Saksham Chauhan

I build with AI to take real work off people.
Writing what I learn from the actual work, the wins and the misses.

[Saksham Chauhan @SakshamCh2302]

Hot take: most people calling themselves "Product Managers" in 2020 wouldn't survive the job today. Not because they got worse.

The job changed under them.

2020: PM meant specs and roadmaps. Execution over judgment.

2021: Remote work killed the hallway decision. If you couldn't write it clearly, it didn't happen.

2022: Metrics ate the room. No number, no decision.

2023: AI walked in quietly. The question stopped being "what do we build" and became "what do we even need to do ourselves"

2024: PM, design, and eng started blurring. Prototyping and prompting stopped being nice to have.

2025: Teams got lean. One PM doing what used to take three, because AI absorbed the grunt work.

2026: The PMs winning right now aren't the roadmap managers. They're the ones who can read a system and know exactly where AI still needs a human.

Same job title. Completely different job.

The ones who kept up were never managing outputs. They were reasoning about systems the whole time.

[Saksham Chauhan @SakshamCh2302]

Looking for recommendations on zero forex credit cards.

I already have @getscapia, but want a solid backup.

What would you recommend?

[Saksham Chauhan @SakshamCh2302]

Most of the time, the tool is not the problem.

I have watched teams buy new systems, set them up carefully, and still end up with the same work patterns two weeks later.

The software sits there. The habits do not move. Everyone quietly accepts it.

Buying something feels like progress. Changing how people actually work does not.

That second part is harder, slower, and easier to avoid. So it gets skipped.

I used to think the right tool would force the change. It does not. It just makes the lack of change more visible.

I wrote a longer piece on the difference between buying a tool and changing how the team works.

https://t.co/fA7YuMDbdI

Curious if others have seen the same pattern.

[Saksham Chauhan @SakshamCh2302]

Sales teams lose more deals in one moment than any other.

The buyer brings up a competitor, and the rep says:
"Let me get back to you."
The information was already in the CRM. The battlecard already existed. The same objection had already appeared in last quarter’s calls.

What didn’t exist was readiness when the deal needed it.

We built https://t.co/PsOEIei01j to fix this.

It detects which competitors are active in your deals, turns that intelligence into deal-specific briefs, lets reps rehearse the exact objections they’re about to face, and gives them the right response in real time during the call.

No more guessing. No more preventable hesitation.

Just reps who walk into competitive conversations already knowing what they’re going to say.

https://t.co/PsOEIei01j

[Saksham Chauhan @SakshamCh2302]

I automated a mess once.

Two small clients, a construction firm and an interior design studio, came to me for marketing. The real problem sat one layer under the content calendar: nobody could see what was actually happening across the business. So I built a dashboard.

For the first two weeks, it made things worse. I'd digitised the chaos, not removed it.

Klarna learned the same lesson at a much bigger scale. Cut nearly 700 support jobs to AI, called it a win, then spent 2025 quietly rehiring once service quality on the hard cases started costing more than the automation saved.

Different scale. Same mistake. The tool worked fine. The workflow underneath it didn't.

I wrote the full breakdown, the UPS and IBM numbers, why strategy beats tools five to one, and the Klarna story in full: https://t.co/ZM01B7Cbui

[Saksham Chauhan @SakshamCh2302]

A lot of companies are using AI now. The part that sticks with me is that 81% of them are still seeing basically no real impact on the bottom line.

The tools are working. People are genuinely getting hours back. But in most organisations, that time just disappears. It gets filled with more meetings and whatever else was already there. Nobody really planned what to do with the capacity.

It's not a tech problem at this point.

The question I keep seeing leaders ask is still “how do we adopt more AI?”

The ones actually getting somewhere are asking something different.

What would our work actually look like if we designed it around what AI can do right now, instead of just layering it on top of the old way of doing things?

That’s the real gap. Most places haven’t crossed it yet.

[Saksham Chauhan @SakshamCh2302]

You used AI this morning and saved 3 hours.
You filled those 3 hours with the same meetings as yesterday.
Nothing changed. Because nothing was designed to change.

This is the quiet failure nobody is talking about. Employees are saving real time. Entire workdays are recovered every week. But that capacity has nowhere to go inside an organisation that was built before AI existed.

The time does not disappear. It flows back into approvals, updates, and work that fills space rather than creates value.
The tools are doing their job. The organisations around them are not.

There is a name for this ceiling. And a way through it.
https://t.co/9H6pcI2Wpf

#digitaltransformation #aiautomation #shadowAI #AIStrategy #autonomousagents

[Saksham Chauhan @SakshamCh2302]

@dimitarangg CLAUDE

[Rouze @Sal_Leone]

The $300/day agent problem does not seem to be an AI problem, it's an engineering problem. Agents that "waste tokens constantly" are poorly designed. The companies burning cash on runaway API bills skipped the step where you actually optimize your systems. That's like blaming electricity costs because you left every light in the building on 24/7

[Rouze @Sal_Leone]

It's now front and center - the next decade belongs to whoever can turn raw AI capability into real business outcomes for the 33 million small and mid-sized companies that will never hire an in-house AI team.

These businesses already have the data. They already have the problems. What they've been missing is a bridge - something that sits on top of their existing tools, understands their messy workflows, and quietly automates the unscalable work running in the background.

That bridge is here. AI agents that plug into everyday systems - email, CRM, calendars, ATS, billing, support - learn how a business actually operates, and orchestrate work like a full-time, cross-functional teammate. No forcing companies to bend to rigid SaaS. The agent molds around the business in real time: prospecting, following up, qualifying, updating records, drafting outreach, coordinating handoffs. Zero changes to the underlying stack.

This is Service-as-Software: customizable intelligence sold as a service, not a dashboard. The winners won't just be the people who understand models. They'll be the integrators who deploy agents into non-tech companies and make them money in week one - no re-platforming, no "become an AI company" pitch.

The old SaaS playbook is running out of runway. This is what replaces it. SaaS is done ...

[Saksham Chauhan @SakshamCh2302]

@reachvaldo Art

[Saksham Chauhan @SakshamCh2302]

@jspujji GA

[Jesse Pujji @jspujji]

I hired an ex McKinsey consultant to compile all my sales materials to document how GrowthAssistant company reached $22M in ARR.

He collected:

- Recordings of sales calls
- Sales scripts
- SOPs
- Lead gen systems
- etc

100s of top companies paid me for access to it.

Today I'll give it away for free.

RT + reply "GA" to get a copy in DMs.

[Saksham Chauhan @SakshamCh2302]

Product Marketing Strongest Storyteller

[SocialRipple.ai @SocialRippleai]

Product Marketing Strongest Storyteller

[Saksham Chauhan @SakshamCh2302]

How often are you rotating your API keys, and do you have automated alerts for unusual usage patterns?

#OpenAI #securitybreach #artificialntelligence #buildinpublic

[Saksham Chauhan @SakshamCh2302]

If you're building with AI APIs, last Friday's #OpenAI breach should make you check your security setup. Right now.

Attacker compromised an employee account, moved laterally, exposed ~10 accounts' API keys and conversation histories. Contained within hours.

Three lessons🧵

[Saksham Chauhan @SakshamCh2302]

If you're managing products with AI integrations: audit your access controls this week, review incident response plans, ensure your team understands the attack surfaces