Step 01
Assess
People and data. I start with the operating reality, not a generic framework.
operating layer for operators & their backers
AI inside the operating plan, not next to it. Decades operating businesses across engineering, finance, sales, marketing, and the CEO seat. Now I help founders, enterprise transformation teams, and PE operating partners turn AI from scattered experiments into real operating leverage. In weeks, alongside whoever's already on your strategy work.
sks/ai ❯
The approach
01The operator story
Decades operating businesses, transforming them digitally. I've sat in your seat, running e-commerce, distribution, services, scaling teams, owning P&L.
That perspective is the substance. I can help you determine what works inside a real operating plan, and what is likely noise. The hype cycle moves fast, but organizational transformation requires planning and commitment
AI inside the plan, not next to it. The operating system underneath is the same. What changes is the scope it gets pointed at.
A founder's growth plan, a fund's value creation plan, or an enterprise transformation roadmap. Same approach. Different scales.
// the process. How the approach runs.
Step 01
People and data. I start with the operating reality, not a generic framework.
Step 02
Themes and use cases. Sequenced by impact, by feasibility, by what the business can absorb.
Step 03
Business cases that drive decision-making: What it costs, what it returns, who owns it, when it lands, how it sustains.
Step 04
Prototypes on real data. Production solutions if necessary.
Principle 01
What gets shipped is what gets measured. Working systems on real data, not slideware. The output shows up on the P&L or in the value creation plan.
Principle 02
Project and deliverable basis. Not hourly, not on a meter. The person you meet is the person who builds. No layered associates, no status meeting handoffs.
Principle 03
If you've already hired a strategy partner, keep them. The roadmap stays the strategy. I'll help alongside and make things a reality.
How we help
02How we help
Pick the one that fits your situation. The operating system underneath stays the same.
For founder-led operators ready to grow the business or cut operating cost. Sometimes the answer is a great off-the-shelf tool. Sometimes it needs to be built around the way the business actually runs. I figure out which is which, without upselling.
m01 · strategy.tooling
online
I learn how you run things and recommend the right mix of existing tools and custom solutions, so you invest in what actually grows the business or cuts cost, not what looks good on a slide.
❯ diagnose --scope=workflow
m02 · workflows.integration
online
Connect the tools you already use (CRM, marketing, ops, spreadsheets) into smooth automated workflows. Less copy-pasting, fewer mistakes, more time for the work that actually grows the business.
❯ connect crm + store + ops
m03 · build.agents
online
When off-the-shelf falls short, I build custom AI agents shaped around your exact operations. Solutions that fit the way you think about your business, not the other way around.
❯ build agent.<your-business>
// what you get
Every engagement is oriented to growing the business or cutting operating cost. Not automation for its own sake.
// in the fieldA founder-led retail business runs its whole operating stack on a platform I built in ten weeks. Six phases shipped, four roles, daily use. Spreadsheets and paper retired.
For enterprise teams already running an AI or transformation program with MBB, Big 4, or a strong internal IT function. I work alongside them to bring results faster, going deep within a single group or function.
The strategy partner has the roadmap. IT has the backlog. The gap is the layer in between. Someone who turns slide twelve into a working agent on real data before the next quarterly review.
That's where I sit. Embedded delivery, in weeks. Use cases mapped against the existing playbook. Prototypes built on real data. A sequenced delivery plan handed to your team. Plus executive coaching for the VP-level leaders driving the AI initiative inside the business.
// engagement · cadence
03 stages
01
Use cases scored against the existing roadmap.
● week 1
02
Two prototypes on real data, in 4 weeks.
● weeks 2–4
03
Sequenced delivery plan, owned by your team.
● week 5
Map → Build → Ship
// what you get
// in the fieldA ~$20B public building-materials supplier signed a scoping engagement with a global strategy firm already in the building. Fifteen AI use cases mapped onto the client's own commercial flywheel, executive workshop delivered, prototypes scoped. Complementary to the incumbent, not a duplicate.
For mid-market PE firms thinking about AI as a value creation lever at the firm level and across the portfolio. I work with funds and operating partners on the strategy that decides where AI shows up in the VCP, and where it does not. Plus the firm-level capability development that makes the play repeatable.
The interesting question isn't "can we use AI?". It's where AI changes the math on sourcing, diligence, LP narrative at the fund level, and on revenue, margin, M&A repositioning, and the exit story at the portco level. Most portfolios are doing the first version of that work right now. Most of it is shallow.
Real adoption is the signal. Slide-deck adoption is not. I sequence the operating-model layer before tools, decide what to build, buy, govern, or avoid, and stay long enough that the work compounds at the LP level.
