The first useful business Pouyan Salehi helped build delivered in thirty minutes or less. His family had arrived in Minneapolis from Iran during the Iran-Iraq war with two suitcases, about $500 and the kind of reset that makes a résumé seem ornamental. Both parents had been engineers. His father washed dishes, delivered pizzas and became a Domino’s manager. Eventually the family bought a franchise that was struggling, then turned it into a strong store. Salehi began working there around age nine. By his mid-teens, he was managing it.
He likes to joke that the shop taught him more about product and business than Harvard Business School. The line works because it carries a serious operating idea inside it. A customer did not simply order pepperoni and cheese. The customer might be a parent home late, buying an evening without cooking. The product was pizza. The value was convenience. Once you learn to see the second thing, every workflow starts confessing what it is really for.
From mechanical systems to human ones
Salehi studied mechanical engineering at the University of Minnesota, then earned an MBA at Harvard. He joined Apple in operations after business school. It was an excellent place to learn how a company with exacting standards moves, and an equally good place to discover that he did not want to remain inside a large organization. In 2010, he left to co-found StackMob, a mobile backend startup later acquired by PayPal.
Then came Lera Labs in 2012, a period of prolific construction that he recalls without polishing every experiment into a triumph. Plenty of ideas never saw daylight. Yet each attempt sharpened the question behind the next one. While trying to find customers for one product, Salehi kept an Excel sheet with roughly 2,500 prospects and hacked together daily mail merges. The indignity was productive. He and Cyrus Karbassiyoon built PersistIQ around the pain of outbound prospecting.
Salehi and Karbassiyoon had met years earlier at a Bay Area technology event. They bonded over design and a shared Iranian connection, kept meeting for tea, and became repeat co-founders. Their partnership is one of the quiet continuities in Salehi’s story: products change, but the people who know how you think become part of the machinery.
“My belief now is try to actually sell before you build.”Pouyan Salehi
PersistIQ joined Y Combinator in 2014 and reached a million dollars in revenue quickly and with relatively little capital. The clean retelling would proceed directly to an exit. Salehi’s version lingers on the awkward middle. Timing was slightly off. Competitors looked similar to buyers. Winning could become a contest of marketing budgets, and the venture-scale route did not feel like the right one. The team was no longer the fashionable company in the room. It kept building anyway.
Mobile infrastructure
Experiment widely
Outbound workflow
Revenue workspace
A spreadsheet was trying to tell them something
PersistIQ was acquired by Wishpond in 2019. That year, Salehi and Karbassiyoon founded Scratchpad after noticing a contradiction in the sales stack. Companies paid handsomely for Salesforce. Their account executives did much of the real work somewhere else. Notes lived in Evernote or Apple Notes. Pipeline thinking happened in spreadsheets. Tasks wandered between tools. Then, shortly before a forecast call, the rep spent hours reconstructing the official account for management.
Most people saw poor compliance. Salehi saw an interface problem. The unofficial tools were not evidence that sellers were careless; they were evidence that the formal system did not fit the work. He and Karbassiyoon sat with salespeople, observed their routines and mapped the repeated bits of friction. No thunderbolt arrived. Patterns did.
Scratchpad began as a faster workspace on top of Salesforce, designed for the person entering the information rather than only the manager consuming the report. The company’s early test was charmingly severe: find ten people in three months who loved the product and behaved like it. They found one. One out of ten is a poor batting average and, in a young company, potentially magnificent evidence. That person cared so much that the founders gave themselves another three months.
Early conviction does not always arrive as a crowd. Sometimes it is one user whose behavior changes enough to justify the next experiment.
The approach explains Salehi’s insistence on selling before building. He is not arguing against craft. He is arguing against a year spent answering an imaginary question. A sales conversation makes a founder compress the promise, encounter objections and hear the vocabulary customers use when they are not trying to flatter a product team. His “mom test” is simple: can he explain Scratchpad to his mother in a way that lands and produces a real signal?
The human unit hiding beneath productivity
Scratchpad announced a $13 million Series A in 2021 and a $33 million Series B in 2022, with Craft Ventures leading both rounds. Funding is the conventional milestone. Salehi’s favorite evidence tends to be more intimate. One user had blocked the final hour of each workday to update Salesforce, then told him that Scratchpad made the updates happen in real time. The hour came back. She could have dinner with her daughter.
Salehi, a father of three, understood the unit immediately. Software companies count active users, retention and revenue. People count evenings. The distance between those measurements is where much enterprise software becomes absurd. A product can improve a dashboard while making the person feeding it miserable. Scratchpad’s original proposition was that the two outcomes did not need to be enemies.
His public manner matches that product instinct. He is plainspoken about missed targets and failed experiments, alert to the comedy of corporate rituals, and willing to let a co-host tease him on the Beyond Quota podcast. The humor rarely drifts far from the work. Sales is full of rejection, status anxiety and manufactured urgency. A little levity is not decoration. It is ergonomic.
“There were no flashes of brilliance. Instead, there were a series of observations that eventually led to identifying patterns.”Pouyan Salehi
After making the record easier, make it whole
By 2026, the software industry had acquired a new abundance. Capable models were everywhere; trustworthy company context was not. A customer existed as a row in Salesforce, a voice in a call transcript, an email thread, a calendar invitation and a Slack discussion. An AI asked to prepare a seller for the next meeting might see one fragment and answer with serene confidence.
Salehi and Karbassiyoon’s next layer is Clearskies, operated by Scratchpad and described as a context layer for revenue AI. It resolves those pieces into a connected, current account history that can be used across models and tools. Salehi frames the context in four parts: the customer, the team, the sales process and the business itself. Models do not arrive knowing any of them.
History and risk
Roles and results
How selling works
Priorities and rules
It is a natural deepening of the same career-long argument. PersistIQ reduced the manual work of finding and contacting prospects. Scratchpad reduced the manual work of keeping the CRM current. Clearskies tries to reduce the repeated work of gathering and reconciling context before an AI can reason. The layer may be less visible than the agent on top, but plumbing has always been where serious software earns its keep.
Salehi also introduced NexusGTM in 2026, a four-week service aimed at helping revenue organizations embed AI into how they operate. The premise is blunt: buying access to a model is not organizational transformation. Teams still have to choose useful cases, shape their data, evaluate results, create feedback loops and redesign routines. Technology supplies possibility. Operators turn possibility into Tuesday morning.
Pay attention to what people have learned to tolerate
A founder’s biography invites a false straight line. Iran to Minneapolis. Pizza shop to engineering. Harvard to Apple. StackMob to Lera Labs to PersistIQ to Scratchpad. Each logo begins to look inevitable from the safety of the next one. Salehi’s own accounts resist that convenience. The transition from PersistIQ was messy. Scratchpad’s first test missed. The central problem did not reveal itself to two founders merely because they had spent a decade in sales technology.
They had to watch. They had to discard the authority of experience long enough to notice salespeople building private systems beside the official one. They had to interpret avoidance as design feedback. They had to let one enthusiastic user outweigh nine indifferent ones for just long enough to learn more.
The pizza-shop education survives because it was never really about pizza. Look at the long day behind the order. Look at the forecast meeting behind the frantic CRM update. Look at the missing customer history behind the AI’s polished answer. The most interesting work is often one layer below the product everyone can see.
Salehi’s aspiration for years was to build the workspace a revenue team would expect to find when it joined a company. The language has evolved from workspace to brain, from easier updates to persistent context. The instinct has not. Give people back the time that bureaucracy quietly borrowed. Make the useful behavior the easy behavior. Treat a workaround as a note from the future, written by the person who needs the product most.