Nectar's Lahore team powers 10 million conversations a week, but its $30 million will hire in California and New York
The AI marketing company, founded by sisters who left Meta, says Pakistani engineers share its codebase. Its published figures show what the business claims for its customers and what each reply might cost to run.

At the volume its Lahore built technology now carries, every cent Nectar Social spends on one customer conversation would add $5.2 million a year to its costs. The company, founded in 2023 by the sisters Misbah and Farah Uraizee, said on 13 May that it had raised a $30 million Series A. Two months later Misbah Uraizee told ProPakistani that Lahore is one of its core engineering hubs and that the Lahore team's technology powers more than 10 million brand conversations a week. The company's own announcement names the Bay Area and New York as the places where the new money will hire, and does not mention Lahore.
Nectar is headquartered in the United States and is not a Pakistani company in the legal sense. It belongs in this magazine because part of its engineering is done in Lahore, and because the figures it has published let a reader test the business.
What Nectar sells
Nectar describes its product as an AI native operating system for social, community, creator and commerce work. Its newest piece is Nectar Agent, an autonomous agent that answers customers in a brand's own voice. Brand teams keep control of voice, strategy, permissions and approvals, and can move between letting the agent act alone and reviewing what it proposes.
The company groups the product into three jobs. The first is community management: responding to comments and messages, moderating, routing and resolving them. The second is social intelligence: listening across comments, direct messages, video, creator content and conversations that never tag the brand, to find sentiment, trends and brand safety risks. The third is commerce attribution, which links a social interaction to a customer profile, a purchase intent and eventually a sale.
In outline, a platform delivers each comment or message to Nectar through its programming interfaces, the software classifies it, answers it or hands it to a person, and records what the customer did next. That makes platform access the foundation of the business, because a platform that narrows what outside software may read or post narrows the product. Nectar says it has official data partnerships with Meta, TikTok, LinkedIn, Reddit and X. Its named customers are e.l.f. Beauty, Babylist, Figma, Graza, Liquid Death, Kosas and Caraway. The announcement gives no pricing.
The round was led by Menlo Ventures through its Anthology Fund, which was created with Anthropic. True Ventures, GV and Kinship Ventures also took part. No valuation or earlier funding was disclosed.
Reading the traction numbers
The company reports more than 10 million conversations a week, up 5 times in three months, more than 50 million consumers engaged to date, $100 million in revenue attributed to social, and a share of more than 80 percent of its customers' brand social interactions. All of these are the company's own figures.
The growth claim can be turned into a rate. Five times over 13 weeks is a weekly compound rate of about 13 percent, since 5 raised to the power of one thirteenth is about 1.13. It also means volume three months earlier was around 2 million a week, so the base was small. Growth at that pace cannot last: 13 percent a week for a full year would multiply volume by roughly 600. The more useful number is the annual run rate at today's level, which is 10 million times 52, or 520 million conversations a year.
The attribution figure needs a careful reading. $100 million divided by 50 million consumers is $2 of attributed sales for each consumer engaged. That is sales the customers link to social conversations, not Nectar's revenue. Nectar's fee is some fraction of it, and the fraction is not disclosed. The 80 percent claim describes how much of a customer's social traffic the software handles, not how much of the market Nectar holds.
The cost of every reply
Autonomous replies shift cost from salaries to computing. Every conversation the agent reads or answers consumes model processing and infrastructure, so cost scales with volume rather than with the number of customers. Nobody outside the company knows what it is per conversation, and the announcement does not say which models Nectar runs on.
A simple sum shows why it matters. At 10 million conversations a week, every cent of cost per conversation is $100,000 a week, or $5.2 million a year, which is 17 percent of the $30 million Series A. At a tenth of a cent the annual bill is $520,000. These are illustrations, not reported costs. The point is that a fivefold rise in volume multiplies the bill fivefold unless Nectar charges by the conversation, or the cost per conversation falls as fast as volume rises. Whether it prices per seat, per brand or per conversation decides if growth improves its margin or squeezes it.
The approval controls the company describes carry a cost of their own. A brand that reviews every proposed reply pays for human time on top of the model, and a brand that lets the agent act alone takes on the risk of a wrong public answer in its own voice. The mix customers choose determines how much of the 10 million conversations a week need a person.
What Lahore does and does not get
The ProPakistani interview, published on 15 July, says engineers in Lahore work on the same products, the same codebase and the same production systems as colleagues in Silicon Valley and New York. It adds that senior engineering and product leadership is concentrated in the Bay Area, which Uraizee describes as intentional, to develop early career engineers into future leaders. No source gives a headcount for Lahore.
Uraizee named Careem and Motive as other companies that have invested in Pakistani engineering. Her policy asks, as TechJuice reported them, are steady regulation, stronger intellectual property protection and tax policies that keep engineers in the country. TechJuice also reports that the company "never set out with a Pakistan strategy," which fits what the numbers suggest: Lahore is where engineers were found, not where the product or the customers are.
The split matters for who gains. Lahore engineers get production responsibility on a product with named international customers, and a route to leadership that the company says it is building deliberately. The equity, the customers and the next wave of hiring sit in California and New York. The Series A is also absent from Pakistani funding tallies: the i2i Ventures review of the first half of 2026 counts $59 million raised by five Pakistani startups and does not list Nectar.
What platform access and cost decide
Revenue, pricing, valuation, customer count and the size of the Lahore team are not public. Two risks follow from the model. The product depends on platform access, and a change in the terms of any of the five partners would affect it directly. Autonomous replies in a brand's voice also carry reputational risk at scale, and the approval controls are what customers will judge Nectar on.
The announcement says the money will go to engineering and applied AI hiring in the Bay Area and New York. TechJuice reports Uraizee as saying that money alone will not build Pakistan's AI ecosystem, which is also a fair summary of what the Series A does for Lahore.





