Vijay Kumar
Vijay Kumar

I build AI products that solve real business problems, and drive revenue.

AI Product & GTM Lead · AI Product Manager · Co-founder, Convogenie AI

I take AI products from problem discovery and product definition through AI-enabled execution, launch and customer adoption.

vijayproject73@gmail.com·linkedin.com/in/vijay-convogenie-ai·+91 99720 25973·WhatsApp

Convogenie Lead Manager
AI lead management platform
176
paying businesses in 7–8 months
~$10.5K
MRR

Remote from India · 4–6 hours of US business-hours overlap (ET, CT, PT)

Featured case study

Convogenie Lead Manager

The problem

B2B sellers in India get leads from many places: IndiaMART, JustDial, TradeIndia, Alibaba, their website, email and WhatsApp. Leads sit in different inboxes, replies are slow (often hours, never at night), follow-ups are forgotten and buyers move on to the next supplier.

The opportunity

These businesses already pay for leads. The cheapest growth they can get is converting more of the leads they already have. An AI layer that replies instantly from the company's own product data, follows up every lead and shows the owner where leads leak is directly tied to revenue, which makes it easy to sell and easy to measure.

My role (what I personally owned)

  • Customer discovery with business owners and sales teams
  • Product decisions, roadmap and priorities
  • Specifications and UX for every module
  • AI workflow design: what the AI may answer, when it must hand off, what needs approval
  • Architecture choices for a multi-tenant platform
  • Directing AI coding agents (Claude Code, Hermes agents) to build from my specs; reviewing and releasing
  • Pricing, packaging, launch and customer onboarding

Who did what

Product, specs, AI design, GTM and onboarding were mine. Code was written by AI coding agents working from my specs, with my review and release decisions.

The solution — two products in one platform

AI Lead Manager

  • Unified lead capture
  • AI lead scoring
  • WhatsApp AI sales agent grounded in each customer's catalogue, prices and documents
  • AI follow-ups with human handoff
  • AI quotations
  • AI voice calls (India)

AI Marketing Manager

  • ICP data automation
  • Drip sequences on WhatsApp and email
  • AI Social Studio for on-brand posts and creatives

Results (as of Oct 28, 2026)

MetricValueSource
Paying businesses 176 in 7–8 months (latest 38 in 6 weeks) Company-reported
MRR ~$10.5K Company-reported
Renewal after first quarterly plan 73% Company-reported
Leads managed 173,713 Platform data
Messages handled 942,392 Platform data
AI replies sent 439,239 Platform data
Median AI reply time ~5 seconds (last 30 days) Platform data
AI replies outside business hours 33% (outside 9 am–8 pm IST) Platform data

Platform data = measured in the live product. Company-reported = customer count, MRR, renewal.

Lessons learned

Real customer language breaks “perfect” AI flows

A sales rep typed a product model number followed by “with 10% off”, and the AI read the model number as the quantity, producing a quote that was wildly wrong. We added rules that separate product codes, quantities and discounts, and send any unusual discount for approval.

Lesson: put clear guardrails around numbers and money; never leave them to the model alone.

Test on real customer data, not demo data

One customer’s website showed several products per page. Our catalogue reader expected one, and a photo filter dropped real product images because names like “silicone” contain the word “icon”. After fixing both, that customer went from 29 to 268 products in their AI catalogue.

Lesson: the product is only as good as its handling of messy real-world inputs.

Answer only what was asked

Buyers on WhatsApp want one clear answer, not a brochure. We tuned the AI to answer the exact question from the customer’s own data and keep the conversation context.

Lesson: shorter, grounded answers convert better than impressive ones.

Product walkthrough

Real product screens from live workspaces

Customer names, numbers, emails and websites are replaced with sample data.

