Deshu Goyel

Deshu GoyelAug 4, 2026

Transition https://lnkd.in/gD8Q6CFz

Deshu GoyelJun 5, 2026

I built an AI agent that replaced a $1,200 manual consulting job. Here's exactly how I did it using Hyperagent. Background: There's an entire industry of "pricing consultants" who charge local businesses $1,200–$3,600 to manually research competitor pricing and find where they're undercharging. Each audit takes 3–4 hours. The consultant goes to Google Maps, finds competitors, scrapes their websites, builds a comparison, and presents the findings. I automated the entire process. I built PriceGap AI inside Hyperagent — a multi-agent system that does the same $1,200 job in 5 minutes, automatically, every week. Here's the exact architecture I built: The Architect (meta-agent) The first agent interprets a business owner's requirements in plain English and designs the full research strategy. It decides which search queries to run, what signals to look for in competitor websites, and how to structure the analysis. I taught it in plain text — no code, no API configuration. The Research Agent Connects to Brave Search MCP server to find 8–12 real competitors in the target city. Then scrapes their pricing pages using Puppeteer. Every Monday at 6AM, it runs automatically for every active subscriber. The Analysis Agent Takes the raw competitor data and calculates the exact revenue gap. If a cleaning business charges $89/job and the market average is $124 — it calculates $35 × 35 jobs × 12 months = $14,700/year just from underpricing. Then it adds missing services and no package structure to get the full hidden revenue number. The Critic (self-improvement loop) After every report, this agent scores its own output on accuracy, completeness, and actionability. If it scores below 7/10, it rewrites the relevant prompt and A/B tests the new version on the next run. The reports get measurably better every week without me touching anything. The Librarian (memory) Stores market data from every audit — pricing trends, seasonal patterns, what works in each city. By run 10 in Austin's cleaning market, the agent knows things about that market that no human analyst would track manually. What Hyperagent made possible that I couldn't do before: → Connected all 5 agents with plain English instructions → Telegram integration for instant alerts when major gaps detected → Scheduled runs — every Monday 6AM, automatic → Auto mode — runs without me being present → Full knowledge access across all agents The result: A local cleaning business owner pays $97/month. The agent finds them an average of $26,940/year in pricing gaps. The business gets a PDF report in their inbox every Monday morning. I do nothing after setup. The system I built inside Hyperagent went from idea to working product in one week. The same system, built manually, would have taken months and required a developer.Hyperagent #Hyperagent #AIAgents #AgentFirst #SaaS #PriceGapAI

Deshu GoyelSep 5, 2026

😂 LinkedIn These Days… Anyone Else Noticing This? LinkedIn has changed a little too much lately. 😅 Here’s what keeps showing up on my feed: 1️⃣ Zomato / Blinkit / Zepto Especially the newspaper ads. Every other day, there’s a new hiring story. 👀 2️⃣ HR: “What are your expected CTC?” Candidate: “12 LPA.” HR: “Our budget is 7 LPA.” Candidate: “Thanks, but I’ll pass.” Meanwhile… The same candidate gets a 14 LPA offer somewhere else. 😂 👉 Lesson: Know your worth. The right company will value it. 3️⃣ Rapido user experience stories Apparently, everyone has a story. 😂 4️⃣ “Learn AI or be homeless.” 🤖😂 Because apparently, there’s no middle ground anymore. Just a few observations from my LinkedIn feed lately. What’s dominating your feed these days? 👀 Drop your answer in the comments. #LinkedIn #Hiring #Recruitment #Jobs #JobSearch #Career #Workplace #HiringNow #FreshersJobs #AI #CorporateLife #LinkedInIndia #JobAlert #HireJobHub