Rishi Desai
Founder of InnovateX. Nine years across MBB strategy consulting and in-house strategy and transformation leadership — and an AI specialist who has deployed agentic AI into live commercial operations rather than presented on it.
I spent a decade learning how elite firms turn confusion into committed action. Then I spent three years finding out which parts of that survive contact with a real operating business.
The first half of my career was MBB. Consultant, then Engagement Manager — strategic due diligence and M&A advisory shaping multi-billion-dollar investment decisions, corporate restructuring and cost optimization for C-suite teams, and cross-functional teams of ten-plus consultants on the kind of engagement where the answer has consequences. Cumulative revenue impact advised across those engagements ran past $200M. One digital transformation drove more than $16.5M in incremental e-commerce revenue.
The second half has been operating. Senior Manager and then VP-level roles owning strategy, global operations and business transformation inside software and commerce businesses — building multi-year strategic roadmaps, owning pricing and packaging across segments, launching go-to-market expansion into APAC and the Middle East, and advising leadership on cost structure. That work moved real numbers: 8 points of EBITDA margin in one role, 20 points in a portfolio company through pricing and cost structure together, 15% ARR growth from pricing and packaging, 12% conversion improvement in a retail segment, and onboarding churn from 62% down to 38%.
The reason both halves matter is simple. Advisory teaches you how to find the answer. Operating teaches you what happens to the answer on Monday — which recommendations survive an actual budget, an actual org chart and an actual team that already has a full quarter.
The AI half of the practice
Most people selling AI consulting have read about deployments. I have run them.
In an operating role I established and led a company-wide AI Council — the governance body that decides which workflows go to production, on what standard, with what oversight. I rolled agentic tooling into daily operations across the organisation, not as a pilot with an audience but as the way work got done. And I redesigned a customer-facing function around it, embedding agentic AI directly into onboarding, with a measurable result: onboarding churn fell from 62% to 38%, with corresponding improvement in adoption and lifetime value.
That experience produced a strong opinion, which is on this site in several places: AI almost never fails on capability. It fails on deployment. No owner after the pilot team disbands. No removal of the old path, so people quietly keep using it. No baseline, so nobody can prove the thing worked. No governance answer, so security stops it at scale. Those are operating problems, and they are why AI enablement here is an operations engagement rather than a technology one.
- AI CouncilEstablished and led a company-wide AI governance body — the standard a workflow had to meet before production.
- DeploymentRolled agentic tooling into daily operations across the organisation, not as a pilot with an audience.
- 62% → 38%Onboarding churn, after embedding agentic AI directly into the onboarding function.
- CertifiedAI Product Manager Professional Certificate, IBM.
Track record, in numbers
Outcomes from advisory engagements and in-house operating roles. Every figure below is one I owned or advised directly.
Figures are drawn from advisory engagements and in-house operating roles. Client and employer names are withheld under confidentiality obligations. Case studies on this site are anonymized or composite and labelled accordingly.
Four commitments, and the reason for each.
These are not values statements. They are constraints on how the practice operates, and each of them costs something.
- Two engagements at a time, maximum. The constraint is the product. It is also the reason there is no bench to sell you.
- Answer first. Page one is the verdict, not a build-up to it. If the honest read is that you have a serious problem, that goes on page one.
- Never invent a number. If a figure is not in the source material, it is flagged as missing. An absence — no documented ICP, no forecast accuracy tracking, no win/loss data — is frequently the most diagnostic finding available.
- The person you meet does the work. No handoff to a team after the sale. This is the whole reason a one-person practice can compete with a firm.
Background
MBB, 2018–2023
Consultant then Engagement Manager. Strategy and operations, M&A and due diligence, cost optimization and corporate restructuring. Market penetration work across North America, APAC and the Middle East.
In-house, 2023–present
Senior Manager then VP-level roles in strategy, global operations and business transformation across software and commerce. Multi-year roadmaps, pricing and packaging ownership, GTM expansion, AI enablement.
MBA · CFA Level II
Top-tier MBA. CFA Level II complete. Bloomberg Market Concepts. Google Project Management Professional Certificate. IBM AI Product Manager Professional Certificate.
On confidentiality. I carry obligations from prior employers and clients, and I honour them without exception. No client name, no proprietary material and no engagement detail from any prior role appears on this site or in any work produced here. Where outcomes are cited, they are stated without attribution. If that makes a claim harder to verify, that is the correct trade — and it is the same discipline your own information will receive.
Also
Thirty minutes, on your constraint.
No pitch. If the honest answer is that you do not need an advisor, that is what you will hear.