Your reps lose hours every week to assembly: researching a prospect across four tabs, rebuilding last week's forecast, and checking quotes against fuzzy discount rules. Below are five AgentSpot library builds that take assembly off your go-to-market (GTM) team, each with a copy-paste prompt one person builds and everyone runs. Once they're running, your reps walk into calls briefed, and every quote gets checked against your commercial rules before a customer sees it.
Why Do Sales Reps Lose Selling Time?
Your reps lose selling time because every call, forecast, and quote draws on separate systems, and someone has to stitch them together by hand. Account history sits in Salesforce, while your commercial rules live in docs and Slack threads that only a few people know to check. What does that cost you?
Every hour of assembly comes out of selling time, which is one of the numbers you're measured on, alongside forecast accuracy and quote turnaround. The work also has a telling profile: it repeats every week and looks the same no matter who does it.
Which kind of AI handles work like that? The humble kind, built to do one recurring job reliably.
Before you pick a build, though, pick the shape.
When Should You Use an AI Agent?
Use an agent when the question changes and a workflow when the job repeats; when many people need one live view, build an app. A lot of first-build friction comes from a shape mismatch, like a chat-style agent for a job that should simply run every Monday. If you match the shape to the job up front, you skip a rebuild later.
| Shape | Use it when | How it starts |
|---|
| Agent | The question changes every time | You ask it |
| Workflow | The job repeats the same way | A schedule or a trigger |
| App | Lots of people need the same live view | Someone opens it |
The five builds below include two agents and two workflows, plus one workflow paired with a live app. For a deeper look at what agents need to act, see this guide on AI agent tool access.
What Should an AI Prep Note Include?
An AI prep note should include an account snapshot, the three whys, a short talk track, and the likeliest objection. With eight minutes before a call and tabs open on Salesforce, the prospect's newsroom, and the enablement folder, your rep can only triage.
What if an agent did the assembling? With the library's SDR Prospecting agent, built for sales development reps (SDRs), a rep types "Brief me on Acme Corp before my 2 p.m. call." The agent then pulls the account from Salesforce, reads your enablement content, and checks the web for recent news before emailing a prep note.
The piece to copy is the three whys, which a rep would otherwise improvise live:
Why change? What the current situation is costing this account today, in their own language.
Why now? The trigger, such as a leadership hire, a funding round, or a missed quarter.
Why us? The one capability that maps to their problem.
If the note can't answer all three, your rep is heading into a discovery call, and it helps to know that before dialing. Paste this prompt into AgentSpot to build your own:
> Build an agent that writes call prep notes. When I name an account and a meeting time, pull the account and open opportunities from Salesforce, check our enablement content for relevant positioning, and search the web for news from the last 90 days. Write a one-page note with an account snapshot, the three whys, a talk track with three discovery questions, and one objection to expect. Email it to me before the meeting.
Once one person builds it with a structure that works, every rep can run it. And when enablement updates the positioning, the next note picks up the change.
Prep notes cover the calls already on your calendar, but what about the deals that changed stage overnight?
How Do You Score Deal Readiness Automatically?
You score deal readiness automatically with a scheduled workflow that grades each qualified deal against a weighted rubric, with a reason for every score. Deals have a habit of changing stage at 4:47 p.m. on a Friday, so by the handoff your SDR and account executive (AE) are starting from zero with a gut feeling.
The library's Deal Research workflow watches for opportunities that move into your early stages. Each weekday morning, it researches them using Salesforce, your governed analytics data in ThoughtSpot, Slack, Confluence, Google Drive, and the web. Once that research is done, the owner gets a prep note, and qualified deals also get a readiness score from 0 to 100.
What makes a score worth trusting? A rubric that shows its work. Rate each deal from 0 to 5 on five dimensions, then apply the weights:
| Dimension | Weight | Question to answer |
|---|
| Champion strength | 25 | Is someone inside the account actively selling for you? |
| Business case clarity | 20 | Can the buyer state the cost of doing nothing? |
| Deal momentum | 20 | Are next steps dated and happening on time? |
| Competitive position | 20 | Do you know who else is in the deal and where you stand? |
| Technical fit | 15 | Does your product meet the requirements already on the table? |
To get the score, multiply each rating by its weight, add the results, and divide by 5. For example, a deal rated 4, 3, 4, 3, and 5 in the order above scores 75. The bands break down like this:
| Band | Score |
|---|
| Strong | 80 to 100 |
| Progressing | 60 to 79 |
| Early | 40 to 59 |
| At Risk | Below 40 |
Every score needs a one-line reason, because a score without a reason is just a hunch wearing a number. One build lesson from the library applies to yours as well: query each stage separately. Otherwise, a busy week of early-stage deals can crowd the qualified ones out of the results entirely.
> Build a workflow that runs every weekday at 7 a.m. Find opportunities that moved into our discovery or qualification stages in Salesforce in the last 48 hours, querying each stage separately. Research each account in Salesforce, our analytics data in ThoughtSpot, Slack, Confluence, Google Drive, and the web. Write a prep note with stakeholders, pipeline history, signals, talking points, and risks. For qualified deals, add a 0 to 100 readiness score across champion strength, business case clarity, technical fit, deal momentum, and competitive position, with a reason for each. Send each note to the deal owner on Slack and email.
Once you know which deals are real, how much of the quarter do they add up to?
Can a Sales Forecast Build Itself Weekly?
