AI Workflows for Companies

What most do

“How can we add AI to what we already do?”

Applying AI on a broken process will give you a faster process that remains broken

This is why AI spend shows no ROI

What works

“What should this workflow
be in an AI-native world?”

Rethink the operation first, then apply AI where it actually compounds

Measurable impact on P&L

Let's talk →

30 mins. We discuss your priorities and decide whether there’s a fit

Prefer email? p12siddharthm@iima.ac.in

How does this work?

1 Diagnose

I map how the work actually happens across functions
Identify the bottlenecks limiting revenue, cost, or speed

2 Redesign workflows

Remove non-value-adding work and make process changes
Identify high-value workflows that can benefit from AI

3 Build and execute

I build the redesigned workflow on your existing data and tools
Your team learns it as it's built, so they can own and run it

4 Measure impact

Track results against an agreed baseline
Iterate until the workflow moves revenue, cost, or speed

Selected case studies

A D2C water-filtration brand: Cost and revenue wins in 3 weeks

I mapped the existing workflows end to end across functions, working with the founder directly. Executing the following:

Evolving customer support into the company's highest-intent sales and product signal by mining support transcripts at scale and plugging revenue leaks
Increasing service levels by building a unified knowledge graph across the company, flagging egregious misses, and rebalancing CS agent load

Live engagement. Delivering on both cost and revenue

AdvocateDesk: Property title deed verification that took days, now drafted in minutes

Reviewing property title means reading 50-year-old deed bundles that are vernacular, handwritten, and error-prone, typically 5 to 7 days per case. I built a system that:

Ingests the full deed bundle (EC, patta, parent documents) as scans or PDFs, vernacular or English, typed or handwritten
Traces the chain of title, reconciles shares, and tables every document, however old
Drafts a title verification opinion, grounds every finding to source, and flags what needs the advocate's eye

Live at advocatedesk.app. Used by practicing advocates, saving 30+ hours a week each

CiteGap: Audit pipeline that runs 1,000+ real-time queries to increase brand visibility on AI Search

Buyers now ask ChatGPT, Claude, and Google's AI for recommendations, and most brands have no idea whether they show up, who's cited instead, or what to fix. I built CiteGap to close that gap.

Runs 1,000+ live queries across four AI engines, mapped to six real buyer personas
Shows exactly which competitors get cited instead of you, and the signals you're missing
I work with the brand to close the highest-impact gaps through a 30-60-90 day action plan built around key priorities

Live at citegap.com · 10+ audits across D2C, healthcare, education & SaaS, built solo with AI

Why me?

I bring a unique combination of product judgment at scale, operator context, and building hands-on with AI

13+ years across business, strategy, operations, and product; IIM Ahmedabad
Ex-Amazon Principal PM-Tech. Built products and workflows at 100M+ user scale
Hands-on AI builder with multiple shipped products end-to-end
Enabling founders, CEOs and senior leaders to build practical AI judgement

You work directly with me. I move fast, and focus on outcomes over frameworks or decks

Full profile

How engagements work

1

Diagnostic

Completed within one week
I work with you and your team to identify opportunities and jointly prioritise the highest-value ones

2

Implementation

Once prioritised, we scope the work and build
Results begin within weeks; a recent engagement delivered both cost and revenue wins within three weeks

Common questions

What kinds of problems are a good fit?

Business problems that materially affect revenue, costs, or speed of execution are a good fit. Identifying the underlying bottleneck is part of the diagnostic work. If there is a problem you are already working on, we can start there.

What do we actually get?

You get measurable revenue or cost gains, or a step-change in execution speed beyond what conventional process improvements can typically deliver. I deliver this by diagnosing your bottlenecks, redesigning workflows, and building and executing the solution using your data and tools.

What do you need from me and my team?

I need a founder or senior sponsor, an appetite to challenge the status quo, access to the people who own the workflows, and access to internal tools and data on a need-to-know basis. I typically diagnose and redesign in person. Your team stays involved through the build so they can own and run it.

AI has not given us the ROI we expected. How is this different?

AI often fails because it starts with a tool instead of the business bottleneck. I remove non-value-adding work and make process changes first. If AI does not offer leverage in the redesigned workflow, we do not use it.

What does the solution usually involve?

The goal is the simplest path to measurable impact. Sometimes that is a process change. Sometimes an existing tool is enough, and I recommend that before doing anything custom. If neither is sufficient, I build and integrate a custom workflow into your existing setup.

How do you handle our data?

We agree on data boundaries and access upfront. I use your existing environment and accounts wherever possible, limit access to what the workflow needs, and make any third-party tool or model usage explicit. Your team can see and run what is being built.