Paytm DevRev Discovery & baselining
Paytm DevRev
Discovery & baselining

Resolving work in seconds, not days

DevRev is an agentic platform for Paytm's non-sales operations. To size what that speed is worth to each business line, we need a picture of how the work runs today — the teams, the tools and the time.

Nothing is mandatory
Answer only what you know
Around 10–15 minutes
Stop wherever you like
Send as often as you like
Multiple submissions welcome
Welcome back. We restored your saved answers from this browser.
00 · Optional

Who's responding

Helpful so we can follow up on the right threads — but leave it blank if you'd rather not say.
01 · The big picture

Operating baseline for non-sales

High-level only. The goal is one agreed baseline number and scope — it becomes the denominator we measure outcomes against, not the headline.
1.1Non-sales strength & cost

The technology, operations and support teams outside the sales/field org. Headcount and annual fully-loaded cost (salary + benefits), split on-roll vs. contract if you have it.

Population
Headcount
Annual fully-loaded cost
On-roll
Contract
Total non-sales
1.2Split by function

Approximate is fine. Include large contractor / BPO spend doing the same work.

Function
Headcount
Annual fully-loaded cost
HR Ops / People Ops
IT Ops
Finance Ops
Data & Analytics
Business Ops
Other non-sales functions, with rough headcount and cost.
1.3Tooling & infrastructure
One rollup number: annual software + cloud + AI/token spend these ops teams run on. Itemised later, in section 02.
1.4What success looks like
Beyond headcount reduction — what business impact over the next 18–24 months would make this a win? Faster turnaround, freed capacity moved to revenue work, better customer or employee experience. Cost is one input; business value is the anchor.
02 · Detail

What the data platforms really cost

Three-way total cost of ownership across Trino, MetaQL and Prism — infrastructure, AI spend and the people who build and maintain each one.
Trino
 
 
Infrastructure / software
AI / token
People cost
Users writing raw SQL
MetaQL
 
 
Infrastructure / software
AI / token
People cost
Users
Prism
 
 
Infrastructure / software
AI / token
People cost
Users
03 · Detail

The people layer around the tools

Where analysts sit, and how much of their day goes into preparing data rather than using it.
Data and business analyst headcount by department and by workload — loan distribution, offline payments, product, growth marketing, finance, business ops.
Teams that prepare or transform data before it's usable by decision-makers, and roughly how much of their time that takes versus actual analysis.
Where critical numbers get computed outside the systems of record — e.g. finance revenue schedules built from partner data in Google Sheets, later reconciled to a single SAP entry.
04 · Detail

How the work actually gets done

What ops teams spend their days on, and what faster resolution would be worth to the business.
Could Paytm share a 5% ticket sample for AI process mining? We'd classify each ticket by category and by the service or business line it belongs to — not simply automatable vs. human-worked.
For the main ticket categories: what does slow resolution cost — SLA breaches, blocked downstream work, customer or employee impact, revenue at risk? And what would instant resolution be worth?
For the JSM automations already in place (e.g. 48 on HR-admin, 22 on marketing-design) — is each built natively in JSM, or does it call an external platform?
Step-by-step ops work living elsewhere: invoice processing, legal contracting, recruitment and onboarding (Genesis, ~10K field hires/quarter), the ~1.5M-device asset system, sales-ops on WhatsApp and spreadsheets.
05 · Detail

Systems access & cost attribution

What telemetry we can draw on to see which systems people actually use, and where spend lands.
Usage data for a sample of agents.
Document and collaboration activity.
Can the LLM gateway attribute tokens by workload or use case? If not, is AWS cost-by-service available as a fallback?
06 · Detail

The CRM question

Understanding the driver behind LeadSquared before anyone proposes a solution.
Genuinely discovery — we'd rather understand the need first than assume the answer.
Sales and lead work has historically run on Excel and WhatsApp groups. What changed? What problem is a CRM meant to solve that the current way can't?
Capturing leads, tracking follow-ups, nudging users who dropped off, hand-offs between teams — what matters most?
The newer lines (Paytm Money, Insurance, Bank), or does it extend to core payments, wallet and UPI?
Implementation status and who owns it.
07 · Detail

Where we should start

The two or three places a first proof of concept would prove the most.
For example HR-separations ticket automation, finance spreadsheet-to-SAP reconciliation, or a conversational surface for sales ops.
Send us what you have
No field is required, and you can submit as many times as you like — we'll collect every response. Partial answers now are more useful than complete answers later.