Client feedback (overall impression)
"My first impression is that this is trying to do too many things and the previous knowledge hub is more appealing and easier to use." And: the old one's "Research" surface "looks and feels more polished" (client reserving judgement, early days).
Concrete signal
The earlier prototype the client liked (bolt.host "Sample API Application") has 6 focused surfaces (Dashboard, Research, Knowledge, Graph, Agentic Retrieval, Self Assessment). The current build has 10 (Explore, Search, Library, Assistant, Investigations, Generate, Assessment, Graph, Help, Manage). The breadth is the likely source of "too many things".
For discussion (Monday)
- Decide the default/first-run experience and which surfaces lead vs sit behind (or fold together).
- Study what the earlier "Research" surface did well (polish, snippets, confidence) and bring it to the current Assistant/Search.
- Note: several polish + UX items are already improved on the current fly.dev build (merchandised cards, collapsed evidence, not-found page) but the client evaluated an earlier deploy - ensure they review the current version.
Client feedback (overall impression)
"My first impression is that this is trying to do too many things and the previous knowledge hub is more appealing and easier to use." And: the old one's "Research" surface "looks and feels more polished" (client reserving judgement, early days).
Concrete signal
The earlier prototype the client liked (bolt.host "Sample API Application") has 6 focused surfaces (Dashboard, Research, Knowledge, Graph, Agentic Retrieval, Self Assessment). The current build has 10 (Explore, Search, Library, Assistant, Investigations, Generate, Assessment, Graph, Help, Manage). The breadth is the likely source of "too many things".
For discussion (Monday)