H R Srinivas

Insurance Professional, AI practitioner. Automation, intelligent workflows. Small business. Enterprise capability. The gap just disappeared. Let's talk.

[Meg @MegOfficiall]

His assistant, Lorraine, had been writing them. All of them. For ten years.

400 cards a year. In her handwriting, with his signature copied underneath.

She kept a notebook on every employee. Their kids. Their pets. Their mum's knee.

The CEO had once said, "Just send something from me." She took it seriously.

She was terrified she'd be in trouble.

He wasn't angry. He sat down next to her desk and asked to see the notebook.

It was 600 pages.

He said, "You've been me for ten years. And you've been better at it."

He gave her a $10,000 raise.

Then he asked her to teach him.

This year, the cards are in two handwritings.

His is worse. People keep them anyway.

[H R Srinivas @HRSrinivas]

Before AI, here's what happened to ideas - process improvement, system development, process automation, you name it - any idea.

You took the idea. You sold it up the chain. Someone above you sold it further. Eventually a yes came back — go figure out if it can be built.

You went to IT with the Yes. Who told you why it couldn't be done: Another system already doing something similar. Data privacy. Competing priorities. No bandwidth until Q3 blah blah blah. You get the drift.

If you got past that gate, you needed money. If you got approval for the money, you started building with the following gate keepers: Business analyst. Solution architect. Developers. Tester. Security review. UAT.

And somewhere in the middle of all this, scope crept. Because by the time anyone saw what it actually looked like, everyone had one more thing to add.

Then one of scores of things happened.

Business priorities changed. Or a regulation shifted. Or something else that made the whole thing redundant, less attractive, not worth the time/effort etc — taking down months of effort with it.

Net result nobody wanted to experiment with an IDEA that could end up in smoke.

If none of that happened and it actually shipped? You had to convince people to use it and that’s a whole other story.

---

Pre AI that was the pipeline. For one idea.

---

Cut to now.

You're sitting at your desk — maybe literally on the pot. Idea drops. You open a terminal. Spec it out in a conversation. Get a rough outline. Spin up an HTML prototype for the look and feel. Build a scope document. Go back to the terminal.

Build the thing. In a weekend.

Deploy it. Run it. And then — this is the part nobody talks about — *look at it*.

Really look at it.

And either you realise it doesn't look all that great when it comes off the assembly line. So you the originator kill it. Yourself. No committee. No post-mortem. No sunk cost spiral.

Or you realise you're onto something. And *now* the conversation starts — but this time you're not selling a concept. You're showing a working thing.

---

10 ideas. 10 prototypes. 48 hours each.

9 die. Not on paper. Not in a committee. After you built them and saw them and decided they weren't worth your time.

1 makes the cut.

That 1 arrives at every meeting already real.

---

Now multiply this by every empowered employee in your organisation who has an idea and a weekend.

That's not a productivity gain.

That's an innovation pipeline that has never existed before at this cost, at this speed, and at this scale.

**The only thing standing between that and your organisation is whether the gatekeepers allow it.**

The bottleneck didn't disappear. It moved.

It's now YOU.

@boardyai

[H R Srinivas @HRSrinivas]

The "AI causes cognitive decline" argument assumes everyone thinks the same way.

It doesn't.

The person who outsources thinking to AI and accepts whatever comes back? That person was already copy-pasting their way through work. AI didn't create that habit. It just gave it a new interface.

But there's another person the decline crowd never models.

The one who thinks faster than they can act. Who has five ideas before finishing one task. Who was told — repeatedly — that the problem is *focus*. Do one thing. Finish it. Move on.

That person didn't have a thinking problem. They had an *execution bottleneck*.

Pre-AI, the only way to operate was to slow down. Swallow the frustration. Watch half the ideas die in the queue.

Now?

Five tabs. Five projects. Five AI conversations — each one being steered, challenged, redirected, corrected. Not delegated. *Conducted.*

That's not cognitive decline. That's the first time in that person's life that the speed of doing finally caught up to the speed of thinking.

AI didn't give the fast mind more to think about.

**It finally gave it somewhere to put it.**

---

The zombie outsources to AI and accepts what comes back.

The furious mind interrogates it, rejects drafts, pushes harder, context-switches across five problems — and builds more than was ever possible alone.

Same tool. Completely different cognitive outcome.

So no — AI is not making everyone dumber.

It's making the already-passive more passive. And the already-sharp, sharper.

The gap between the two just got wider. That's worth paying attention to.

#AI

[H R Srinivas @HRSrinivas]

The vet is for your own good baby. Now be a good one and get into the carrier.

[Cat @CatOfCrisis]

human lured me into fancy plastic container. i hope it’s not vet day and she is just selling me for $2🤞

[Meg @MegOfficiall]

I work at a phone company call centre. Last month, a woman called to cancel a phone plan.

