Problem: Currently, Glean agent fields are constrained to accept only one type of input: either text, multiple choice answers, or attachments. Users who remember the flexibility of simple prompts, which allowed fields to accommodate both text and attachments simultaneously, find the current limitations frustrating. This restriction hampers the efficiency and versatility of agents, leading to a less streamlined user experience.
Use Case Example: Consider a scenario where a project manager is compiling notes and relevant documents for a team meeting. They want to input critical text commentary and/or attach the document where they took their notes. Under the current Glean agent setup, they are forced to use separate fields for text and attachments. This bifurcation complicates the data entry and agent making process, making it cumbersome and time-consuming.
Another example of our business: We are a recruitment firm that creates summaries of candidates for our clients. Our associates could be working for multiple partners who like their summaries formatted differently. Having the ability to have one "format example" field where they could either enter in text how they want their summary formatted, or feed in an example summary document, would be extremely helpful.
Solution: To enhance user experience and restore the functionality seen in simple prompts, we propose enabling Glean agents to support multi-input capability. Specifically, an agent field that can accept both text and attachments within the same field. This change would streamline the information-gathering and submission process, making data entry more efficient and flexible. Allowing dual inputs in agent fields would also align Glean's current capability with its earlier user-friendly features, addressing user needs and improving overall satisfaction.
