needs to provide higher quality outputs sometimes rather than extensive searching. It can also help by implementing permission-controlled access, so when logging in, it can check permissions rather than asking for API keys or raw credentials later.
For how long have I used the solution?
I have been using the solution for over two months.
What do I think about the stability of the solution?
The platform is truly amazing. I have never seen any breakdown or error with Glean Platform. It is always working and always on point.
What do I think about the scalability of the solution?
Glean Platform is scalable because it offers amazing features ranging from chatbot to building any agents. It helps with orchestration and any technical demonstrations that I need, providing end-to-end detail for everything.
How are customer service and support?
Customer support is amazing. It starts with AI, and if I have a query that requires connecting with customer support, it is always available and accessible.
Which solution did I use previously and why did I switch?
I have used many solutions, but Glean Platform is the best, and I would truly appreciate it if everyone considers this platform. Through its stimulating and straightforward approach, this is really impressive.
We had many other options such as DeepSeek, Claude, ChatGPT, Copilot, and Gemini model. However, this feature is very rare because Glean Platform is an overall 360-degree product that helps from A to Z, rather than addressing only specific areas.
How was the initial setup?
The pricing, setup cost, and licensing were all free for us because it was a hackathon and the company came down to our campus for the event. We were the winners of the challenge they provided.
What about the implementation team?
The pricing, setup cost, and licensing were all free for us because it was a hackathon and the company came down to our campus for the event. We were the winners of the challenge they provided.
What was our ROI?
I have already appreciated this platform and have given a demonstration of this project to my colleagues and friends. They are using it end-to-end daily.
What's my experience with pricing, setup cost, and licensing?
The pricing, setup cost, and licensing were all free for us because it was a hackathon and the company came down to our campus for the event. We were the winners of the challenge they provided.
Which other solutions did I evaluate?
We had many other options such as DeepSeek, Claude, ChatGPT, Copilot, and Gemini model. However, this feature is very rare because Glean Platform is an overall 360-degree product that helps from A to Z, rather than addressing only specific areas.
What other advice do I have?
After using Glean Platform, it does not give me any irrelevant answers or leave me with second thoughts. Instead, it provides very well-structured and on-point answers to my queries. The main factor is that every token is very costly and each token is very impactful. With less context switching, it helps me implement the queries I have or the routine tasks that exist.
A key area for improvement is that they can implement more models. However, there is always room for improvement because AI does not evolve over a decade but rather evolves every second, every minute, and every day. I rate this product a 9 out of 10.
Which deployment model are you using for this solution?
Public Cloud