Greenwood Cinema Rental: Email Inquiries to Current RMS Draft Orders, Without the Copy-Paste

It is working really well this week. I am basically running 50 of our orders through Runbear just to keep testing it.

Disclosure: Greenwood Cinema Rental is a Runbear pilot customer. This story is written from the implementation we delivered together.

Greenwood did not have a software problem. They had a copy-paste problem. Their team was running a real cinema and event rental business on Current RMS, and every new inquiry arrived as an email someone had to read, interpret, and retype into the system, one line item at a time.

Greenwood is exactly the kind of operator the bigger AI companies overlook. Mid-market revenue, a small operations team, a software stack that works well for inventory and scheduling but offers no native AI for the part of the day that consumes the most time: turning unstructured customer email into a clean, priced, available Current RMS order.

An email-to-quote agent is a Slack-native AI workflow that reads incoming rental inquiries, matches each requested item to the Current RMS catalog, checks availability for the requested dates, and creates a draft opportunity for a human to review and approve.

The Pain: Every Inquiry Was a 15 to 30 Minute Manual Task

In event AV rental, somewhere between half and two thirds of new business arrives by email. A typical inquiry looks like this:

Hi, we're running a corporate event at the Hilton Dec 15-18. We'll need a 12-channel mixer, 4 wireless lavs, 2 PA speakers, a projector, and a 10x12 screen. Can you also do delivery and setup on the 15th? Let me know pricing.

Every one of these meant the same chain of steps:

1. Read the email and figure out what the client actually wants
2. Open Current RMS and start a new opportunity
3. Search the catalog for each item, one at a time
4. Translate informal language ('lav', 'PA', '12-channel mixer') into the right SKU
5. Check availability against the requested dates
6. Set quantities, delivery, pickup
7. Generate and send the quote
8. Follow up

Done correctly, this is 15 to 30 minutes for a simple request and well over an hour for a complex one. Done at the volume a busy rental house actually faces, it eats the working day of whoever is on coordinator duty. Skip a step and you ship the wrong gear or double-book inventory.

The team was not asking for a new RMS. Current RMS was working. They were asking for the slow, manual translation layer between their inbox and Current RMS to stop being a human job.

What We Built

We deployed a Runbear agent that lives in Greenwood's Slack workspace and is connected to both their inbound rental email and their Current RMS account. When an inquiry comes in, the agent runs through a fixed pipeline:

1. Parse the email. Extract client, contact info, item list, rental dates, delivery and pickup details, venue, and special instructions.
2. Match to the Current RMS catalog. Fuzzy match informal item descriptions to real Current RMS products. Handle aliases ('lav' to 'Wireless Lavalier Microphone') and product groups ('PA speakers' to suggestions from the PA group).
3. Check availability. Query Current RMS for each matched product against the requested dates. Flag fully available, partially available, and unavailable items.
4. Create a draft opportunity in Current RMS. Populate subject, organization, contact, dates, delivery address, and matched line items. Attach the original email as a note. The draft stage does not impact stock, so it is safe to review.
5. Post to Slack for human review. A structured summary lands in the orders channel: client, dates, matched items, flagged items, availability status, a link to the draft opportunity, and quick actions to approve, edit, or reject.
6. Human approves. A team member checks the draft, edits if needed, and converts it to a quote. The agent never sends a quote without a human in the loop.

The coordinator's job changed from typing every order to reviewing every order. The system is built for trust, not autonomy.

Edge Cases the Agent Already Handles

Rental inquiries are messy. The pilot was scoped to make sure the agent fails gracefully instead of fabricating a clean order on top of an ambiguous one.

• Partial product match. Posts best guess with a confidence score ('80% match: 12-ch mixer to Yamaha TF1 Digital Mixer. Confirm?') and lists unmatched items separately.
• Custom or non-catalog item. Flags as 'Custom item, requires manual entry' with the original description, without blocking the rest of the order.
• Conflicting bookings or shortages. Checks availability before creating the draft and surfaces shortages in Slack ('Only 2 of 4 wireless lavs available Dec 15-18. 2 available from Dec 16.') with sub-rental or alternative suggestions.
• Multi-day rentals with different items per day. Creates a single opportunity with extended scheduler dates and notes sub-period scheduling in the description.

Why a Slack-Native Agent Instead of a New App

The team was never going to log in to a new dashboard to process orders. We wanted the AI to live where their work already happened, which for them is Slack and Current RMS. Two design choices made that real:

• No new interface for the team. Approvals, edits, and rejections all happen in a Slack thread. The Current RMS draft is the source of truth; Slack is the cockpit.
• No replacement of Current RMS. The agent reads and writes to Current RMS through its API. The system Greenwood already paid for and trained on stays the system of record.

How the Pilot Was Structured

Greenwood signed the Serve the Underserved pilot package:

• A short paid pilot focused on the email-to-quote workflow only.
• Full refund if the agent did not deliver value inside the pilot window.
• A path into the Runbear Business Plan after the pilot, priced for a mid-market operator, not an enterprise.

This structure exists because operators in underserved industries have been burned by software before. The most common objection in this segment is not price. It is 'my team will not actually use it.' The refund guarantee made the decision low-stakes, and the narrow scope made the value question answerable in weeks, not quarters.

Why This Pilot Matters Beyond Greenwood

Greenwood is the first proof point for a wider strategy at Runbear we call Serve the Underserved. The pattern looks like this:

1. Find the gap. Identify industry software that lacks official AI or MCP support but has community-built integrations, a signal that users want AI but their vendor is not delivering.
2. Reach out. Offer a short, paid pilot with a refund guarantee on the most painful workflow in that industry. Low risk for them, high signal for us.
3. Expand. Take what works in one shop and bring it to other operators on the same stack.

Current RMS users were the first vertical because email-to-quote is the universal pain in event AV rental and Current RMS sits in the middle of it. Greenwood validated that the agent can do that work for a real operator without changing the systems the operator already runs on. More rental operators sit in the pipeline behind them.

How It Works on Runbear

Runbear is a Slack-native AI platform that connects to the tools your team already runs on. For Greenwood, that meant connecting to inbound rental email and to Current RMS, then giving an agent in Slack the instructions and the catalog context it needs to convert one into the other.

Setup was non-technical. No custom code, no new app for the team to install, no migration off Current RMS. The operations team stayed in charge of the workflow; the agent just did the typing.

If you run an event AV or cinema rental operation on Current RMS, Rentman, Flex, or even a spreadsheet, and you are spending half your day translating inbox into orders, book a Runbear demo and we will show you the Greenwood pipeline.