# Salesforce prompts your whole team can trust and repeat. > The SyncOnAI 360 Prompt Library is a team library of Salesforce prompts with categories, versions, variables, favourites and execution history. Variables are filled with pickers that read the connected org, prompts run through the existing chat engine with its tools and approvals, and every run is recorded so it can be reviewed or run again. Source: https://synconai360.com/features/prompts ## Key facts - **1**: Engine: prompts run through the same governed chat - **Every**: Run recorded in history, with rerun - **Org**: Pickers fill variables from your own metadata - **0**: Separate runtime to secure or maintain ## Why AI prompts for Salesforce stay personal and inconsistent The prompts that work best live in one person's notes. Everyone else types their own version, gets a different answer, and nobody can repeat what worked last time. - **Copy and paste.** Good prompts are shared in chat threads and documents, then edited until they stop working. - **Wrong field names.** A prompt written for one org mentions objects and fields another org does not have. - **No history.** When a prompt produced a great result, nobody can find exactly what was asked or rerun it. - **Outside the guardrails.** Prompts pasted into general tools run without the org's context or any approval. ## What is in a Salesforce prompt library? The library holds verified prompts organised by category, with versions, so a team uses the same proven wording instead of each person's improvisation. Favourites keep the ones you use most at hand. Open the library from the Prompts button in the chat composer, or from the Prompts page, where each prompt's history is kept alongside it. - Verified prompts by category - Versions and favourites - From the composer or the Prompts page ## How do prompt variables use your Salesforce org? A prompt's variables are filled with pickers that read the connected org: pick an object and the field and record type pickers reload to match it. The prompt runs with real API names from your org, so it works the first time instead of referencing fields that do not exist. When you rerun a prompt from history, its inputs are prefilled, and it will not silently swap in a different org. - Pickers that read your org - Dependent pickers for fields and record types - Rerun prefilled, never a silent org swap ## Do library prompts follow the same approvals? Yes. Prompts run through the existing chat engine, not a separate runtime, so they get the same org context, the same tools and the same guardrails. A prompt that leads to a change produces a proposal with checks and blast radius, and production needs a second admin. Specialist agents can be paired with prompts too, so a Flow prompt runs with the Flow and Automation agent's instructions and tool permissions. - Same engine, tools and guardrails - Changes become proposals - Pairs with specialist agents ## How do teams learn from prompt runs? Every run is recorded in an execution history you can filter, open in detail, annotate with feedback and notes, and rerun. What worked, and exactly what was asked, stays findable for the whole team. Over time the library becomes a record of how your team uses AI on Salesforce, not a pile of personal snippets. - Filterable run history - Feedback and notes per run - One-click rerun ## Which Salesforce jobs suit a library prompt? Repeatable jobs with a known shape, such as documenting an object, reviewing a Flow, drafting a validation rule or explaining a class. Wording that works once can be saved, versioned and run by everyone, with its variables filled from the org. The library is organised by category so the right prompt is quick to find, and favourites keep each person's most-used prompts one click from the composer. Jobs that are new every time are better asked directly in the chat. - Repeatable jobs with a known shape - Organised by category - One click from the composer ## Using a library prompt 1. **Pick.** Choose a prompt from the composer or the Prompts page. 2. **Fill.** Choose objects, fields and record types with pickers from your org. 3. **Run.** It runs in the chat with the org's context and tools. 4. **Review.** Any change becomes a proposal for approval. 5. **Reuse.** Find the run in history, add notes, rerun it later. ## Prompts, personal versus shared | Without | With SyncOnAI 360 | |---|---| | Snippets in someone's notes | A verified library with versions | | Field names that do not exist | Pickers that read the org | | Results nobody can repeat | Every run recorded with rerun | | Run outside any guardrails | Same chat engine, tools and approvals | | Lessons lost | Feedback and notes on each run | ## Who uses the prompt library - **Consultancies.** Turn the practice's best prompts into a shared, versioned library, so every consultant runs the proven wording against each client's org with real API names filled in. - **Salesforce admins.** Run verified prompts for common jobs with pickers that read your org, so a prompt never refers to a field that does not exist. - **Delivery leads.** See what was asked and with what result in the execution history, add notes and feedback, and rerun what worked. - **Platform owners.** Standardise how AI is used on the org: prompts run through the same chat, tools and approvals as everything else, with no second route into production. - **Salesforce developers.** Keep prompts for code review and test writing beside the ones admins use, run with the same org context and recorded in the same history. ## Frequently asked questions ### What is the SyncOnAI 360 prompt library? A team library of verified Salesforce prompts with categories, versions, variables, favourites and run history, opened from the chat composer or the Prompts page. ### Do prompts use our org's field names? Yes. Variables are filled with pickers that read the connected org, so prompts use your real objects, fields and record types. ### Do prompts run differently from normal chat? No. They run through the same chat engine with the same tools and approvals. ### Can we see past runs? Yes. Every run is recorded in history, with filters, detail, feedback, notes and rerun. ### Can a prompt change our org? Only through a proposal with checks and approval, the same as any other change. ### Can prompts be triggered with a slash command? Not yet. Open them from the Prompts button in the composer or from the Prompts page. ### Is the prompt library on every plan? Yes. Every plan includes every feature. ### Do prompts work with specialist agents? Yes. Pick an agent in the composer and the prompt runs with that agent's instructions and tool permissions. ### Do prompts run against the right org? Yes. Variables are filled from the connected org, and a rerun from history keeps its original inputs and will not silently switch to a different org.