Sravan Nallana
Resume
Sravan Nallana, Senior Product Leader in agentic AI and digital commerce

It looks like a conversion problem. Underneath, it’s a customer experience problem.

I find the customer need hiding behind the metric, then build the product that meets it. For 18+ years at Verizon and Optum, that habit has turned into measurable growth. Most recently, I led Verizon’s move from conversational GenAI to agentic AI across a digital sales funnel serving tens of millions of customers.

See my work Email me
Every story here started as one problem and turned out to be another.
$50M+revenue influenced
30M+users served across telecom and healthcare
+9%conversion from agentic AI at Verizon
+23 ptsNPS turnaround in 12 months

Customer first. Outcome always.

I’m a Senior Manager of Product Management at Verizon, where I own the end-to-end digital sales funnel: discovery, personalization, promotions and checkout. Before that, I spent six years building healthcare platforms at Optum and seven years in consulting.

Customer-obsessed

I start with the customer behind the number. My signature move is the diagnostic reframe: separating the problem that shows up in the dashboard from the one actually causing it.

Outcome-driven

I set the north-star metric and OKRs before the roadmap. I measure success in conversion, revenue, retention and NPS, never in features shipped.

Agentic AI builder

I took Verizon’s VzGPT from summarizing information to completing multi-step shopping tasks, adding autonomy only once customers had reason to trust it.

Agentic and generative AI

LLM products, RAG and grounding, AI evaluation, responsible AI, human-in-the-loop design.

Customer discovery

Journey mapping, behavioral analytics, root-cause analysis, user research, cohort analysis.

Outcome-driven execution

OKRs and north-star metrics, experimentation systems, cross-functional squads, PM coaching.

Selected work

Each story follows the same arc: what the problem looked like, what it actually was, what I did, and what changed. Open any story for the detail.

Featured: agentic AI

VzGPT: from answering questions to finishing the job

Verizon, digital sales
Assumed
Customers want a smarter chatbot.
Real problem
They want the work done for them: find, compare, decide, buy.

Customer problem

Buying a phone meant weighing devices, plans and promotions across many screens. Every extra step sat between a customer’s intent and their purchase.

Insight

Trust in AI is earned in stages. Each layer of autonomy had to prove it was accurate and useful before the next one shipped.

What I did

Set the staged roadmap and aligned engineering, data science, design and marketing around one outcome framework, moving the platform from conversational GenAI to agentic AI.

  1. SummarizationCondense product, plan and offer detail into answers customers can act on.
  2. RecommendationsSuggest the right device, plan and promotion for each customer.
  3. ConversationLet customers ask, refine and follow up in their own words.
  4. Agentic actionsAgents run the multi-step task: search, compare, recommend and assist at checkout.
+9%conversion after agents began completing multi-step tasks across the funnel

One outcome, four squads

Verizon, AI commerce strategy
Teams need to ship more features.Four squads were chasing four different goals.
$12Mincremental annual revenue

Problem

Engineering, data science, design and marketing each defined success differently, so the customer experience moved in four directions.

Insight

Alignment beats throughput. One shared outcome turns four roadmaps into one.

What I did

Owned the 12-month strategy for AI-powered digital commerce, set the OKRs and north-star metric, and aligned all four squads to them.

Outcome

+15% conversion and $12M incremental annual revenue. Digital share of sales grew from 18% to 22%, beating gross-adds targets by 5–8%.

The reminder that knew where customers stopped

Verizon, cart abandonment
Customers forgot to finish.One generic reminder ignored where each of them got stuck.
+11%conversion, against a ~2% target

Customer problem

A static cart reminder sent everyone the same message, earning 5.5% click-through and 1.7% conversion.

Insight

Someone who left at plan selection needs a different nudge from someone who left at payment.

What I did

Replaced it with reminders personalized to the exact funnel step each customer left. I didn’t own the personalization model, so I brought the owning team, design, legal and marketing behind the test.

Outcome

+64% engagement and +11% conversion, more than five times the ~2% target.

Protection at the moment of purchase

Verizon Mobile Protect
Customers don’t want device protection.The offer was getting lost at the point of sale.
+10%attach rate, ~$5M incremental revenue

Customer problem

Device protection attach rates lagged at checkout, leaving customers uncovered and revenue on the table.

Insight

Journey mapping and behavioral analytics showed the drop-off came from the experience, not from a lack of demand.

What I did

A/B tested targeted UX changes with design and engineering at the exact steps where customers dropped off.

