Skip to main content
Every week, before the grants pipeline review, someone has to pull together where their grant requests stand — what’s moving, what’s stalled in an approval stage, what’s gone quiet. That means opening Salesforce, running or re-running a report or list view, skimming every open request by hand, and holding it all in your head long enough to speak to it. It’s not hard work; it’s the same tedious 20–30 minutes, every single week. This walkthrough uses Aperium’s live connection to Salesforce to do the compiling for you in plain language. A generic chatbot has never seen our Salesforce org — it doesn’t know your grant requests exist, let alone which ones are stuck in approval — so it can only produce a plausible-sounding “how to prioritize your pipeline” answer that isn’t about your requests at all. For the why, see Connected AI vs. a generic chatbot.
Hillspire’s Salesforce is a grants-management org, not a sales org. A grant request is what the system calls an “opportunity,” and a grantee organization is what it calls an “account.”
  • Open / in-flight stages: New Request, Concept, Program Review, Program Director Review, Legal Review, Senior/Executive Director Approval, Grant Agreement, GPS Triage.
  • Terminal stages: Grant (awarded) and Closed.
  • Staleness is judged by last-modified date — the “last activity” field in this org is unreliable, so don’t lean on it.
As Aperium works, look for a step showing it querying Salesforce — a tool or action indicator, separate from its written answer. That’s the “do” half of talk-and-do. The exact visual cue depends on your Aperium build; if you’re unsure, ask Aperium what it just did.

​
The walkthrough

1

Pull your open grant requests, grouped by stage

Ask it the way you’d think it — no report to configure, no list view to build:
Show me my open grant requests, grouped by stage.
Aperium queries Salesforce, scopes to your own requests, and returns them organized by stage. Saying “my” scopes it to you automatically — see the caution below on what that means.
2

Flag what needs attention this week

Stay in the same conversation — no need to re-describe the list:
Flag anything that hasn’t moved in 30+ days or is stuck in an approval stage.
Aperium highlights the subset that actually needs attention, using last-modified date for staleness. Sanity-check the flags against what you know; if one looks off, confirm it’s reading last-modified date and not the unreliable last-activity field.
3

Get a one-liner per flagged request

This is the part that usually takes longest by hand — turning a row of fields into a sentence you’d actually say:
Give me a one-line status on each so I can speak to it in the review.
Each line is grounded in the real fields — stage, amount, last-modified date — not generic filler.
4

(Optional) Draft a pipeline summary — then review it

Draft a short pipeline summary for the review.
This is a starting point to bring into the room, not a finished deliverable. Read it against what you actually know before you rely on it or forward it to anyone.

​
This is also an automation

The question here never changes — only the underlying data does. That makes it a textbook candidate for a scheduled job: the same review-prep can run every Monday morning and land in your inbox before you’ve had coffee, instead of being asked for live each week. We ran it live so you could see every step; it doesn’t have to stay a live ask. See Workflows for how scheduled jobs work.

​
Before you speak to it in the review

Verify before the meeting. Check any figure, stage, or last-modified date against Salesforce itself before you say it out loud — the same misquote risk as with any AI-produced summary.
  • It acts as you. This runs against the logged-in user’s grant requests and permissions, so it only ever shows what you’re allowed to see. Someone else running the same request gets their own book, not yours — that’s by design.
  • A draft is a draft. The optional summary is a starting point to bring into the room, not a finished document to forward or read verbatim without a look-over first.
For the habits that apply across every connector, see Responsible use.

​
Try it yourself

Run this exact request against your own grant requests before your next weekly review. If it’s useful, set it up as a scheduled job so it’s already waiting for you next Monday.