Yes: a scheduled workflow can read your open pipeline every Monday, flag risk deal by deal, and email the report before your forecast call. Without one, someone in sales ops exports the pipeline, reapplies the stage weights, and chases reps whose close dates have drifted into fiction. By the time the tables are rebuilt, you're reading Tuesday's forecast.
The library's Sales Forecasting workflow runs every Monday morning, reading the quarter's open opportunities in Salesforce and applying your stage ladder. Win rates and sales-cycle trends come from your analytics data.
So which part should you take for your own build? Two columns and eight flags.
Calculate weighted pipeline twice: once with your company's stage probabilities and once with the probability each rep entered. Wherever those columns disagree is where your forecast call should spend its time. Then flag every deal that is:
Slipping past quarter end
Past its close date
Quiet for 30 days
Quiet for 60 days
20 or more points off your stage ladder
Worth $100K or more and carrying any other flag
Closing within 30 days at 25% probability or less
Sitting at 50% or higher with no activity for 21 days
> Build a workflow that runs every Monday at 7 a.m. Pull all open opportunities closing this quarter from Salesforce. Calculate weighted pipeline using our stage ladder, plus a second column using each rep's Salesforce probability. Flag slipping, past-due, inactive, and mismatched deals. Break it down by segment and region, add an owner leaderboard, and email me the report with a Slack summary of the top three risks.
By Thursday, your managers will want to slice the numbers by region or owner. Describe that view, and AgentSpot builds it as a live app on the same data. Because the workflow and app read the same pipeline, everyone works from one set of numbers.
Forecasts catch risk inside the pipeline, while quotes carry risk straight to the customer.
How Can Deal Desk Automate Quote Checks?
Deal desk can automate quote checks with a daily workflow that reads your rules doc and flags every exception before a quote goes out. Your rules may live in a doc, three Slack threads, and one veteran's memory. When quarter end arrives and volume spikes, steps get missed.
Every weekday, the library's Daily Deal Desk Quote Sweep pulls new quotes from Salesforce and checks them against your rules doc. It flags expired terms, Closed Lost opportunities, and discount or currency violations, then posts the exceptions to Slack. The habit to adopt? Write the rulebook so a machine can check it:
Write one rule per line.
Use numbers in place of adjectives, so "more than 30 days" replaces "unusually long."
Name an owner for every rule.
Write the exceptions down.
Here's a pleasant side effect: rules clear enough for software are often, for the first time, clear enough for a new hire.
> Build a workflow that runs every weekday at 8 a.m. Pull quotes submitted in Salesforce since the last run. Read our rules from the Deal Desk Rules doc in Google Drive, check every quote against every rule, and post exceptions to #deal-desk, grouped by rule, with the quote, the owner, and what needs fixing.
When finance changes a threshold, someone edits one line in the doc, and the next run uses it.
Once the deal closes, though, where does all the account context live?
What Goes into a Customer 360 Brief?
A Customer 360 brief puts the relationship status, open commercial items, open support issues, and recent conversations on one page before you meet the customer. Customer context scatters fast across Salesforce, Slack, and Jira, with a different person on the account owning each piece. Meanwhile, the executive joining the quarterly business review (QBR) has none of it and 20 minutes to catch up.
The library's Customer 360 agent pulls from Salesforce, Slack, Atlassian, Google Workspace, and your enablement content, then sends a single brief before any customer meeting. At its core sit five questions every brief should answer:
Where does the relationship stand?
What's open commercially?
What's open with support?
What have they said lately?
What three things does this meeting need to answer?
If the brief answers the first four, you get to spend the meeting on the fifth.
> Build an agent that briefs me before customer meetings. When I name a customer, pull their account and opportunities from Salesforce, recent Slack mentions, open Jira tickets, and recent Google Drive docs. Write a one-page brief covering the relationship, what's open commercially and with support, who's changed, and three things to ask.
All five builds share one trait, and it matters more than any single prompt.
What Makes Multiplayer AI Different for GTM?
Multiplayer AI means one person builds an agent and shares it in one click, so your whole team runs the same version. When AI use stays single-player, every rep keeps a private stash of prompts. The cost shows up when someone leaves, since their best prompt walks out with them.
Shared builds keep that prompt in place. Side by side, the five above form a small multi-agent system: each build owns one job and draws on the same Salesforce records and rules.
Want to see what each one did? Every request and tool call is logged, so you can check.
You can read more about the thinking behind that approach in the AgentSpot launch post.
Before you build anything, run the multiplayer test with three questions:
Would three or more people run it every week?
Does it pull from systems more than one team owns?
Would it get worse if everyone built their own version?
Three yeses mean you've found your first build.
Which AI Agent Should You Build First?
Build the one that ruins your Monday, and leave the other four for later. Starting small keeps your time-to-value short: paste the prompt and run it on yourself for a week before sharing it with three people. Watch what they change, because their edits tell you what the build still needs.
| If this ruins your Monday | Start with |
|---|
| Reps walking into calls without context | Prep note agent |
| New-stage deals sitting unresearched | Deal readiness workflow |
| Rebuilding the forecast by hand | Monday forecast workflow |
| Quote errors at quarter end | Quote sweep workflow |
| Executives arriving at QBRs without the history | Customer 360 agent |
Build Your First GTM Agent Today
Busywork around selling repeats every week, which is exactly why those hours are the easiest ones to win back. Pick the build that costs you the most time and paste its prompt; once it works, share it with three teammates.
See AgentSpot in action: try it now, free, no card needed.
