Should bookkeepers focus on building custom AI workflows in Claude, or just rely on AI features already built into their bookkeeping software (e.g., Double)?

One member raised the idea that a bookkeeping platform called Double is adding many AI tools directly into the product, raising the question of whether bookkeepers should instead focus on learning to use built-in AI tools rather than building custom Claude workflows. The response acknowledged there's a lot of curiosity right now around building custom tools, and that as the barrier to building things with AI keeps getting lower, more specialized software will likely emerge. There was general agreement that small 'connector' workflows — moving data between systems, removing PII, controlling team access to information — are a great use case for AI-built tooling, distinct from relying solely on vendor-embedded AI features. No firm consensus was reached on which approach bookkeepers should prioritize.