Director of Product Management (Agentic AI)

Director of Product

• Department:

Product \& Technology

• Reports To:

CEO

• Location:

Gurugram, with regular US travel and time inside dealerships

• Work timings:

India hours, with committed US overlap

• Team Stage:

Series B

• Experience:

9\+ years in product, including 3\+ years managing product managers

• Scope:

One of 3 lines: Studio AI, Vini AI, or Platform. 2 to 5 product managers, plus design and analytics partners

• Key Partners:

CTO, Associate Directors of Technology, COO, CRO, Customer Success, Lot Services and Contact Center leadership

About Spyne

Spyne builds AI software for US car dealerships. Two products are on the market. Studio AI turns phone-shot vehicle images into retail-ready merchandising in minutes. Vini AI is a conversational agent suite that handles dealership sales, service, recall and acquisition interaction: inbound calls, web leads, follow-up, scheduling.

• Dealerships live:

2,000\+

• Revenue from the US:

80%\+

• Systems integrations:

70\+ across CRM, DMS, inventory, service, finance

• Team:

200\+, primarily India-based

• Capital raised:

$25M\+ from Accel, Vertex Ventures, and Storm Ventures

The Business You Would Build For

You can’t make good product calls here without an in depth dealership P\&L knowledge.

• The car sale barely makes money, and sometimes loses it.

It’s a volume event that creates 2 profitable ones. So every value case gets argued in cost and days removed. A roadmap justified on selling more cars won’t survive a dealer conversation.

• F\&I is high margin and sits in a separate office.

Finance, service contracts and GAP carry much of the gross on a deal. Sales, service, F\&I and parts run as separate P\&Ls under one roof, and nobody gets paid on the handoff between them. That’s where deals leak.

• Service and parts throw off about half of dealership gross profit on an eighth of revenue.

Franchised dealers did over $164 billion of it last year. The service drive is the bigger long-term prize and the incumbents barely serve it. Vini’s service side is harder to build than the sales side and worth more.

• Inventory turn drives everything else.

A car sitting on the lot costs the dealer $37 to $75 a day. Turn the same inventory twice as often and annual gross roughly doubles with no extra customers. That’s Studio’s whole value case.

What You Own

• The outcome:

You carry one metric a dealer can feel: days from acquisition to listed, automation rate, kept appointments per rooftop, attach rate, or ACV expansion, depending on the line. You report the trend monthly with the reason behind it.

• Discovery you do yourself:

You’ll be in US dealerships, on the phone with general managers and fixed ops directors, and listening to recorded calls every week.

• Pricing and packaging:

What you charge for is what you optimize. Per seat, per conversation, per resolution, per vehicle and per outcome each produce a different product and a different set of incentives. You own the product input into pricing across your line, including how the services scope gets sold. The CRO owns the number. You own the shape of what’s sold.

• Quality as a specification:

You define what good means before engineering builds it, and you co-own the eval set that enforces it. What counts as a resolved conversation, an acceptable image, a correctly booked appointment. What blocks a release.

• The configuration surface:

Thousands of dealers each wanting something different can’t mean per-dealer engineering. Every hour a deployment engineer spends on one customer is a defect in your product.

• The team:

2 to 5 product managers, plus design and analytics partners. You hire them, level them, and give feedback in days. Develop at least one into someone who could run a line.

Product in an Agentic Company

• You own the eval set with engineering.

Golden cases from dealer traffic, a quality bar written before the build, and a release gate you enforce. You can separate a regression from noise, and you know when a model grading a model needs human labels underneath it.

• You plan for model change.

Your roadmap survives 2 model generations. You’ve shipped through a deprecation or an upgrade that quietly changed behavior. Version 1 of an agent is easy. Holding it steady for 6 months is work.

• You know which metrics lie.

High containment with frustrated callers is worse than lower containment with satisfied ones. Task success gets measured against the system of record. An agent that says it booked the appointment and didn’t is a failure that reads as a success.

• You spec the failure paths.

Escalation, confirmation before anything irreversible, and what the agent does when it doesn’t know. Vini speaks on the dealer’s phone line, so a bad call lands on the dealer.

• You trade latency and accuracy against cost.

A decision that adds 400ms or doubles token spend is a P\&L decision. Holding gross margin above 70% while 2 human-delivered service lines scale is partly your job.

• You own automation rate.

You set the definition and the quarterly target for contacts resolved and units captured with no human touch. It gates expansion on both service lines and it is the margin on both.

• You say no to a demo, including one an executive loves.

Anyone can build a convincing prototype in an afternoon now, so this is a product skill rather than a personality trait.

• You do your own data work.

SQL or the BI layer yourself, your own cohorts and funnels, and suspicion of a number that looks too clean.

What We’re Looking For

• A revenue-bearing product line you owned

, with its own metric, roadmap and customers, where a bad quarter was visibly yours.

• AI products shipped to paying customers.

You can name a production quality failure, what it cost, and the product change that killed the whole class of it.

• Quality defined before the build

, run against a real eval set, with a release you held.

• Built for an operator with a job to do.

Vertical SaaS, fintech, healthtech, logistics, contact center and field ops all translate. Auto retail is a bonus, and we’ll teach the domain to someone who has proven they can learn one.

• Pricing and packaging you’ve moved

, and lived with what it did to the funnel.

• Your own analysis.

You find the bad number before the dashboard shows it.

• Writing that provokes questions about edge cases

instead of confusion about intent.

• PMs you’ve managed, hired, and grown

into bigger roles.

• A product decision that came from sitting in a customer’s workplace.

You’ll be on the road to US dealerships regularly.

• A prioritization philosophy that holds under pressure.

We’ll test it with a real conflict from our roadmap.

What We Offer

• Ownership:

A full product line with revenue attached

• Autonomy:

Your own roadmap, your own hires, your own quality bar, working directly with a founder who wants a partner rather than an executor

• Problems:

Agents running unsupervised on a business phone line, vision output publishing without human review, and a services margin only a product decision can move

• Compensation:

Benchmarked to senior product leadership at fast-scaling AI companies in India

• Equity:

An ESOP grant under a published policy with band-based multipliers and periodic refresh

• Trajectory:

Head of Product for the line, with a path to VP Product or GM as the business scales