Everyone today will tell you
which tools to use.
I want to talk about why most of it won't matter.
unless you fix what's underneath first.
The premise this entire session is built on
Marketing is not a department.
It never was.
We invented the marketing department the same way we invented the HR department — to manage something that was never supposed to be managed. When marketing becomes a department, it means the rest of the company opted out. The product opted out. The ops team opted out. The founders opted out. And then we wonder why campaigns don't convert.
And B2B vs B2C is the wrong axis. The real axis in 2030 is H2H → A2H → A2A: human-to-human, agent-to-human, agent-to-agent. Every surface you ship is one of those three. Most brands only design for the first.
01 / 04 — Reframe
Every Touchpoint Is a Marketing Act.
The onboarding email. The API error message. The invoice PDF. The sales-call follow-up at 11pm. The founder's reply on LinkedIn. These are not ops. They are not support. Every ops surface is a marketing surface — and increasingly, the reader on the other side isn't human. H2H, A2H, A2A — all three at once.
→
H2H — Sales-call follow-up speed, founder reply on LinkedIn. Still the table stakes.
→
A2H — Onboarding email latency, in-product nudges, AI-drafted CS replies. Where most "AI marketing" lives today.
→
A2A — API docs, structured catalog, schema.org, machine-readable pricing. The reader is a buying agent, not a buyer.
→
The brands stuck only optimizing H2H are watching A2A quietly eat the funnel — and don't yet know it.
Blaze — How this shapes how we build
Every product we ship is designed for H2H, A2H and A2A from day one. The onboarding is the campaign. The first reply is the ad. The docs are the deck — and the API is the salesperson the agent will actually talk to.
02 / 04 — Reframe
The Best Brands Have No Marketing Department.
Apple. Patagonia. Zomato in its early years. None of them won on ad spend. They won because marketing wasn't a function — it was a founding decision about what the company would be and how every part of it would behave.
Product
Is the product itself the message?
Ops
Does how you deliver say something?
People
Do your hires embody the brand?
All of it
This is where marketing actually lives.
The Blaze way
36 products. Zero ad agency. Every rupee of revenue came from the product doing the talking — because we designed it that way from day one.
03 / 04 — Reframe
Agentic Brands Don't Run Campaigns. They Have a Nervous System.
A campaign has a start date and an end date. A brand has a nervous system — a continuous, consistent expression of what it believes, across every channel, every interaction, every moment. AI doesn't give you the nervous system. It amplifies the one you already have. If you don't have one — AI just makes the noise louder.
H2H loop — the old game
Founder-led signal, community, referrals, SDR outbound. Real, but it doesn't scale past the founder's calendar.
A2H loop — the current edge
Triggered product-signal email. Churn-risk intervention. Personalised onboarding. The agent works while you sleep — the customer still sees a human-shaped surface.
A2A loop — the next decade
Your docs, catalog, feeds, MCP endpoints designed to be read by buying agents. Structured intent surfaces. The customer never sees them — their agent does.
What AI actually changes
It doesn't just speed up H2H. It spawns A2H and A2A as net-new surfaces. Brands designing for all three compound. The rest just send louder emails.
04 / 04 — The Most Dangerous Reframe
The Most Dangerous Thing AI Does to Bad Marketing Is Make It Faster.
If your brand has no clear point of view — AI will help you produce more of nothing, at scale, faster than ever before. Most brands are using AI to do louder H2H. The unlock is designing for A2H and A2A from day one. The brands destroyed by AI aren't the ones that ignore it — they're the ones that adopt it without fixing the foundation first.
What AI accelerates — if your foundation is broken
Generic content. Inconsistent voice. Campaigns that look like every competitor. Faster race to the bottom.
What AI amplifies — if your foundation is solid
A clear POV expressed across every channel simultaneously. A brand that never sleeps. Compounding signal. Defensible identity.
What this philosophy produces — when you actually live it
7 Years. Zero Capital. 36 products still running.
36+
AI-native products live across 5 clusters
₹0
External capital. Customer-funded. Always.
