Case Study: SaaS Company Automates Competitor Research, Saves 20 Hours Per Week

Company: Propel CRM (pseudonym), growth-stage CRM platform Team size: 35 (12 engineers, 8 sales, 6 marketing, 5 product, 4 operations) Challenge: Manual competitor research consumed 20 hours/week of product team time Result: Fully automated competitive intelligence, 5 minutes per weekly report, product decisions 3x faster

---

In competitive SaaS markets, knowing what your competitors are doing isn't optional -- it's survival. But tracking pricing changes, feature launches, positioning shifts, and customer sentiment across 8+ competitors takes enormous time.

Propel CRM's product team was spending 20 hours per week manually researching competitors. That's half a full-time employee just reading competitor blogs, checking pricing pages, scanning review sites, and synthesizing findings into reports.

They automated the entire process with an AI research squad on Ivern. Now, a weekly competitive intelligence report takes 5 minutes to generate. The product team uses those saved 20 hours for actual product work.

Related: AI Research Assistant: How It Works · How to Automate Research with AI Agents · How to Build an AI Competitive Intelligence Workflow · AI Research Assistant Tools

The Problem

Propel CRM competes in the crowded mid-market CRM space against 8 direct competitors. Their product team needed to track:

Research Area	Frequency	Time Spent
Pricing page changes	Weekly	3 hours
Feature launches & updates	Weekly	5 hours
Review site analysis (G2, Capterra)	Weekly	4 hours
Content & positioning changes	Weekly	3 hours
Customer case studies & testimonials	Monthly	3 hours
Funding & company news	Weekly	2 hours
Total		~20 hours/week

The product manager and one senior engineer rotated research duty. It was thorough but slow, and it pulled them away from product strategy and roadmap planning.

The AI Research Squad

Propel built a 4-agent research squad in Ivern. Each agent specializes in a different aspect of competitive intelligence.

Agent 1: Market Scanner

• Model: Gemini 2.5 Pro (free tier)
• Role: Scan competitor websites, blogs, and news for updates
• Schedule: Runs weekly
• Prompt:

  "Analyze the following competitors for updates in the past 7 days: [competitor list]. For each competitor, check for: new feature announcements, pricing changes, blog posts, press releases, and significant homepage or positioning changes. Return a structured summary organized by competitor."

Agent 2: Review Analyst

• Model: Gemini 2.5 Pro (free tier)
• Role: Analyze customer reviews on G2, Capterra, and TrustRadius
• Prompt:

  "Analyze recent reviews (last 30 days) for [competitor name] on major review platforms. Identify: most praised features, most criticized features, common switching triggers (why customers leave), and sentiment trends. Compare against previous month's analysis for changes."

Agent 3: Feature Comparator

• Model: Claude Sonnet 4
• Role: Deep-dive feature comparison based on scanner findings
• Prompt:

  "Given the competitor updates identified by the market scanner, perform a detailed feature comparison with our product [product name]. For each new or changed competitor feature, assess: what it does, how it compares to our equivalent, what gaps it reveals in our product, and recommended response (match, differentiate, or ignore). Present as a priority-ranked action list."

Agent 4: Intelligence Briefing Writer

• Model: Claude Sonnet 4
• Role: Synthesize all findings into a weekly executive briefing
• Prompt:

  "Synthesize the market scanner, review analyst, and feature comparator outputs into a weekly competitive intelligence briefing. Structure as: (1) Executive Summary (3 bullet points), (2) Competitor-by-Competitor Updates, (3) Feature Gap Analysis, (4) Recommended Product Actions (priority ranked), (5) Market Trends to Watch. Keep under 1,500 words. Write for the product leadership team."

The Weekly Process

Step	Who	Time
Run Market Scanner	AI agent	3 minutes
Run Review Analyst	AI agent	2 minutes
Run Feature Comparator	AI agent	3 minutes
Run Briefing Writer	AI agent	2 minutes
Human review & add context	Product manager	5 minutes
Total		~15 minutes

The product manager reviews the briefing, adds context from customer conversations and sales calls, then shares it with the leadership team. Total human involvement: 5–10 minutes per week.

Results After 90 Days

Time Savings

Metric	Before	After	Change
Weekly research time	20 hours	15 minutes	-99%
Time to decision on competitor moves	2–3 days	Same day	-80%
Research coverage (competitors tracked)	5 of 8	All 8	+60%
Reports produced	1/month (ad hoc)	1/week (automated)	+300%

Strategic Impact

Metric	Before	After
Competitive features responded to within 2 weeks	40%	85%
Product decisions informed by competitive data	30%	90%
Pricing adjustments per quarter	1	3
Feature gaps identified proactively	2/quarter	8/quarter

Cost

Item	Cost
Gemini 2.5 Pro (scanning + reviews)	$0.00 (free tier)
Claude Sonnet 4 (comparison + briefing)	$0.25/week
Weekly cost	$0.25
Monthly cost	$1.00
Annual cost	$12.00
Product team time saved (20 hrs/week × $75/hr)	$78,000/year

How It Changed Their Strategy

1. Faster Response to Competitor Moves

Before automation, Propel learned about a competitor's major pricing change 10 days after it happened. Now, they know within 24 hours and can adjust their own positioning the same week.

2. Data-Driven Roadmap Prioritization

The weekly competitive reports revealed that 3 of their planned features were being commoditized by competitors. They reallocated that development time to differentiating features instead -- saving an estimated 6 weeks of engineering time.

3. Win/Loss Analysis Improvement

Sales teams now reference the competitive briefing in discovery calls. When prospects mention competitors, the sales rep has current data on feature differences, pricing gaps, and known weaknesses. The close rate on competitive deals improved from 22% to 31%.

4. Proactive Positioning

Instead of reacting to competitor content, Propel's marketing team uses the competitive gaps identified each week to create targeted content. Blog posts addressing weaknesses in competitor products now rank for competitor-branded keywords.

Lessons Learned

1. Structured Output Is Essential

The first version of the briefing writer produced narrative paragraphs that were hard to scan. After reformatting the prompt to produce structured sections with clear headers and action items, the briefings became immediately actionable.

2. Human Context Still Matters

The AI misses things that a human product manager wouldn't -- subtle positioning shifts, the significance of a VP leaving a competitor, or the implied strategy behind a feature rename. The 5-minute human review catches these nuances.

3. Free Models Are Good Enough for Research

Gemini 2.5 Pro's free tier handles the scanning and review analysis perfectly well. Propel only pays for Claude Sonnet on the comparison and briefing steps, where deeper reasoning produces better output.

4. BYOK Means Zero Usage Markup

Propel uses their own Google and Anthropic API keys. Ivern doesn't add a markup. A full year of automated competitive intelligence costs less than a single hour of product manager time.

Build Your Own Research Squad

1. Sign up free at ivern.ai/signup
2. Add API keys -- Google (free) and Anthropic ($5 credit)
3. Create a 4-agent research squad with the roles above
4. Run your first competitive intelligence report this week
5. Set it to run weekly and review the briefings

The free tier's 15 tasks cover 3 complete research cycles -- enough to validate the workflow.

Ready to automate your competitive intelligence? Create your research squad →

---

This case study is based on aggregated patterns from Ivern users in SaaS companies running competitive intelligence automation. Results represent typical outcomes for product teams tracking 5+ competitors. Individual results vary based on market complexity and competitor landscape.