Skip to content
All posts

I Build AI Solutions. Here's How I Actually Use Them.

March 27, 20263 min readDhruv Jain

I tell companies how to deploy AI into their operations. Fair question: what does my own setup actually look like?

Not the client work. The boring, daily, unsexy stuff that keeps my one-person consultancy running from a desk in Hong Kong.

Three things. What they save me. Where they still fall apart.

Email triage that bought me 90 minutes a day

I get somewhere between 40 and 80 emails a day. Client threads, invoices, LinkedIn notifications, newsletters I subscribed to at 2am and never read, cold pitches from people who clearly didn’t look at my profile before hitting send.

I built a Claude routine that reads everything each morning and drops it into three buckets: respond now, respond later, archive. Before this, I’d open my inbox at 8am and come up for air at 9:30 having done nothing except sort messages. Now the sort takes about 4 minutes. The urgent pile gets handled first. The “later” pile gets a dedicated 30 minutes after lunch.

That’s 7.5 hours a week I used to lose to inbox paralysis.

It still misjudges tone, though. A passive-aggressive client email lands in “respond later” when it should really be “respond now and choose your words carefully.” I scan the full list every morning before trusting the sort. That check takes two minutes and has saved me twice already this month.

Research synthesis that cut a 4-hour task to 40 minutes

This week I received 15 minutes of voice messages covering six different topics. A CRM recommendation, a homework assignment, a tech question, a deployment update, scheduling, and something confidential I can’t get into here.

Old process: listen to each message 2-3 times, scribble notes, spend an hour organizing everything into something actionable. New process: transcripts through Claude with a single prompt asking for every action item, every decision, every piece of intel, and every open question, organized by priority.

I got back 8 items, each tagged with status and next steps. Cross-checked against the originals. Two needed corrections. The rest was right.

Here’s the thing people skip: verification. Claude occasionally merges two separate points into one, or drops a throwaway comment that turns out to matter three days later. If you don’t check, you will miss something. The 40 minutes includes checking. Without it, you’re just generating organized-looking fiction.

ICP analysis that showed me what I couldn’t see

This one wasn’t about saving time. It was about seeing clearly.

I pulled my last 10 LinkedIn posts and ran them through a cold analysis. No context about my business, my goals, my positioning. Just the raw posts, as a stranger would read them.

The result was uncomfortable. My posts attract AI-curious founders who want to learn about automation tools. My LinkedIn profile sells done-for-you outreach services. And neither one reflects the direction I’m actually building toward, which is enterprise AI deployment for larger companies in APAC.

Three audiences. Zero overlap. I would not have caught that on my own. When you’re inside your own content, you read what you meant to say. A cold analysis reads what you actually said. Those two things are different more often than anyone wants to admit.

The real lesson isn’t about productivity

Everyone selling AI tools talks about speed. “10x your output.” “Save 20 hours a week.” I save maybe 10-12 hours. That’s meaningful. It’s basically a free day and a half every week. But it’s not the revolution the LinkedIn hype machine promises.

The thing that actually changed how I work was the ICP analysis. It didn’t save me a single minute. It showed me I was building content for an audience that would never buy what I sell. That’s not a productivity gain. That’s a course correction that affects every hour I spend on the business going forward.

Speed is what people buy AI for. Perspective is what makes it worth using.

I’m Dhruv. I help enterprise teams in Hong Kong deploy AI systems that take the repetitive knowledge work off their plate. If that sounds like a problem you’re dealing with, reply to this email. I read every one.

Request an AI Readiness Review

For CTOs, operators, department heads, and compliance leaders who need a practical path from scattered AI usage to governed adoption.

20-min review — exposure, use cases, next step
Your data stays yours — NDA on day one

Opens Cal.com to select your slot

Need context first? Read the proof, case studies or subscribe to the weekly essay.

Q2 AI readiness window

Find the shadow-AI risk before it becomes policy debt.

In 20 minutes, we'll identify the department to review first, the AI usage surface you can't see yet, and whether a readiness audit, workshop, or private AI pilot is the right next step.

NDA-ready20-minute executive reviewNo tool pitchFor regulated or data-sensitive teams

Best fit: CTOs, operators, and compliance leads who need a governed first AI use case.

Review output

Your first governed AI use case

Actionable
01

First department to review

Where AI usage is already creating leverage, risk, or hidden process drift.

02

Shadow-AI exposure surface

The workflows, data paths, and approval gaps leadership cannot currently see.

03

Approval-worthy next step

A readiness audit, workshop, or private pilot scoped for governance first.

The urgency is not hype. Once teams normalize ungoverned AI habits, cleanup becomes policy debt, retraining, and slower approvals.