Anush Bulusu · Enterprise AI, Product Leadership, Transformation Strategy

I work where AI has to become real business value.

My work sits at the intersection of product strategy, enterprise transformation and AI-native product craft, turning emerging AI capabilities into scalable products, operating models and measurable outcomes.

AI × Product × Enterprise Transformation

Explore the three tracks LinkedIn
  • 16+ yearsEnterprise technology, product and transformation
  • 3 GenAI platforms21 product modules
  • 9 PMs5 direct + 4 indirect
  • 45+ engineersUnder functional product leadership
  • $10M+ ARREach across two scaled platforms
  • 3× adoptionAcross scaled products

Selected examples are intentionally presented at a high level and exclude confidential client, architecture and commercially sensitive information.

My point of view

01

AI does not create value because a model works. It creates value when the surrounding product, workflow and operating model work.

That belief shapes how I approach AI.

I start with the value equation, understand the workflow that needs to change, determine where intelligence genuinely creates leverage, build the right product system around it, and create the governance and operating mechanisms required for adoption at scale.

The model matters. But the product around the model matters more. And in an enterprise, the system around the product often matters most.

  1. Capability
  2. Product
  3. Workflow
  4. Adoption
  5. Business value

Three tracks

02

Three questions I keep working on.

Track 01 AI × Product Value Can AI create measurable product and business value?

I am interested in the distance between an impressive AI capability and a product people will actually adopt, trust and pay for.

This track is about 0→1 product creation, product-market fit, adoption, economics, portfolio choices, GTM and value realization.

 Explore → Track 02 AI × Enterprise Transformation Can the enterprise actually absorb and scale AI?

AI transformation is rarely constrained only by technology.

Operating models, fragmented workflows, ownership, incentives, governance, change management and organizational capability often determine whether AI becomes infrastructure or remains a collection of pilots.

 Explore → Track 03 AI-Native Product Craft Are we building AI products differently enough?

Probabilistic systems cannot be designed exactly like deterministic software.

AI-native product management requires new thinking around evaluation, agent behavior, human judgment, orchestration, observability, guardrails, trust and iteration.

 Explore →

Track 1: AI × Product Value

03

Start with the value equation, not the model.

A technically impressive AI capability is not automatically a good product. The questions I care about are more fundamental.

  • What user or business behavior changes?
  • What becomes faster, cheaper or materially better?
  • Will customers trust it?
  • Will they adopt it?
  • Does the economics work?
  • Is the opportunity large enough to justify building it?

The four proofs

Capability Workflow Trust Economics

If any one approaches zero, so does the product.

From capability to commercial product

I currently lead product strategy and P&L direction across three GenAI platforms spanning 21 product modules. Two have scaled beyond $10M ARR each, with approximately 20% YoY growth and 3× user adoption. A third has moved from 0→1 into early commercial traction.

My role spans the entire journey:

  • Opportunity identification
  • Product thesis
  • Customer discovery
  • Product architecture
  • MVP / 0→1
  • Enterprise validation
  • GTM
  • Adoption
  • Expansion
  • Portfolio investment decisions
  • $10M+ARR, each of two platforms
  • 20%YoY growth
  • 3×User adoption
  • 0→1Early commercial traction

Value beyond revenue

Enterprise AI often creates value by changing the economics of an existing workflow. Across AI-led workflow transformation initiatives, outcomes have included:

  • 50%Lower operational costs
  • 40%Faster go-to-market
  • 75%Reduction in processing effort

The important question was never “Can AI perform this task?” It was: “Can we redesign the workflow so the entire system performs better?”

AI should not merely compress a task. It should improve the economics or experience of the whole workflow.

Track 2: AI × Enterprise Transformation

04

AI transformation is an operating-model problem disguised as a technology program.

The first generation of enterprise AI programs focused heavily on use cases. The next challenge is different: how does an organization repeatedly identify valuable opportunities, build them safely, deploy them quickly, measure value and progressively move ownership closer to the business?

That requires more than an AI roadmap. It requires an operating model.

Maturity, as it actually progresses

  1. PilotProve that something works.
  2. ProductizeEngineer reliability, governance and repeatability.
  3. StandardizeCreate reusable patterns, tools and controls.
  4. FederateEnable business teams to own more of the outcomes.
  5. IndustrializeOperate AI as an enterprise capability.

Central leverage. Distributed ownership.

