AI Spotlight

The businessperson turned AI engineer

A go-to-market advisor on why the advantage isn’t knowing the “right” AI tool. It’s learning to think like a systems person close to the work.

Meet Nick Bhavsar

AI is moving at a speed that makes even the most plugged-in operators feel behind.

Nick Bhavsar isn’t pretending he’s above that. In our conversation, he said it plainly: it’s impossible to keep up with everything that changes week to week.

But his approach to staying current is different from most.

Instead of collecting tools and hot takes, Nick spends his time sitting with teams, watching how work actually gets done, and then using AI to compress the feedback loop between “this is the problem” and “this is a working solution.”

He previously founded Velocity Engine, an early MarTech / AI go-to-market company, and today he’s consulting with B2B teams on AI-powered GTM: upskilling, use-case discovery, and building practical implementations that can survive contact with the real world.

What I liked most about Nick’s lens is that it refuses to romanticize AI. It treats it like a new operating capability. Useful only when it plugs into actual workflows.

“You don’t need to know how to code. You just need to know how to scope a problem really well, put something up, see if it works, and give it feedback.”

- Nick Bhavsar, Principal @ Bhavsar Growth Consulting

The real shift: business people can now build for themselves

Nick shared an example that would have been a full internal project a year ago.

One of his clients is in financial services. Their team gets leads from a regulatory database (the SEC), but the workflow to find the right prospects was painfully manual: go to a site, run a search, apply filters, open PDFs, look for specific signals, keep the good ones, discard the rest.

Nick sat with a sales rep and documented the work step-by-step. Then he used Claude Code to replicate the workflow.

The output was not a “demo.” It was a prioritized list of prospects the team can generate on demand.

The detail I keep thinking about is that Nick doesn’t present himself as a developer. He said the last time he coded before this was in college, decades ago.

What changed is not that every GTM operator suddenly became an engineer. It’s that the distance between “I understand the problem” and “I can ship a solution” just collapsed.

Nick framed it as a systems problem, not a tooling problem:

In the old world, the loop was: business user → product manager → engineer → back again. Slow. Expensive. Easy to lose intent.

In the new world, if the business user can think clearly and iterate, they can build something functional and refine it in near real time.

Just-in-time code beats permanent systems

Nick also had a useful counterweight to the idea that you’re building permanent assets.

A lot of what teams build with AI right now is closer to “just-in-time code” than it is to product.

You solve a real business problem quickly. It works. And then the landscape changes again.

Tools evolve, models improve, and sometimes the “right” implementation six months from now is just different.

That’s not a failure. It’s the trade: speed and adaptability over polished permanence.

Where AI gets really interesting: when it starts buying

Most of the market conversation is about selling: write better outbound, generate more content, improve targeting, raise conversion.

Nick’s more interesting question is what happens when AI starts buying.

Not in the micro sense (where AI already bids in auctions, like paid social). In the macro sense: monitoring a business, noticing a problem, and initiating procurement behavior.

He described a future where an AI monitors a company’s P&L, identifies a shift, sends RFPs to vendors, evaluates responses, and brings back recommendations.

That’s not tomorrow. But it’s not science fiction either.

And once that starts happening, the “how do I get my message into the LLM” conversation becomes table stakes.

The hot take: SaaS pricing won’t stay the way it is

Nick’s clearest conviction was about business models. He doesn’t think the current SaaS pricing model survives AI. His expectation is a move toward outcome-based pricing: pay when you get paid.

He pointed to Intercom’s model as an early indicator (pay per resolved ticket instead of a big flat software bill), and said we’ll see more vendors try to tie pricing to attributable results.

It’s easy to dismiss this as “obvious.” The hard part is making attribution real enough to price it.

That’s why it’s worth watching: the vendors who can do it will be interesting businesses.

“Everything right now is focused on the selling side of AI… I’m curious what it looks like when the AI starts buying.”

- Nick Bhavsar, Principal @ Bhavsar Growth Consulting

A practical place to start: stop training everyone the same way

Nick’s advice for teams trying to roll out AI internally was not “pick a tool.” It was “organize for maturity.”

He described cohorts moving through a maturity curve:

Cohort 1: chat-based usage (basic prompting and day-to-day assistance)
Cohort 2: advanced tools (agents / coworker-style workflows)
Cohort 3: building (code, automations, apps)

His warning is simple: if you throw cohort 1 and cohort 3 into the same training environment, everyone loses.

The advanced folks get frustrated by how slow it feels.

The beginners get overwhelmed and shut down.

Treat it like adult education. Create small pockets where people can try things, break things, and share what works.

Then move the maturity line forward on purpose.