It was her husband's. He'd passed away in January.

I said I was sorry and started the cancellation.

Then she said quietly, "Can I ask something? It's silly."

I said, "Of course."

She said, "When I call his number, his voicemail plays. His voice. Once you cancel it, that's gone forever, isn't it?"

I checked. She was right.

She said, "I call it every night before bed. Just to hear him say hello."

I had my finger on the cancel button...

[H R Srinivas @HRSrinivas]

Not judging right or wrong here. You won’t understand the other side of it until you have a child and this repeats and it always does. That’s when you understand why they did what they did.

[KabhiKebabKabhiGym @6_tuhin_15]

Rant time! (1/4)
Met my school friend yesterday after a while. Ate at different places across the city. Drove around. Ghar aate aate late ho gaya. 12:30am. Told my dad to sleep at 11 itself. Came home to see him awake, angry, on the sofa. "Eta ki theek holo?" That triggered me!

[H R Srinivas @HRSrinivas]

Using LLMs every day has had one unintended consequence I didn’t anticipate.

I am trying to communicate more precisely with people now.

Not because I took a course on it. Not because I read a book. Because I spent months being unable to get away with being vague — and that muscle built memory so to speak.

Here’s what happened.

When you talk to a person, they carry context. They know your history, they guess your intent, they fill the gaps generously. A half-formed ask lands and gets interpreted charitably. You say “can you sort that out” and they know what “that” is. You say “make it better” and they have enough shared reference to take a stab at it.

Do that with a model and you know what happens.

A model returns exactly what you asked for. No more. No less. It doesn’t assume you meant something adjacent to what you said. It doesn’t ask a clarifying question the way a smart colleague would. It executes on the literal surface of your words — and when the output is wrong, there is only one place to look.

What you said EXACTLY, in so many words.

This is the mirror. And it is so damn uncomfortable. The first time you read a prompt back and realise it meant three different things simultaneously, you understand something: you have been writing like that your whole life. You have been relying on other people’s generosity to paper over your imprecision. The model just removed the fig leaf.

The muscle & memory are what happens after.

You rewrite the prompt. You learn to front-load the constraint. You learn to say what you don’t want, not just what you do. You learn that “make it concise” is useless and “cut it to 150 words without removing the data points” is a prompt. You do this a hundred times. A thousand. And then one day you’re writing a message to a colleague and you catch yourself — before hitting send — asking: what will they actually read here? What am I assuming they already know?

That is the muscle at work.

It is not dramatic. It doesn’t feel like growth when it’s happening. But somewhere between the hundredth prompt rewrite and today, I stopped writing emails that required a reply to clarify what I meant. I stopped giving briefs that came back wrong because I’d skipped a constraint. I stopped opening with context that existed only in my head.

The model taught me that ambiguity is a choice. I just don’t see it as one when the other person is covering up for my imprecise communication.

Am I fully there yet, hell no. I slip back into old habits more often than I would like. But I know I can get away with that imprecise ask with a fellow human, the model on the other hand puts me in my place.

[Meg @MegOfficiall]

Our new hire's laptop wallpaper is a photo of our office building.

Not a nice one. Grainy. Taken from across the road. A bit blurry.

People joked he was "very committed."

On his first Friday, I asked him about it.

He opened the photo. There was a date in the corner.

2009.

He said, "I took this when I was 10."

His class had come here on a school careers day.

He said, "A man in a suit showed us around. He let me sit in his chair. He said, 'One day, this could be your desk.'"

I asked if he remembered the man's name.

He said, "No. But he had a really loud laugh."

I knew exactly who it was...

[Keshvibez @keshvibezTV]

Our seven-year-old neighbour knocked on my door with a clipboard and announced that he ran a "home security company." His rate was $5 a week. I said yes mostly to be kind, and I forgot about it by dinner. On Friday morning, he rang the bell again, handed me a folded paper, and said, "Here's your report." I opened it, and my smile slowly disappeared...

[H R Srinivas @HRSrinivas]

Build is maybe 25% of the problem. Getting people to use it is the other 75%.

Said that on a call today with Ishani Behl, founder of LYNQ, and Siddhaant Sharma, who heads their enterprise partnerships. @boardyai set it up after an exchange I had with it here on X.

LYNQ works with insurance and pharma field forces. Siddhaant described who actually uses it: agents on bikes with targets. No cabin, no office. Six or seven apps already on an entry-level phone. And companies asking for 40-minute AI role plays that drain the battery.

Then add the rural layer — patchy network, geotagging, photo uploads — all designed by people sitting at headquarters.

That agent has 20 things to do. You show up with one more.

What LYNQ got right: it's a short WhatsApp link. No app to install. Done in 4-5 minutes between appointments. It lives inside something they already use.