Central AI and product teams

Create leverage through:

  • Shared platforms
  • Reusable components
  • Governance
  • Standards
  • Evaluation frameworks
  • Architecture
  • Capability building
  • Partner ecosystems

Business teams

Should increasingly own:

  • Use-case outcomes
  • Adoption
  • Workflow redesign
  • Change management
  • Operational performance
  • Continuous improvement

Transformation across the enterprise

Before my current GenAI portfolio, I led strategy and delivery across eight internal enterprise products spanning:

  • Finance
  • Legal
  • HR
  • Operations
  • Sales
  • Quality
  • Risk
  • Compliance

These were not independent applications. They were interconnected workflows cutting across organizational boundaries. The challenge was therefore as much about stakeholder alignment, operating models and process redesign as it was about software.

One transformation replaced fragmented manual workflows with integrated enterprise platforms connecting Sales, Delivery, Finance and Operations.

$283M+In revenue recognition enabled

What this taught me

Enterprise transformation succeeds when you can do all of these at the same time:

  • Understand the workflow
  • Structure ambiguity
  • Align competing stakeholders
  • Create clear ownership
  • Build the right product
  • Change how people work
  • Measure whether the transformation actually created value

Track 3: AI-Native Product Craft

05

If the system is probabilistic, the product cannot be designed as if it were deterministic.

AI-native products introduce an entirely different set of product questions.

  • What should the model decide?
  • What should remain deterministic?
  • When should an agent act autonomously?
  • When should a human intervene?
  • How do we know whether the system is improving?
  • What happens when it is confidently wrong?
  • How do we make the system observable?
  • How does trust change the UX?

These questions change the craft of product management.

Where the craft changes

  • Agentic system design

    Agent orchestration, tool use, agent boundaries, deterministic versus autonomous execution and multi-agent patterns.

  • Human-in-the-loop

    Designing the correct boundary between automation, judgment, accountability and regulatory oversight.

  • Evaluation

    Moving beyond conventional QA toward task-level evaluation, quality thresholds, failure analysis and continuous improvement.

  • Guardrails & governance

    Model and prompt governance, explainability, policy controls and responsible deployment.

  • Observability

    Understanding what agents did, why they did it, what tools were used and where the system failed.

  • Knowledge & grounding

    Enterprise search, retrieval, knowledge graphs, integrations and the context required to make AI useful inside real workflows.

Changing how product teams build with AI

I created and institutionalized an Agentic Product Development Lifecycle to explore a simple question: if AI changes the product we are building, should it also change the way we build products?

The lifecycle introduced AI into discovery, problem framing, PRD development, design and execution while keeping human judgment and product accountability explicit.

  • 30–50%Reduction in discovery and design cycles
  • 15%SDLC capacity reclaimed
  • 18+Product Managers trained

The objective was not to make PMs use more AI tools. It was to create a better product-development system.

AI-native product craft is not about putting an agent into every workflow. It is about understanding where intelligence creates genuine leverage, then designing the rest of the system around that reality.

Current scale

The environment the work happens in.

  • 9 Product Managers5 direct, 4 indirect
  • 45+ engineersFunctional product leadership
  • India, UK, USGlobal operating scope
  • 3 GenAI platforms21 product modules
  • Fortune 500Enterprise environments
  • Regulated workflowsHigh-stakes decision contexts
  • Partner ecosystemIncluding Adobe Firefly integration

About Anush

06

A curious mind, most at home where structure meets ambiguity.

Building with purpose: bringing enough structure to chaos to create momentum that matters.

Anush Bulusu
Beyond work
Life is deliberately less structured.
Family
Keeps me grounded.
Sports
Cricket, tennis, Formula 1 and table tennis are rarely far away.
Road trips
Long drives are one of my favourite ways to reset.
Photography
Began as a hobby, unexpectedly turned into a small business, and taught me more about customers, entrepreneurship and human behaviour than I expected.

I have always been drawn to messy problems.

The kind where technology is only one part of the answer. A workflow crosses multiple teams. Everyone can see that something is broken, but no one owns the entire problem. Or a new technology suddenly makes something possible that simply was not possible six months earlier.

Somewhere between strategy, technology, operations and customers, somebody has to make sense of it. That is usually where I am happiest.

I am naturally curious, and probably ask “why?” more often than is convenient.

  • Why does this workflow exist?
  • Why would someone change their behaviour?
  • Why are we solving this problem at all?
  • What would have to be true for this to work at 10× the scale?
  • And increasingly: what should AI actually do here, and what should it not do?

