Aakash (@aakash558) header
Aakash (@aakash558) profile picture

GTM & AI Operator B2B Saas, GTM, StartUps, Demand Gen Ex Director/Head Marketing - https://teapawt.net/t.co/9aO5hpLdWZ, @SpotdraftHQ | ex @join_ef, @gocashfree, @Belongco

Bengaluru, India
Joined November 2011
Photo by Aakash (@aakash558)Photo by Aakash (@aakash558)Photo by Aakash (@aakash558)Photo by Aakash (@aakash558)Photo by Aakash (@aakash558)Photo by Aakash (@aakash558)Photo by Aakash (@aakash558)Photo by Aakash (@aakash558)Photo by Aakash (@aakash558)Photo by Aakash (@aakash558)
A few months back I started the GTM & AI Operators Show with one simple idea. AI is changing how we build, sell and service customers, and I wanted to understand it properly. This week it ran in the press. Not very big names, but something to cherish and be thankful for. Thank you to the guests who said yes when there was barely anything to say yes to. We are just getting started. youtube.com/@GTMAIOperators
1
93
I asked Kushagra Tiwaria for one line of advice per GTM function. No explanations allowed. As I expected, he didn’t hold back. Content Writer: "Don't use AI." Mid-market CEO: "Stop pivoting." Most operators would round these off. Add a "depends on your stage," a "generally speaking." He didn't hedge, and that's exactly why they landed. There are more questions in this rapid fire, including his two-word diagnosis for founders who think AI can do everything. I wear a few of these hats myself including founder, writer, operator, so a couple of these hit closer to home than I'd like to admit. KT, I'd like to state for the record that I use AI "responsibly." 😅 Full clip below if you're ready for some truth bombs.
1
3
58
Posted about an outbound email playbook last week on LinkedIn. Genuinely thought it would get a few likes and disappear. Got a bunch of comments and DMs instead. Here’s a gist of everything I keep relearning as I run this: - Email by itself is a cold ask from a stranger. Nobody owes a stranger a reply. - Email plus a webinar with 2 to 3 external speakers: 50 to 100 registrations on the first one, from people who had never engaged with us before. - Email plus attending a conference: a VP who wasn't even attending replied anyway, asking for reference customers. The conference was the excuse. The reply was the win. - Email plus a small offline event: 25-35 people at a standalone mixer. A Head of Infra at a global bank replied "please count me in" the same day. No pitch anywhere in that email. So email was never the channel doing the convincing. It's just where the reply lands after other channels did the task of authority and brand building. I put it all into one playbook. One webinar, one conference, one offline event, LinkedIn running alongside, email underneath all of it. 25 pages, 12 templates, under $1K a month in tooling. One rule in there I'd flag upfront: don't let AI write your first batch of emails. Write them yourself. Watch what actually gets a reply. Then teach the AI to sound like that. Comment and I'll send it over.
76
I was watching Kushagra Tiwari’s journey last year and cheering. It was a proud moment for all of us when he got to the Clay World Championship - Top 4. In between all of that, he kept helping me here and there, whenever I asked for help. When I started the podcast, he was on my list from Day 1. So, happy we could do this earlier this month. I asked him bluntly whether the GTME survives to 2027. His answer: the role started outbound. It's now spreading into inbound routing, scoring, events, even PLG. It's not dissolving. It's finding its place inside more functions. Episode 06 is out on the GTM & AI Operators Show. Below is a small byte and Link to the full video is in the first comment. PS - This episode is 90 mins long, post editing and has a rapid fire at the end as well.
1
1
36
Kimi K3 was all over the news last week, so I decided to test the platform. My plan: Start with the free model on a basic data classification task, then upgrade to K3 for heavy lifting. The result? Stuck in an infinite loop 🌀 Disappointing given the Kimi Claw hype. Anyone in GTM/RevOps using K3? Is it worth upgrading?