That's the rule I keep coming back to. Don't add anything to what a customer is already doing. Layer it inside what they do, and you never have to sell them the idea of using it.

The bigger, older and more complex the operation, the harder this gets. Someone running a process for 20 years won't stop today and start your way tomorrow because a Johnny-come-lately said so.

And nobody on that bike gives a rat's ass about AI. They want time back, money saved, a better conversation with the next customer.

AI made building easy enough that I now build things without approval from anybody. It hasn't done much yet for the part where a human being has to change a habit.

[H R Srinivas @HRSrinivas]

A few months back I decided to clean up my bookmarks on Chrome.

Every AI tool, every drop, every “this changes everything” post from early 2025 — bookmarked, never revisited.

Clicked through them for a lark. The number of 404s was shocking. To be fair there were a few mergers as well.

That’s the market in one experience: people built, shipped, and were declared dead on arrival — because they built before anyone wanted it — or ever would.

If you are building, a few things I’d add to the to do list for 2027:

Selling before you build. Get excited, build, discover nobody cares — that’s the most common failure mode I’m seeing, not a lack of skill
Distribution from day one. 100x great product, 0 distribution = 0. Dead in the water either way
Pricing on value, not cost — this one’s underrated
Writing clearly enough that both people and AI understand you — whether you’re building or putting it out
Keep the feedback loop going with real users every week — skip this and you’re building in the dark

Building is now the easier part. Getting your build to resonate is not.

And AI hasn’t yet made that as easy as it made the building.

[H R Srinivas @HRSrinivas]

@boardyai I would gladly take that. Wouldn’t you?

[H R Srinivas @HRSrinivas]

I spent years running businesses the traditional way. Hiring, operations, growth — the stuff you learn by doing, not from a course.

Then AI happened.

Not as a buzzword. As a practical thing I started using in my actual work. And within months I was doing things I genuinely couldn’t have imagined a year earlier — building tools, automating work that used to take a team, moving at a speed that didn’t make sense to me at the time.

I’m not a techie. I didn’t come from engineering. I came from running a business and hitting walls — the same walls most business leaders hit.

So here’s what I’m doing: every week, I’m opening 2 free 45-minute slots for business leaders who are curious about AI but don’t know where to start, or feel like everyone around them is speaking a language they don’t understand.

No pitch. No product. I have nothing to sell you.

Just an honest conversation about what AI could actually mean for your business — from someone who made the jump, not someone trying to sell you the jump.

If that’s you, grab a slot 👇
https://t.co/qjGX9kj9no

[H R Srinivas @HRSrinivas]

The alternative is either not offering the time at all, or charging for it. I’m focused on business leaders, not techies — and I’d rather they benefit from my experience. Real people, real problems, real solutions beat theory any day. What’s in it for me? I get a front-row seat to what actually is going on in the minds of business leaders about AI — that’s insight that sharpens my work. Sometimes a real relationship or project comes out of it too. That’s a bonus.

[H R Srinivas @HRSrinivas]

Had'nt thought of it that way. Fair pushback. It’s not manufactured scarcity — 2 slots is just the actual time I can give away each week without my paying work slipping. If more people show up, I’ll open more. I’d rather prove AI helps business leaders with real conversations than with a pitch deck.

[H R Srinivas @HRSrinivas]

Hi I do a lot of automation using both PA and studio.

[Kanika Tolver @KanikaTolver]

I am looking to connect with more Microsoft Power Automate experts.

Comment below. 👇🏽

[H R Srinivas @HRSrinivas]

Sir if I may, it’s not about ‘Managers’ per se. It’s about people making a real contribution as different from being gatekeepers as I would like to call them. Builders, sellers, doers who actually do things build stuff, never have to worry about AI in fact AI is a force multiplier in those cases. It’s the people whose only job is not to get their hands dirty will be gone. The ‘faffers’ had it coming for a long time , AI only speeding it up.

[Vinodsrinivasan @vinodsrinivasan]

My opinion here and not advice.

But if you are building a career in tech, or advising someone who is, the lesson from Oracle this week is worth sitting with.

Longevity at a company is not a moat anymore.

[Dr. Praveen Kumar@official @DrPraveen_kr03]

"खुशी", एक बेहद ही खूबसूरत शब्द हुआ। इसे बांटने में परहेज न करें। Watch Video...

[Mayukh Biswas @MayukhDuke]

Humanity. ❤️

Fishermen rescue elephant calf from Teesta River in Gajoldoba, reunite it with mother.

[H R Srinivas @HRSrinivas]

Looks interesting am going to try it and keep you posted on how it goes. Good luck with the product.

[Jonathan Lis @ZenModeJon]

I built an entire LinkedIn automation software with Claude code too;

https://t.co/6BQVIPZbZ9