With backing through a16z Speedrun and angel investors associated with OpenAI, Anthropic, DeepMind, KKR, and other companies, Tarun Odedra is building Desk One around an end-to-end approach to sales and marketing for private equity portfolio companies. 
Technology

Tarun Odedra Is Building Desk One Around a Different Model for Running Growth

Written By : Arundhati Kumar

When Tarun Odedra and his co-founders enter a company, they are not starting with the assumption that another piece of software will solve its growth problems. Desk One, the company they are building, works with private equity firms and their portfolio companies to run sales and marketing functions directly. 

Tarun’s premise is that increasingly capable AI models have changed what small teams can execute,  Which means that the velocity at which things are shipped is considerably higher. That makes the toughest and the final frontier that needs to be solved: distribution.

That distinction has shaped Desk One from the beginning. Tarun chose to approach growth across the broader customer-acquisition process instead of building around one isolated part of sales or marketing. His team works inside each business to understand its customers and existing workflows, then builds systems designed around that specific context. The goal is to leave the company with agents it owns while Desk One’s operators remain accountable for the commercial work being performed.

“The models are already capable of doing far more than most companies are asking them to do,” Odedra says. “The difficult part is everything around the model. What information does it have? What decisions can it make? Where does a person need to step in? Those questions determine whether the system is actually useful.”

Desk One has attracted backing through a16z Speedrun, along with angel investors associated with OpenAI, Anthropic, DeepMind, KKR, Citi, ElevenLabs, Brex, Scaling Europe, and Magic. [Tarun, when was Desk One accepted into a16z Speedrun, and was there a disclosed investment amount?]

The company’s model reflects a broader change in how AI is entering commercial work. Many software companies are automating individual functions such as prospecting, content creation, or research. 

Odedra made a different strategic choice by building Desk One around responsibility for a broader portion of the growth operation, using AI as part of the execution layer while keeping human operators accountable for performance.

That creates a different standard for success. A tool can be evaluated according to whether it completes the task it was designed to perform. 

An operator responsible for growth eventually has to answer whether the work improved commercial performance. Odedra argues that this is where the value of the model will ultimately be judged.

“Companies do not ultimately want more activity,” Odedra says. “They want somebody to take responsibility for getting useful work done and improving the outcome. 

AI changes the economics of how that work can be delivered, but it does not remove the need for judgment or accountability.”

Desk One begins by learning how a portfolio company currently sells and markets its products. The systems it creates need context about the business, its customers, positioning, and what is happening in actual sales conversations. 

Odedra argues that even a strong underlying model becomes far less useful when it lacks the information necessary to understand the company it is representing.

That problem becomes especially important in outward-facing work. A software failure inside an internal process can sometimes be identified and corrected before a customer sees it. A poorly written email, inaccurate marketing claim, or inappropriate customer response is already external once it has been sent. Odedra has therefore made human responsibility part of the operating model instead of treating automation as a reason to remove it.

“There is a major difference between showing that an agent can generate something and trusting it to represent a company,” Odedra says. “The more customer-facing the work becomes, the more important it is to be clear about context, permissions, and who owns the final result.”

Desk One currently starts with content and expands into other GTM needs as its team learns where it can contribute inside the business. Its larger operating model is designed to serve multiple companies across a private equity portfolio while adapting execution to the customers, market, and commercial situation of each one.

[Tarun, how many private equity firms and portfolio companies has Desk One worked with to date?]

The commercial evidence will ultimately determine how far that model can travel. Odedra says Desk One is designed to grow revenue faster than the human effort required to service it, which makes the relationship between delivery quality and labor particularly important.

[Tarun, please provide one measurable, supportable result Desk One has produced for a client so far, such as qualified pipeline generated, revenue influenced, meetings booked, conversion improvement, cost reduction, or another commercial result. Please also explain what you personally decided, built, or changed that contributed to that result.]

Odedra’s approach draws on experience across several sales environments, although Desk One presents a different operating challenge. He previously worked in AlphaSense’s strategic accounts organization and later built his own GTM company before running GTM at another venture-backed startup. Those experiences exposed him to enterprise sales, founder-led growth, and the pressure of building commercial systems inside young companies. Desk One carries those lessons into a model where the operator is expected to execute and remain accountable for the result.

That matters because AI has made many of the visible outputs associated with sales and marketing easier to produce. Companies can generate more content, conduct more research, and automate more outreach than they could a few years ago. Odedra’s model puts the emphasis on whether those activities are informed by real customer evidence and connected to a commercial objective.

Desk One is already structured around that principle. Odedra has built the company to combine company-specific context, AI-supported execution, and human accountability across a broader portion of the growth function. Its backing gives the team the opportunity to test that model across portfolio companies, while the next measure of its strength will come from the commercial results those systems produce.

Best Crypto Apps with Fast KYC Verification in 2026: 7 Compared

Mudrex vs CoinSwitch: Fees, Features and Crypto Trading Options Compared

Top 10 Crypto Tax Software Platforms in 2026

Bybit raises USDe rewards to 6.25%, but trading risks persist

Crypto News Today: BTC Hits Four-Month Low, ETH Upgrade May be Pushed to 2027