Free AI sales forecasting that shows what will really close this month
Add the free Arrows Sales Skills to Claude and it scores every deal the CRM has closing this month from what the buyer said and did, and puts that next to the CRM’s own number.

Get it running in 3 steps
Free, with no signup and no email gate.
Step 1: Add the free Arrows Sales Skills to Claude
In Claude Desktop, open Settings, then Connectors, then Add custom connector. Name it Sales Skills by Arrows, paste https://skills.arrows.to, click Add and restart Claude. Prefer files? Download the skill’s SKILL.md from GitHub and upload it under Settings, then Skills.
Step 2: Connect your CRM, email, calendar and call recorder in Claude
Turn on the Claude connectors for the tools your team uses. The more it can read, the better the read. Calls matter most here: without them, tiers that need a buyer’s own words are judged from email alone. No CRM connector? It asks once for a CSV export of open deals.
Step 3: Run it, or have it ready before every forecast call
Once the skills are added, type the prompt below in Claude. (In a Claude without the skills, the same words just get a generic answer.) It covers this month; name a quarter or another month to change the period. On the first run it offers to have the forecast ready every Monday at 7am. Run “Build my sales profile” once first (about 5 to 10 minutes) and it uses your own stages and methodology. Without a profile it still runs, with stages straight from the CRM.
With the skills added, type in ClaudeRun my Arrows forecast
What’s in the report
One page, the same layout every run, so people learn where to look. Every deal shows the CRM’s number next to the evidence, and every low number comes with a move. The chat reply leads with the decisions: which close dates to move, which deals to close out, and the one buyer action that would move a deal up a tier.
- 1
The header
Deals the CRM has dated this period, what the CRM says, what the evidence says and what’s likely to slip, with a bar of commit, likely, possible and unlikely by value.
- 2
What will really close, and when
Every deal, most likely first: its label, the buyer’s own date if they gave one, the one thing the buyer hasn’t done yet, and the one move on the seller’s side that would get it done. Open a deal for the evidence.
- 3
CRM ahead of the evidence
Deals whose stage probability or close date is ahead of what the buyer has done, ranked by money at stake, with the one fact behind each.
- 4
Further along than the CRM shows
Deals the evidence rates higher than their stage, plus deals dated just after the period that could pull in, with the buyer fact that earned it.
- 5
Likely to slip
Deals whose evidence date lands after the period, and the value that leaves the period with them.
- 6
By rep
For leaders: each rep’s CRM total next to the evidence, and one question about a specific deal for their 1:1. Open a rep for their deals.
- 7
How the forecast works
The rubric in plain words, so anyone can check a number.
How it judges what will really close
Every deal gets the same rubric, so the same evidence always gives the same number. It takes the highest tier the buyer’s actions have earned, then applies the caps. Only what the buyer said or did counts: a seller’s note like “verbal yes” needs a buyer message, call or meeting behind it.
| Rule | What it means |
|---|---|
| Commit90% | The contract or order form is with the buyer and the buyer named a signature date, or the budget holder gave a verbal yes in their own words with paper in motion. |
| Commit75% | The budget holder agreed to pricing or terms in writing or on a recorded call, and the buyer’s paper process has started. |
| Likely60% | The budget holder has met the seller, the buyer gave a decision date or process in their own words, and a next step is booked. |
| Possible40% | A champion is engaged and replying and a next step is booked, with two or more buyer contacts involved or the budget holder engaged. No buyer-stated decision date yet. |
| Possible25% | Two-way contact in the last 14 days, but no booked next step, or single-threaded, or a timeline that exists only in the CRM or the seller’s notes. |
| Unlikely10% | No buyer reply in 14 or more days, or no next step and no buyer reply in the last 7 days. |
| Not this period0% | The buyer said there’s no budget or chose someone else, or hasn’t replied in 30 or more days. |
| Then the capsCaps | No budget holder identified: at most 40%. Quiet 14 or more days: at most 25%. Close date moved twice or more: at most 40%. A competitor the buyer named in the last 30 days, with no stated preference: at most 60%. A buyer promise missed by more than 2 business days: one tier down. A new deal with no call held: at most 25%. |
- The evidence close date is the later of the date the buyer gave and the date the remaining steps could finish at the team’s usual pace. A deal that lands after the period counts as likely to slip and adds nothing to this period’s number.