98
The Man, the Myth, the Legend. Kushagra Tiwari - Top 4 Clay World Cup | GTM Engineering Coach @ Clay Bootcamp | Founder, RevenueGenie; and currently Principal GTM Engineer at Starbridge. Just spend the entire day with him recording the next episode on How does GTM and the GTM Engineer look in 2027? KT is one of the pioneers of the GTM Engineering ecosystem globally. This is by far one of the best podcast conversations I had (obviously knowing him from before had a certain comfort level). The depth of stuff we discussed is way beyond what is usually shared on LinkedIn or Youtube. Thank you KT for contributing to the narrative, prep calls, driving down and doing this. It really means a lot. It’s gonna be Legen... wait for it... dary. PS. So many off-camera conversations that will shape my next few months in ways I did not think earlier.
43
Your buyer decided before they got on the call with you. Two operators told me this in two separate conversations. Same diagnosis. Different fix. Mason Cosby came at it from marketing. Brand recall takes three to six months. So you plan for three to six months of intentional, targeted messages. Not a campaign. A presence. You shape how they think about the problem long before anyone gets on a call. Give the playbook away publicly. The only thing left to sell is having you execute it for them. Rahul Wadhwa came at it from sales. By the time the buyer is on the call, they have heard every pitch. They spent a weekend trying to solve it themselves. His line stuck with me - your buyer has evolved more than you have. So the only currency left is a point of view they haven't heard. Show up with a script and you're telemarketing. One shapes the decision before the call. The other has one move left once the call starts. The work that decides the deal happens where you can't see the buyer. Before the call. Off the call. Mason's clip is below. Here's the link to the conversation with Rahul. Scroll to [1:00:20](youtube.com/watch?v=aVDrb7jN…) Your B2B buyer has evolved more than you have.
31
Delayed gratification. Everyone talks about it. Almost nobody does it consistently. I'm in that group too. I believe in it and I try to practise it. But some days and weeks, you just get impatient. You try to fix it by overworking. Or by "taking a break". Neither works. The thoughts stay sitting in the back of your head. Maybe as a system we just don't glorify the boring part. Showing up and doing the work day after day with no appreciation and no immediate outcome. Everything has to be a 0 or a 1. Most of my days are neither. I think 90% of them won't be. But I hope I'll still show up, keep a clear head, and do the work whether or not it pays off that day or that week. Happy Friday everyone. PS - I took this photo in Varkala back in 2022 and thought, this'll make a good LinkedIn post one day with some deep insightful copy.
18
The video on building a Claude skill is out. I build one live, end to end. A copywriting skill trained on my own posts, but the process works for anything. The part most people miss: the skill is not done when you save it. For the first 10 outputs, edit what it gives you and feed the edits back so it updates itself. That loop is what makes it yours. youtu.be/BTDIrfDEuuk?si=Bk_1… via @YouTube
32
I've started to develop an aversion to AI generated content and design. I don't fully know why yet. Maybe because it's everywhere in my feed. Maybe because it tells me the person behind it didn't want to put in the effort. Maybe because it tells me I'm not engaging with a human anymore. The same sentiment keeps coming up on my calls - phone, zoom, in person. I'm watching market leaders again changing their design systems. In-person events are doing well again. Was this just another hype we got tired of too fast? Or does AI actually get to a point where you can't tell it apart from a human. Or maybe we just give in and accept it, the way we all made peace with "how did you get my number." My personal take is that AI systems are really good for efficiency at scale but the last touch point - the human <> human part needs to feel like that. Could be I'm looking at this too narrowly. Curious where everyone else is landing on it.
29
No MCP, no product. It's that simple now. I run everything through my Claude MCP. The people I work with ask bluntly - can't this be done via AI using MCP? MCP isn't just an orchestration layer anymore - the role Clay played for GTM in 2025. It has become the interface for using software. I can't comment with authority on the future of headless software. But one thing I know as a user: I prefer tools that have an MCP server. From the builder side, a lot is talked about subscription pricing being dead but I think if you don't ship an MCP server, you're missing customers. Nitya, founder of HeyAmos (heyamos.com/), summed it up with the oldest line in the book - build for where your customers are. Right now, almost all of them are at an MCP.
