Manthan Tyagi

Growth Consultant @velcor.ai | Helping B2B teams automate sales, outreach, and lead generation

[Manthan Tyagi @manthan2tyagi]

The impressive part is the price to capability ratio.

If Gemini 3.8 Flash can approach larger frontier models on long horizon engineering while staying at Flash pricing, the economics of running coding and GTM agents at scale change dramatically. The model race is increasingly becoming a cost per successful task race.

[Manthan Tyagi @manthan2tyagi]

@julianlehr This is where the agent architecture gets really interesting.

The killer feature may not be individual bots, but a universal GTM inbox that understands the intent behind an input and routes it to the right workflow automatically. Less tool selection, more outcomes.

[Manthan Tyagi @manthan2tyagi]

The most important part is the willingness to slow down when capability outpaces understanding.

The real benchmark for frontier AI will not be how quickly models improve, but whether safety, alignment and governance can keep pace with that improvement. Exciting progress and caution can coexist.

[Manthan Tyagi @manthan2tyagi]

grok bot can now research accounts overnight

that is useful

the leak is still the same

lead comes in
nobody follows up
crm is stale
nobody knows which message booked the call

so teams add another bot

more research
more drafts
more sequences

volume does not fix leaks

the boring version wins:

clear icp
clean data
fast routing
short outreach
approval before send
track booked calls

then grok bot is a gtm tool

not another inbox to manage

[Manthan Tyagi @manthan2tyagi]

The interesting part is that marketing engineers are not just automating marketing tasks. They are building the feedback loops that make growth compound.

The person who can connect customer data → signals → execution → measurement → iteration across the entire GTM system will be incredibly valuable. That is a very different skill set from traditional marketing.

[Manthan Tyagi @manthan2tyagi]

The bigger issue is that companies are hiring for a title instead of an outcome.

A strong GTM Engineer should be able to identify the highest leverage workflow, build it, measure the impact, and make the team self sufficient. That is very different from hiring someone to simply “manage AI tools.”

[Manthan Tyagi @manthan2tyagi]

@Voxyz_ai The impressive part is not having 7 bots. It is the feedback loop. Research → filtering → creation → distribution → performance data → better research.That turns content from a recurring task into an actual system that gets smarter over time.

[Manthan Tyagi @manthan2tyagi]

The interesting shift is that AI Operating Partners are becoming less about “AI strategy” and more about proving operating leverage.

A 30 day window to validate one workflow with measurable ROI is far more useful than another portfolio wide AI roadmap. Once something works, then you build the operating model around it.

[Manthan Tyagi @manthan2tyagi]

The pay per lead model is interesting, especially for teams that do not want another monthly SaaS commitment.

The bigger differentiator is the open source approach. If the workflow is genuinely flexible across channels, it could make outbound infrastructure much more accessible to smaller GTM teams.

[Manthan Tyagi @manthan2tyagi]

The biggest mistake is treating B2B data as a single database problem.

The best stacks are increasingly waterfall based: use multiple sources, verify aggressively, and let each provider fill the gaps the previous one misses. Data quality comes from the system, not just the vendor.

[Manthan Tyagi @manthan2tyagi]

@romanbuildsaas Bot

[Manthan Tyagi @manthan2tyagi]

This is where AI starts feeling less like a chatbot and more like a programmable interface layer.The interesting part is not just summarizing HN. It is being able to reshape any website around the way you actually want to consume information, without needing to build a full extension yourself.

[Manthan Tyagi @manthan2tyagi]

@MaxAIoutreach The real breakthrough is not AI sending more outreach.

It is AI understanding the conversation well enough to know when to ask, when to qualify, when to educate and when to stop. That shift from automation to judgment is what could actually change outbound.

[Manthan Tyagi @manthan2tyagi]

This is a fascinating gap between AI capability and human behavior.

AI has removed a lot of the work around selling, but it has not automatically increased the quality of the seller. The next leap may come from changing how AEs spend the time AI gives back, not from adding more AI tools.

[Manthan Tyagi @manthan2tyagi]

@BrianLaManna_ This is what real sales enablement looks like.

You cannot coach reps effectively from a playbook you have never tested yourself. Leaders who get on the phone, hear objections firsthand and prove the messaging in the field earn credibility that no training deck can create.

[Manthan Tyagi @manthan2tyagi]

@MaxAIoutreach The real milestone is not that AI can send LinkedIn messages.

It is that agents can now interpret intent, adapt the conversation and qualify interest instead of blindly following a sequence. That is when AI SDRs start becoming actual sales infrastructure.

[Manthan Tyagi @manthan2tyagi]

@FidelCacheFlow @ChatAE_ai The best AI SDRs will not win because they can send more emails.

They’ll win because the people building them understand the psychology, workflow and edge cases of selling. AI should amplify great sales judgment, not replace it.

[Manthan Tyagi @manthan2tyagi]

@jasonlk This is the real test of the headless CRM thesis.

[Manthan Tyagi @manthan2tyagi]

@chrisbeals Exactly. The next generation of consulting firms will not just implement software.

[Manthan Tyagi @manthan2tyagi]

@partnerforce The consulting opportunity here is massive.

As agents become part of core enterprise workflows, implementation alone becomes less valuable. The real value shifts toward designing skills, architecture, governance and trust around how those agents actually operate.