- It’s about the CRM and the buyer, never the seller. A gap means the record is ahead of what the buyer has done, not that a rep is wrong.
- It’s read only. It never changes stages, close dates or probabilities in your CRM: it suggests, and you decide.
Use it before your forecast call
- 1
Have it ready Monday morning, so the call starts from the same page.
- 2
Spend the call on the deals where the CRM is ahead of the evidence.
- 3
Leave with the close dates to move, the deals to close out, and each rep’s one question.
How to run a sales pipeline review that changes your deals →
When you want more than a report
A report tells you what’s wrong. Arrows lets you act on it.
- During the review, a leader talks an action into a deal, and Arrows queues it on the deal’s gameplan.
- Arrows runs your team’s Agentic Playbook on every deal, without anyone typing a prompt.
- Reps review and send the drafts it prepares: follow-ups and CRM updates.
- Next week, you can see whether it got done.
Arrows is its own product, not part of Claude. Your team can keep using Claude alongside it.
Every tool. Every deal. No prompts.
An Agent Architect connects your stack and sets what Arrows can do in each tool. Then Arrows does the follow-through on every deal, and your reps approve what matters.
Book an agent kickoffObjective forecast: common questions
What is AI sales forecasting?
AI sales forecasting uses AI to judge which deals will close and when, instead of relying only on the stage and close date typed into the CRM. The Arrows objective forecast is a free Claude skill that does this from buyer-side evidence in your calls, email and calendar, with a fixed rubric, and shows the result next to the CRM’s own number.
Can AI like Claude or ChatGPT do sales forecasting?
Yes, when it can read your deals. Without access, it can only work from what you paste in, and it will judge differently each time you ask. Connected to your CRM, calls and email, Claude with the Arrows objective forecast scores each deal with the same rubric every time, so the same evidence always gives the same number, and shows the evidence under every number.
Do I need to install anything to use it?
Yes, the skills themselves. Add the free Arrows Sales Skills to Claude first: in Claude Desktop, go to Settings, then Connectors, add a custom connector with the URL https://skills.arrows.to, and restart Claude. Or upload the skill file from GitHub under Settings, then Skills. Without the skills, typing “Run my Arrows forecast” only gets a generic answer, not this report. Then connect your CRM, email, calendar and call recorder so it has deals to read.
How can I improve forecast accuracy?
Check the CRM’s number against what buyers have actually done. Before the forecast call, this skill shows which deals the CRM has ahead of the evidence, which are further along than their stage, and which are likely to slip past the period, with the evidence for each. You decide which close dates to move and which deals to close out.
How is this different from the forecast in my CRM?
The CRM forecast comes from stage probabilities or forecast categories and the close dates typed into each deal. The objective forecast sets those aside and scores each deal only from what the buyer said and did, then shows both numbers side by side.
Will it change close dates or probabilities in my CRM?
No. It’s read only. It suggests which close dates to move and which deals to close out, and the person makes the change. That keeps it safe to run on a schedule.
Does it score or rank reps?
No. It’s about the CRM and the buyer, never the seller. Leaders get each rep’s CRM total next to the evidence and one question about a specific deal for the 1:1, with no scores, rankings or adjectives about people.
More free sales skills
For sales leaders
- Weekly pipeline reviewOne shareable page: what’s moving, what’s at risk, what to do this week. Leaders get a rollup by rep.
- Process gaps reportWhat’s falling through the cracks: money at risk, missed follow-ups, and this week’s moves.
- CRM hygiene checkLearns how your pipeline really works, then keeps your CRM clean with fixes from your calls and email.
- Win-loss analysisWhy you win and why you lose, from your closed deals: the moves that win, the real loss reasons, and 3 things to change.
- Sales methodologyPuts MEDDPICC, BANT, SPICED, or your own process into Claude, then shows which wins followed it and which deals don’t.
- Agentic playbook builderFor sales leaders: writes down how your team runs its sales process, play by play, from your real deals and calls.
For reps
- SetupLearns how you sell from your CRM, calls, and email. Run this first.
- Daily briefYour day, what changed since yesterday, and what to do first.
- Pre-call prepA 60-second brief for one call, gaps included.
- Post-callWrites your follow-up and updates the CRM from what was said, after you approve.
- Deal nudgeFinds deals at risk and drafts a short nudge for each, in your voice.