90
Days. Hard break-even target on every build.
7
Years. Same philosophy. Tools changed. This didn't.
The tools were different in 2017, 2019, 2021, 2023, and 2025. The principle — that marketing is structural, not departmental — never changed. That's the only reason 36 products are still alive.
Case Study — IndiaMART
The Platform Is the Pitch.
IndiaMART didn't out-advertise Alibaba or Amazon Business in India. It out-systemed them — across all three layers. H2H: WhatsApp seller-buyer chats. A2H: a buy-lead engine matching 200M+ buyers across 8M+ sellers in 25 languages. A2A: a structured catalog and category schema that ChatGPT and buying agents are already crawling. The matchmaking engine IS the marketing department.
The question for the room
"Is your catalog ready for the agent on the other side?"
IndiaMART across the three layers
→H2H — WhatsApp seller-buyer chats. Human voice, human trust, last-mile close.
→A2H — The buy-lead engine: pre-sales for 8M sellers, 24/7, in 25 languages. No SDR army.
→A2A — Structured catalog and category schema already being read by ChatGPT and procurement agents. They didn't plan for it — they architected for it.
→The brands that win in 2030 won't have bigger sales teams. They'll have matchmaking compounding inside the product itself, designed for both humans and agents.
Blaze Live — CORTEX / RTO Sentinel
Returns Are Churn. Treat Them Like B2B Does.
Indian D2C loses 25–40% of revenue to RTO — orders that come back undelivered. Everyone treats it as logistics. It isn't. RTO is an A2H loop (an agent intervening with a human at the churn moment) feeding an A2A surface (structured returns data downstream agents can read — logistics, underwriting, account-segmentation). SaaS has done this with activation data for a decade. D2C still doesn't.
The reframe nobody makes
"Your churn data isn't just a dashboard. It's an A2A feed waiting to be exposed."
How RTO Sentinel works
01
Score the order
Address quality + order history + COD flag + behavioural signal. Risk score before dispatch.
02
Trigger WhatsApp confirmation
High-risk orders get a personalised confirm flow. Reduces returns before they happen.
03
Feed the marketing layer
The same risk signals map to retargeting suppressions, upsell triggers, lookalike exclusions. Ops data becomes marketing data.
04
Self-improving loop
Every confirmed/returned order retrains the model. Sharper each sprint.
Workflow — Live in Production · A2H today, A2A tomorrow
The SMB Outreach Agent Stack.
A pure A2H system today — no SDRs, no humans drafting. 5-stage loop, brief goes in, paying SMBs come out. Zero ad spend. Next version doesn't email a human — it talks to their procurement agent. That's the A2A jump.
01
Scrape
JustDial / Google Maps / public registries. 150+ Mumbai SMBs per sprint. Zero manual search.
02
Enrich
Business type, size, pain signal, contact. Auto-scored against ICP.
03
Personalise
Each message references their specific business — site, reviews, hiring signal. Not a template. Actual context. This is the unlock.
04
Deploy
WhatsApp-first. Sequenced follow-ups. Reply triggers the next action automatically.
05
Signal
Open, reply, convert. Weekly feedback into the scoring model. System improves.
Receipts150 paying SMBs · 30 days · ₹0 ad spend · 0 SDRs
The only thing I want you to take out of this room
If your marketing department
disappeared tomorrow —
would your customers notice? Or would the product, the experience, and the culture keep doing the work?
and in 2030 — would the agents notice?
Marketing 2030 — The only thing that changes everything
Build the Nervous System. Not the Department.
AI is not the answer. It's the amplifier. H2H is crowded. A2H is the current edge. A2A is the next decade. Systems that span all three compound while competitors sleep.
Ideas are cheap. Execution is fragile. Tools are transient. Systems are rare.
Blaze — blaze.bwe.one — 36 live ventures
Come find me after. Let's talk about your foundation.