48
Sometimes, going off-script is good. In my conversation with Rahul, after a point, I left the questions and just continued talking like we were having a coffee or a beer in a typical Bengaluru setting. These conversations don't just teach me things. They shape how I think as an operator. Rahul is an execution first operator. He did the whole nine yards of the startup hustle, found his 'right to win' and built successful businesses. In a not so formal way, we still covered a lot. → Why every B2B company is quietly becoming a service company. → The 5% rule, that fewer than five people he knows have built anything worth selling with AI. → Why the best of Claude is still a 7 out of 10, and your own judgment is the multiplier. → The Nevara outbound playbook, coffee bags on bicycles and office drop-ins across two cities, pulling 30%+ reply rates. What we kept coming back to was much simpler. AI didn't hand anyone an edge. It exposed who already had one. The episode is live. If you're building and selling in the AI world right now, this one's worth your time. youtu.be/aVDrb7jNyCE?si=ZhIC… via @YouTube
1
78
Still small, still figuring it out. But people are showing up, and that means more than the badge. Thank you if you're one of them. 🙏
1
6
I recorded a session yesterday morning, building a Claude skill from scratch. It started as something for the YouTube channel. By the end it was more than that. It was just me sharing what I've been figuring out. It was genuinely fun. At one point I was talking to the AI. At one point I was talking to whoever ends up watching. I was learning the whole way through. There were also long silences where the AI was working and I was just sitting there. Somewhere in the middle it clicked: this will help some people learn a few things. But it will also put me in front of people who are way ahead of me, and they'll tell me where I'm wrong. A whole Sunday morning and afternoon went into it. I'll edit it now and share it with the wider group. Tried something new today. Really enjoyed it. Thank you to everyone who's been backing me so far. Would love your feedback once the video is out.
53
Anthropic shipped Fable 5 as the most capable model they'd ever put in public hands. Three days later it was gone. For everyone. I lost access too, sitting in Bengaluru. A US export control now blocks every foreign national, inside the country and outside it. Two weeks ago I wrote about this exact split. Frontier AI breaking in two, a public tier and a private one, with relationships and clearances deciding capability, not pricing pages. I think I got the split right and underestimated it. The public tier isn't a tier you buy into. It's a permission someone can pull. Not an outage. Not a price hike. A Commerce Department letter, over the vendor's objection. Any GTM motion built on a single frontier model is a geopolitical dependency now, not just a vendor dependency. The kill switch was never on the pricing page. And it isn't held by anyone you can call.
83
I am genuinely impressed by the shipping speed of @AnthropicAI. Ladies and Gentlemen, we were not ready for this 😅 I was just getting my head around 4.8. Now, we have Fable 5. Amazing time to be alive and excited for AI. 🙌
15
Everyone is talking about agentic SDRs. The model. The personalization. The orchestration. The bottleneck isn't the agent. It's one layer below. If 30% of your B2B list is catch-all and your validator is throwing those away as "risky", your agent isn't reaching a third of your TAM. If your validator is letting them through without real verification, your agent is drafting to dead inboxes. Either way you're scaling errors, not output. Came across OrbiSearch. Built by Alex Loughlin, who shut down his outbound agency to build a validator that resolves catch-alls instead of discarding them. I tested it on a small sample of contacts a client's existing tool had marked as catch-all and tossed. Many came back as valid. The sample was too small to put a number on it, but the delta was real. Three things stood out: → Catch-all resolution. Most validators flag these as risky and tell you to discard. OrbiSearch verifies them. → Dual-layer verification. SMTP check plus proprietary mailbox detection. No test emails, no 72 hour wait. → Real-time lookup, not stale database. Apollo and ZoomInfo serve from cached indexes. OrbiSearch verifies live, the moment you ask. If you're scaling outbound through agents, the layer underneath matters more than the model on top. 500 free validations and 100 free lookups on signup, no card needed. Ping me if you want more to properly test it. Link 👇 naano.xyz/c/aakash-sinha-orb…
36