I spent over a decade using SaaS tools to close B2B revenue. Now I use AI agent swarms to do it. 10+ years carrying quota and building pipe, rebuilding it all on an AI-native operating layer now.
Career B2B SaaS seller who now designs the agents and workflows that run revenue and GTM. Both sides of the table: carried the number, then built the system that helps others carry it.
I'm Agnishwar. Most people call me AB.
I've spent over a decade on both sides of B2B SaaS revenue, vendor and user. I closed $5M+ across 100+ enterprise deals, selling compliance and legaltech SaaS to CXOs and the kind of multi-stakeholder buying committees that treat a signature like a hostage negotiation, at companies like Unilever, Siemens, Tata Group, Mahindra Group, Vodafone, P&G, L&T, and Brookfield. Then I built the GTM engine from the inside: product positioning, competitive battlecards, and enablement that lifted competitive win rate by 40% and supported $5M+ in new ARR at CleverTap. Very few people have sat on both sides of that table, carrying the number and building the system that helps others carry it.
Over the past six to seven months I've been rebuilding how I operate, using AI as a core operating layer for revenue and GTM work rather than a productivity add-on you bolt on and forget. That has meant going deep on agentic workflow design, prompt architecture, and tool orchestration across Claude (Projects and Cowork), ChatGPT, and Gemini, instead of staying at the level of politely asking a chatbot for help.
The point isn't the tooling. It's the combination: quota-carrying sales instinct, GTM systems fluency built from the inside, and AI-native build capability most operators in sales or marketing simply don't have. It's the difference between someone who can describe what AI should do for a revenue team and someone who has built it, shipped it, and measured what it changed.
Most operators have one of the three. A few have two. The edge is having all three pointed at the same problem.
Everyone in revenue now has an opinion about AI. Far fewer have shipped an agent that survived contact with a real pipeline. I'd rather show you a workflow that saved an AE 70% of their prospecting time than a slide that promises it will.
So the rest of this page is mostly things I've built and what they changed. Read the numbers with the healthy skepticism they deserve; then ask me to walk you through any one of them live.
See what I've builtA sample of the functional agents and workflows currently in active use. Figures are as measured; treat them as directional and ask me to show my work.
A Hermes-based agent that manages a production VPS entirely through Telegram. Deploy sites, monitor health, manage DNS. No terminal needed.
Idea to live site in 10 min View buildTurns a PRD into a full GTM launch package: positioning, messaging, and enablement collateral, generated end to end.
100–120 min vs. 2–3 days View buildDecodes buyer intent signals, product gaps, and competitive pressure from call transcripts. Not a note-taker — a signal-intelligence layer that feeds a continuous PMM and sales improvement loop. 100+ calls decoded.
100+ calls decoded View buildDiagnostic study that identifies where pipeline leaks across ICP, outbound, CRM, messaging, and sales process, with a detailed roadmap to improve qualified meetings.
Pipeline leak diagnosisEnd-to-end product listing: visuals, copy, optimization, and loading, run as a single automated flow.
~75% time saved on bulk uploads View buildInvestment-bank stock research, built for the retail investor. Runs 40+ quality and valuation checks, a discounted cash-flow model, and a bear case — then grades every parameter behind the verdict.
40+ graded checks per stock View buildConsolidates company research, executive priorities, industry trends, and recommendations into concise, ready-to-use briefs.
Prep −70% · conversion +50% View buildAutomates account research, pain-point analysis, personalized outreach, discovery planning, and objection handling.
~70% time saved per lead View buildA track record across legaltech, GRC, and customer engagement SaaS, from full-cycle quota-carrying sales to enterprise GTM.
Published work on cyber risk, compliance, and AI-driven customer engagement, written for MetricStream and CleverTap.
What the SEC's finalized cybersecurity disclosure requirements actually demand of public companies, and where most compliance teams are still behind.
ReadA breakdown of what changed in the 2.0 update to the NIST Cybersecurity Framework, and what it means for risk and compliance programs.
ReadOn pulling Snowflake, BigQuery, and Redshift data into live campaigns in real time, so AI agents act on current customer data instead of stale exports.
ReadOn adding up the subscriptions, migrating to a self-hosted VPS stack, and why the case for owning your infrastructure is practical, not ideological.
ReadMore writing → MetricStream | CleverTap
Blunt, mildly quirky, and allergic to taking myself too seriously even when the work is serious. I compulsively check the wifi wherever I land (occupational hazard of running your own stack) and have been known to hug a tree mid-hike without warning. I'm the friend people call for decisions, trip planning, career pivots, the "should I actually do this" ones, because I'll bring the same evidence-gathering instinct to a friend's dilemma as to a client's deal.
Off the clock: chess, space, photography, film, an unreasonably serious coffee ritual, music, and travel built around character and walkability over luxury. I also run a fully self-hosted digital stack and a podcast, AyeBeeTalks. I like owning my infrastructure, literally and professionally.
Always glad to talk revenue, GTM, and AI-native systems. If you're figuring out where agents actually belong in a revenue org, or you just want to compare notes, reach out.