AI Lead Manager

Unified lead inbox: every lead from every source in one place, so nothing gets lost.
Unified lead inboxevery lead from every source in one place, so nothing gets lost.
AI lead scoring and focus leads: the sales team sees which buyers to call today.
AI lead scoring and focus leadsthe sales team sees which buyers to call today.
AI sales agent conversation: instant answers from the company's own catalogue and prices, day or night, with the AI score, summary and next action beside the chat.
AI sales agent conversationinstant answers from the company's own catalogue and prices, day or night, with the AI score, summary and next action beside the chat.
Follow-ups and human handoff: every lead gets followed up based on its score; hot or unclear chats go to a person.
Follow-ups and human handoffevery lead gets followed up based on its score; hot or unclear chats go to a person.
AI agents working across channels: WhatsApp, voice, email and business insights, each with its own rules.
AI agents working across channelsWhatsApp, voice, email and business insights, each with its own rules.

AI Marketing Manager

ICP Data Automation: pick what you sell, who should buy it and where, to build a target list.
ICP Data Automationpick what you sell, who should buy it and where, to build a target list.
ICP Data Automation: shortlisted companies with verified contacts, ready for outreach.
ICP Data Automationshortlisted companies with verified contacts, ready for outreach.
AI Social Studio: on-brand posts and creatives, previewed for each platform before publishing.
AI Social Studioon-brand posts and creatives, previewed for each platform before publishing.
Growth and GTM case study

How Convogenie Lead Manager found its first 176 customers

  • ICP: Indian B2B manufacturers, distributors and industrial suppliers that already buy leads from marketplaces such as IndiaMART. Why: they feel the pain of lost leads every day and can see the value in their own numbers.
  • Acquisition: direct outreach. Business data filtered by category and location to build target lists.
  • Nurture: AI-run WhatsApp and email sequences working together: WhatsApp for quick, personal touches; email for detail.
  • Conversion: AI phone calls (India) and a guided onboarding that connects the customer's lead sources and catalogue so they see value fast.
  • Economics: 176 customers in 7–8 months; ~$60/month per customer; 73% renew after the first quarterly plan. (Company-reported, Oct 2026.)
  • Productisation: the same playbook (ICP data, drip sequences, AI calls) became the AI Marketing Manager inside the product, so customers can run it for their own growth.
Find
ICP data
→
Nurture
WhatsApp + email
→
Convert
AI call + onboarding
→
Retain
quarterly renewal
How I build

How I build reliable AI products

Lead sources
→
Lead ManagerFastAPI + PostgreSQL
→
AI agentsleading AI models from OpenAI, Google and Anthropic, grounded in customer data
→
WhatsApp / email / voiceCloud API
→
Message queueAWS SQS
→
DashboardReact
  • Grounded answers: the AI answers from each customer's own catalogue, price list and documents, not from general knowledge.
  • Human approval and handoff: hot or unclear chats go to a person; unusual quotation discounts need approval.
  • Retries and failure handling: failed message sends are retried automatically and tracked.
  • Multi-tenant isolation: every customer's data is separated, with role-based permissions.
  • Usage and cost tracking: AI usage and cost are logged per customer, so pricing stays profitable.
  • Safe releases: every release has a backup and health checks before it goes live.

How it was built: I write the specs; Claude Code and Hermes agents write the code; I review and decide what ships. This let a small team build and run a multi-module platform.

Ways to work together

Engagements

AI Product Discovery to Launch

Problem definition, product strategy, specs, AI build plan and launch.

Suits: a founder or team with an AI idea that needs to become a product.

AI Agent Implementation

Design and launch sales, support or marketing agents on your data, with handoff and approvals.

Suits: teams who want AI working in a live workflow.

AI GTM & Activation Sprint

ICP, positioning, outbound, onboarding and conversion.

Suits: products with users but weak growth.

Engagement: hourly, fixed-scope project or monthly retainer. Pricing discussed privately.

Background and contact

25+ years building B2B platforms and SME ecosystems at IndiaMART, Sulekha, AskLaila and Groweon. Co-founder of Convogenie AI.

Have an AI product or growth challenge?

I'm available for remote, fixed-scope projects and ongoing engagements, with 4–6 hours of overlap across US business hours (ET, CT, PT).

Book a 30-min call

vijayproject73@gmail.com · linkedin.com/in/vijay-convogenie-ai · +91 99720 25973