- The Ops Digest
- Posts
- The Deals You're Losing Without Knowing Why
The Deals You're Losing Without Knowing Why
A Competitive Intel Problem Nobody Solves
Welcome to The Ops Digest!
Each week, we drop no-BS insights and one AI prompt to cut wasted costs, tighten workflows, and eliminate manual grunt work.
Today: Your competitors are making moves right now. Pricing shifts, new hires, product launches. By the time your reps hear about it from a customer, it's too late. We're building an AI agent that catches it first.
The Competitive Intel Problem Nobody Solves
Sales teams face direct competition in 68% of their deals. But when asked how prepared they feel to handle those situations, the average score is 3.8 out of 10 (Crayon, 2025 State of Competitive Intelligence). That gap costs mid-market companies an estimated $2M-$10M per year in lost deals (Salesmotion).
You know why it happens. Someone builds a shared spreadsheet. Someone else creates a Slack channel. Both are abandoned within six weeks because nobody's job is "go Google our competitors every Monday."
Today we're building a Claude setup that does the Googling, reads your competitors' actual websites for changes, scans your Slack and email for competitive mentions, and compiles everything into a one-page briefing your reps can use before calls.
Quick Insight:
Your competitors are already automating order entry.
While your team is still keying in POs, theirs are processing them instantly.
Build It: The Competitive Intel Agent
Step 1: Connect Your Internal Tools
Open the Claude Desktop app (Mac or Windows). Go to Customize → Connectors. You'll see a list of available integrations. Connect the ones where competitive intel lives in your org:
Slack - so Claude can search channels where reps mention competitors
Gmail - so Claude can scan for competitor mentions in customer emails
Google Drive - so Claude can reference past battle cards or reports
Each one uses OAuth. Click "Connect," sign in, done. About 30 seconds each.
Don't have Slack or Gmail? Skip this step. The web scraping alone is more competitive monitoring than most teams do.
Step 2: Add Firecrawl (The Web Scraping Piece)
This is what takes this from "decent" to "actually useful." Claude's built-in web search only sees search result snippets. Firecrawl lets Claude read entire web pages: your competitor's product catalog, their careers page, their press releases.
Setup:
Go to firecrawl.dev and create a free account. The free tier gives you 500 credits, plenty to test this.
Copy your API key from the Firecrawl dashboard.
Back in Claude Desktop: Customize → Connectors → Add Custom Connector. Name it "Firecrawl" and paste this as the MCP server URL:
https://mcp.firecrawl.dev/YOUR-API-KEY/v2/mcp
(Replace YOUR-API-KEY with the key you copied.)
That's it. Claude can now scrape and search the web like a research assistant with a browser.
Step 3: Pull Your Internal Data
Export your closed-lost deals from the last 90 days. This tells the AI which competitors are actively taking your business, not just which ones exist.
Salesforce: Reports → Opportunities → Stage = Closed Lost, last 90 days. Include "Competitor" and "Loss Reason" fields.
HubSpot: Deals → Filter closed-lost → Export with deal notes, competitor mentioned, close date.
Dynamics 365 / Zoho / Pipedrive: Lost opportunities report with reason codes and notes.
Step 4: Run the Prompt
Open a new Cowork task in Claude Desktop. Upload your closed-lost CSV. Then paste this:
You are the Competitive Intelligence Agent for [Your Company Name], a [manufacturer/distributor] in [your industry].
COMPETITORS TO TRACK:
1. [Competitor A] - [their website URL]
2. [Competitor B] - [their website URL]
3. [Competitor C] - [their website URL]
DO THE FOLLOWING:
## PART 1: WEB INTELLIGENCE
For each competitor, use Firecrawl to:
- Scrape their website's news/press page and product pages. Flag anything new or changed.
- Search the web for their company name + "press release" OR "announcement" OR "partnership" from the past 7 days.
- Check their careers/jobs page. New sales hires in our territory = incoming competition.
## PART 2: INTERNAL SIGNAL SCAN
Search Slack (last 7 days) for mentions of each competitor name. Extract who mentioned them, the context, and whether it was tied to a specific deal.
Search Gmail (last 7 days) for emails containing competitor names. Flag any customer email mentioning switching, comparing, or requesting a price match.
## PART 3: LOST DEAL PATTERNS
Analyze the uploaded closed-lost data. Which competitors appear most often? What reasons come up repeatedly? Are there product categories where we lose disproportionately?
## PART 4: BRIEFING
Compile into a single report:
COMPETITOR MOVES
- One paragraph per competitor, only if there's real news
- Skip anyone with nothing to report
FIELD INTELLIGENCE
- Every competitor mention from Slack/email, with context
- Flag anything tied to an active deal as HIGH PRIORITY
LOSS PATTERNS
- Top 3 competitors by lost deal count
- Most common loss reasons with dollar amounts
SUGGESTED ACTIONS (2-3 max)
- Specific things to do this week
- Be direct: "Call [customer type] before Competitor A's new pricing takes effect"
RULES:
- No filler. If a competitor had no news, skip them.
- Keep the briefing under 500 words.
- Tone: morning briefing, factual, no commentary.Don’t Lose Deals - Reduce Friction
Your competitor isn’t better, they’re just faster.
Y Meadows removes the manual order entry bottleneck so orders flow from email straight into your ERP.
What To Do With the Output
Run this every Monday morning. Takes about 2 minutes once it's set up. Then pick the distribution method that fits your team:
Slack: Copy the briefing into #sales-intel, or ask Claude to post it directly using the Slack connector.
Sales meeting: Use the "Suggested Actions" as your Monday pipeline review opener.
Google Drive: Save weekly briefings to a shared folder. After a quarter you'll have a competitive trend archive that didn't exist before.
An Honest Take
There are dedicated competitive intelligence platforms. Klue, Crayon, Salesmotion. Good products. They also run $20K-$50K per year and require someone in product marketing to maintain battlecards full-time.
My actual opinion: for a 10-50 person sales team at a mid-market distributor or manufacturer, that's overkill. You don't need a CI platform. You need someone to check on three competitors once a week and tell your reps what changed. That's the whole job. And now AI can do it for the cost of a Claude Pro subscription and a free Firecrawl account.
If you outgrow this, buy the platform. But start here.
Case in Point
A regional fastener distributor with 14 outside reps had tried competitive intelligence twice. First, a shared Google Sheet. It lasted six weeks. Then a monthly email from the VP of Sales summarizing what he'd heard at trade shows. That lasted three months.
They set up this workflow tracking 4 competitors. Within the first two weeks, the Firecrawl scan of Competitor B's careers page turned up three new sales rep job listings in the distributor's core territory. Nobody on the team had noticed.
Two weeks later, Competitor B started calling on their accounts. But because the distributor had advance warning, they'd already reached out to their top 20 accounts in that territory. They retained all 20.
Total time: 15 minutes to set up, 2 minutes per week to run.
The Bottom Line
Competitive intel isn't a technology problem. It's a consistency problem. The information is out there, on public websites and in your own Slack channels. Nobody is collecting it.
Set it up. Run it for a month. Then decide if it's worth keeping.
👇 👇 👇
The AI catches the intel. But acting on it is a different problem.
If pulling order history, margin data, and buying patterns takes hours - you’re too late.
Y Meadows automates order entry and structures your data at the source, so when competitive pressure hits, you can actually move.