Over time, I have realised that one of my strongest roles is that of a translator: between business and technology, ambition and execution, emerging capability and something people can actually use.

I enjoy building things, understanding how things work and occasionally taking the longer route simply because the journey looks more interesting.

The journey so far

The titles changed over time. More importantly, so did the kinds of problems I learned to solve.

2005–2012

Learning to think in systems.

Engineering gave me structure. My early years working on large enterprise and regulated systems taught me something equally important: technology rarely exists in isolation.

I learned how complex systems connect, how global teams operate, and why reliability, process and context matter just as much as functionality.

Key lesson

Structured problem solving

  • Engineering
  • Enterprise systems
  • Global delivery

2012–2019

Learning that influence scales further than authority.

As the problems became larger, the work became less about delivering a defined solution and more about aligning people around one.

I worked across global enterprise transformation, financial services, risk, compliance and automation programs, gradually moving from execution into solution and product leadership.

This phase taught me how to navigate ambiguity, work across organizational boundaries and translate technology decisions into business outcomes.

Key lesson

Influence without authority

  • Transformation
  • Stakeholders
  • Solutions
  • Leadership

2014–2020

Learning what customers actually value.

Alongside my corporate career, I built a photography business from the ground up. What began as curiosity became more than 250 customers and a real operating business.

It was my first visceral lesson in product-market fit. Customers do not care how much effort went into something. They care whether it solves a problem, creates an experience or gives them something they value enough to pay for.

That experience permanently changed how I think about products. Around the same period, my formal career also moved increasingly toward product management and intelligent workflow automation.

Key lesson

Value is defined by the customer

  • Entrepreneurship
  • Customers
  • Product
  • Commercial thinking

2020–2021

Learning to zoom out.

Going to IIM Calcutta was a deliberate reset. After years of building and delivering technology, I wanted to better understand the systems around it: strategy, markets, finance, operations, organization design and growth.

It changed the level at which I looked at product decisions. Instead of only asking “Can we build this?”, the questions increasingly became:

  • Should we build it?
  • For whom?
  • Why now?
  • What does it change for the business?

Key lesson

From product decisions to business decisions

  • Strategy
  • Business
  • GTM
  • Organization

2021–present

Building products, teams and operating systems.

The last chapter has brought the earlier ones together: enterprise technology, customer understanding, business strategy, product leadership and organizational transformation.

I now lead product portfolios and teams working on enterprise GenAI and agentic systems, while spending an increasing amount of time on a larger question: how does AI change not only the products we build, but the way organizations themselves operate?

That has taken my work beyond individual products into portfolios, operating models, governance, adoption, capability building and enterprise transformation.

Key lesson

From building products to building systems that scale

  • Enterprise AI
  • Product leadership
  • Agentic systems
  • Transformation

The common thread has stayed surprisingly consistent.

  1. Take something complex.
  2. Understand what really matters.
  3. Bring people around the problem.
  4. Build the right system.
  5. Make the outcome measurable.
  6. And then keep learning.

What I am thinking about

07

Some questions I keep returning to as AI moves from demos into real products and enterprises.

  • The AI demo worked. The product still failed.

    Why a capability that impresses in a demo can still fail every test that matters once it enters a real workflow.

    AI product

    In progress

  • The Four Proofs

    Capability, workflow, trust and economics: why a weakness in any one of them collapses the product case.

    Product value

    In progress

  • Your AI product has a behavioral contract. You just haven’t written it down.

    What users implicitly expect a probabilistic system to do, and what happens when nobody has made that promise explicit.

    Product craft

    In progress

  • Why your eval set is lying to you

    How evaluation quietly drifts away from the tasks, edge cases and failure modes that decide real-world performance.

    Evaluation

    In progress

  • You compressed the wrong step

    Automating the visible task while the real cost, delay or risk sits somewhere else in the workflow.

    Workflow design

    In progress

  • The validation tax

    What it costs when every AI output has to be checked by a human, and how that cost reshapes the business case.

    Economics

    In progress

  • What happens at 10×

    The assumptions that quietly break when an AI product moves from pilot volumes to enterprise scale.

    Scale

    In progress

These will link to longer pieces on LinkedIn and Substack as they are published.

Connect

08

The problems I find most interesting sit between AI possibility and enterprise reality.

I am always interested in conversations around enterprise AI, product leadership, agentic transformation, operating models and building AI-native products at scale.