The output is a value creation plan that doesn't quietly drop AI in week four. AI is wired into sourcing, diligence, revenue, margin, ops, M&A, repositioning, comms, and the exit story from the start.
// portfolio · ai activation map
06 portcos × 05 levers
| Rev | Margin | Ops | M&A | Exit | |
|---|---|---|---|---|---|
| Portco α | ●●● | ●● | ●● | · | · |
| Portco β | ●● | ●●● | · | ●● | · |
| Portco γ | ●●● | · | ●● | · | · |
| Portco δ | · | ●● | ●●● | · | ●● |
| Portco ε | ●● | · | · | ●●● | · |
| Portco ζ | ●●● | ●● | · | ●● | ●●● |
unscoped
scoped
in production
Strategy → Diagnose → Foundation → Accelerate → Govern → Sustain & Scale
// what you get
At the firm level
At the portco level
// in the fieldA lower-middle-market industrial PE firm ratified an AI Ambition I authored through its AI committee, runs a sourcing-intelligence desk I built on its live CRM every day, and then asked me to lead the production build and a portfolio-company pilot as prime.
How we engage
03How we engage
Same approach, different scope. Pick the engagement that fits the work, not the buyer type.
Engagement 01
$25K · 4–5 weeks
Scoping, assessment, prototypes, and business cases. Fixed bid. Scoped at kickoff. Where most engagements start.
Engagement 02
Scoping from $5K · build priced per engagement
When prototypes need to ship to production. Start with a 1–2 week scope. The build phase is priced against the work it takes. Fixed-bid posture inside each phase. No meter.
Engagement 03
Sized to time + commitment
Embedded AI leadership for funds and firms. Days per month, length of engagement, and scope set together at kickoff. Senior on the work, not behind it.
// alternative models. Outcome-based and usage-based pricing are on the table, especially for smaller businesses. Tell me what model fits how you'd rather pay.
//recently deployed
Five engagements since March 2026, from a $10M boutique to a ~$20B public company. Same operating system underneath; the scope is what changes.
// active deployments
05 instances
01
Fractional CAIO. Firm-wide assessment, an AI Ambition ratified by the firm's AI committee, a sourcing-intelligence desk on the live CRM in daily BD use, an LP-intelligence desk for the next raise. Now leading the production build as prime.
● in production build
02
AI opportunity assessment and a sales-knowledge capture pilot at a ~$95M equipment distributor. Capture tool built and deployed on privacy-preserving inference, in use by the company's own trainer. Product-knowledge brain prototype in build.
● pilot in flight
03
Scoping how AI scales a commercialization pilot at a ~$20B, 61-market company, alongside the incumbent strategy firm. Fifteen use cases mapped on the client's own flywheel. Executive workshop delivered, prototypes scoped.
● scoping
04
Sales and buying intelligence layer over a legacy ERP with no API. Live since April: customer follow-up email under 13 named guardrails, a daily ranked call queue for 16 reps, buying intelligence for the owner, a vendor upload portal. Transitioned to a successor engineering vendor as a going concern.
● in production · handed off
05
Replaced the manual operating stack: spreadsheets, paper, per-designer Excel. Six phases shipped in ten weeks. Daily use across four roles: order pipeline, client CRM, per-designer P&L, an AI linesheet parser at cents per file.
● in production
04About Sonesh
I'm Sonesh Shah. Two decades operating and transforming businesses. Not consulting on them.
I've held the seat at every altitude AI gets pitched to. Global President of an international industrial brand. VP Marketing and Head of Digital at a flagship power-tools business. Director of Digital, eCommerce, and IoT building the operating stack end to end. Engineering. Finance. Sales. Marketing. The P&L itself.
Industries. Industrial manufacturing, power tools, construction, B2B distribution, e-commerce, IoT. Scale. From country-level operations to a global business unit.
Since March 2026. Five client engagements, from a $10M boutique to a ~$20B public company. Roughly 120,000 lines of production code shipped. Working systems in daily use by ~26 client users across two companies.
The breadth is the substance. I know what works inside a real operating plan, and what is likely noise.
I've shifted my full attention to AI because the value at stake is bigger than anything I've worked on. I'm here to help operators capture it.
I don't sell hours. I don't sell decks. I embed for a defined window and ship the working pieces. The ones that show up on the P&L or in the value creation plan.
When the answer is a tool you can buy, I say so. When it isn't, I build it. When you've already hired a strategy partner, I work alongside them. Not in place of them.
No hype. No black box. No unnecessary builds.
// the sks/ai approach
I sit inside operating plans as the AI substance. A founder's growth plan, a fund's value creation plan, or an enterprise transformation roadmap.
// sks/ai operating principle
05Get in touch
Tell me what you're working on. A short conversation about where AI fits, and where it doesn't.