Outcome

10% improvement in attach rate, worth about $5M in incremental revenue.

Every support call is a product signal

Verizon, self-serve
Support needs more capacity.Customers were calling about things the product should handle.
−18%call volume, $1.5M saved a year

Customer problem

High-volume support tickets meant customers were waiting on calls for answers they wanted in seconds.

Insight

The top ticket types pointed to gaps in the product, not in the call center.

What I did

Ran a root-cause analysis of the highest-volume tickets, turned it into a self-serve roadmap and shipped three product interventions.

Outcome

18% fewer calls and $1.5M saved every year.

Turning around a negative NPS

Optum, ScriptSwitch (UK)
Users are unhappy with the product.The teams behind it had no shared picture of the customer.
+23 ptsNPS, from −3 to +20

Customer problem

A UK healthcare product had a negative NPS, and the teams building it worked in silos.

Insight

Trust had to be rebuilt step by step: quick, visible wins first, long-term platform improvements after.

What I did

Turned multi-stakeholder research and quantitative data into a sequenced roadmap, and rallied teams around a shared CX north star.

Outcome

NPS moved from −3 to +20 within 12 months.

Rebuilt for 7x scale

Optum, OptumRx PBM platform
The platform just needs to scale.It was missing capabilities the market already expected.
20M+members on the rebuilt platform

Customer problem

A pharmacy benefits platform needed to onboard a major new client far beyond its existing scale.

Insight

Market and competitive gap analysis showed scale was a capability problem, not only an infrastructure one.

What I did

Led the gap analysis and product redesign of the platform to close the capability gaps.

Outcome

Platform supports 20M+ members, and onboarded a major client at 7x its previous scale.

$2Mannual revenue from redesigning promotions discovery, with 10% less transaction friction
75 bpslower customer churn within 6 months of redesigning a promotions offer
30% → 52%experiment win rate after building an experimentation operating system
100%team retention across 6 PMs over two years

How I lead

I own the end-to-end funnel. Each PM owns one part of the customer journey with goals I set, so every team can see how their work moves the number that matters.

6 PMs

A team that stays

I recruited, mentor and manage six product managers. Structured career frameworks and quarterly growth conversations kept retention at 100% over two years.

+40%

Experiments as a habit

I introduced hypothesis-driven experiment briefs, pre-mortem reviews and a shared learning library across the PM org. Experiment velocity rose 40% and the win rate went from 30% to 52%.

3 plans → 1

Executing a strategy I didn’t set

When Verizon simplified three plans into one, I led the funnel-wide customer experience change across US and India teams, covering impact analysis, design, metrics and launch with marketing.

How I work

Four habits that show up in every product I lead.

Start with the customer, end with the metric.

Every roadmap item has to trace back to a customer problem and forward to a measurable outcome.

Diagnose first, then build.

The first explanation is often a symptom, not the cause. Finding the real problem early keeps the team from building the wrong thing well.

Ship AI in stages.

Earn customer trust first, then add autonomy. Each stage has to prove itself before the next one ships.

Every experiment teaches something.

A losing test with a clear hypothesis is worth more than a winning one nobody can explain.

Experience

18+ years across telecom, US and UK healthcare, and consulting, including 12+ years in product management.

  1. Oct 2020 – present

    Senior Manager, Product Management

    Verizon

    Own the AI-powered digital commerce and personalization platform across mobile, web and app, serving tens of millions of customers through Verizon’s primary growth channel.

  2. Apr 2014 – Oct 2020

    Product Manager

    Optum Global Solutions (UnitedHealth Group)

    Led product strategy and platform evolution for healthcare systems serving 20M+ members, spanning onboarding, analytics and decision platforms in the US and UK.

  3. 2007 – 2014

    Consulting and business analysis

    Accenture, Cognizant, IIC Systems, Tech Mahindra

    Delivered business cases and process redesigns for major US health plans and providers, including $1.2M in savings on the largest US health plan BPO engagement.

Certifications

Advanced AI Program, Product Space (2025) Google Cloud Generative AI Leader (2025) SAFe PM/PO (2022) Pragmatic Certified PM (2016)

Education

  • MBA, Marketing and Finance
    Osmania University
  • BSc, Computer Science
    Andhra University

Let’s build AI products customers love.

Open to conversations about product leadership in AI, commerce and platforms. Based in Hyderabad, India.

sravan.nallana@gmail.com +91 91549 02481 LinkedIn View resume