Adam Jafer spent years at Voi watching spreadsheets hold together a fast-scaling micromobility operation. The technical co-founder eventually decided there had to be a better way—and that building it would require starting over. So he left in December 2025.
What emerged on May 7, 2026, was Pit.com, an AI platform designed to observe how companies actually operate, then build custom internal software to replace the Rube Goldberg machines of Excel files, email threads, and inflexible SaaS tools that most enterprises still run on. The Stockholm-based startup announced it had raised $16 million, led by Andreessen Horowitz.
The round was led by a16z general partner Alex Rampell and included Lakestar, the founders themselves, and a constellation of angel investors from OpenAI, Anthropic, Google, Deel, and Revolut, plus backing from Sweden's Stena and Lundin industrial families. European outlets pegged the figure at roughly €13.6 million. Pit didn't formally designate a stage in its press materials, though some observers called it a seed round.
The Pitch: AI That Builds Your Operations Software
Pit describes itself as an "AI product team as a service." The platform watches workflows inside a company, then assembles operational systems tailored to those specific processes. Crucially, the company insists its output isn't prototypes or templates—it's production-grade software with structured workflows, monitoring, and rollback capabilities built in.
The target is that messy operational middle layer where most companies—even large, sophisticated ones—still depend on spreadsheets, shared inboxes, and manual handoffs between disconnected systems. Pit Cloud, the infrastructure layer underneath, handles tenant isolation, single sign-on, role-based access, and full audit trails. The company holds ISO 27001 certification and says it's built for GDPR, NIS2, and the EU AI Act, with data residency options for regulated sectors.
Technically, the system is model-agnostic, choosing what Pit describes as the best model for each specific task. For customers in regulated industries, EU-hosted models are available. The system integrates with existing tools rather than demanding migration. According to company materials, deployments typically go live within days or weeks—a claim that, if true, would set it apart in a market where enterprise software rollouts often drag on for months.
A Team Pulled From Swedish Tech
Jafer assembled a group drawn heavily from Swedish companies known for operational complexity at scale. Two founding engineers came from Klarna: Anton Öberg and Fredrik Olovsson, both former Senior Principal Engineers. The company's materials describe the founding team as "founders and CTO/AI leads behind Voi, Klarna and iZettle," though the full co-founder roster isn't spelled out explicitly on the website. (One outlet named five founders; the site's team imagery shows Andreas Hjelm among the founding team. Reporting variance, perhaps, or quiet team changes.)
A LinkedIn post from late April, written by one of the founding engineers, referred to Pit as a "13-person startup." By launch, the company size was listed at 11 to 50 employees. The company says it's designed in Sweden but lists San Francisco as headquarters—standard for a Delaware-incorporated company with European engineering roots.
Early Metrics and Named Customers
Pit launched with four named customers: Voi, Tre (3 Sweden), Stena Recycling, and Kry. The company claims these early deployments delivered an 85% reduction in campaign execution time, with typical engagements saving more than 10,000 hours annually. One large industrial customer—likely Stena Recycling, given context—reportedly saved over 10,000 hours per year with zero validation errors. Invoice automation for another customer hit 99% acceptance rates.
Those numbers, if they hold up under scrutiny, suggest Pit is tackling genuinely high-volume manual work: contract processing, campaign management, cross-system data reconciliation, approval workflows. The kind of operational busywork that scales poorly in mid-to-large enterprises and drives operations teams quietly insane.
Crowded Territory
"Replace your spreadsheets" isn't exactly a fresh pitch in enterprise software. ServiceNow and UiPath have built massive businesses around workflow automation. Low-code and no-code platforms promise faster internal tool development. More recently, competitors like Equals have positioned themselves as "AI-native spreadsheets," while vertical-specific entrants—Firstshift in supply chain, Conduit in dock and yard operations—are making similar bets on eliminating manual processes.
Where Pit claims differentiation is in producing fully deployed, tailored systems rather than handing teams tools to build their own. Rampell's quote in the launch materials framed it as "selling speed that holds up for years," calling it "a new category." Jafer's positioning emphasized software that adapts to how a company operates, not the other way around.
The timing tracks with broader industry currents. A February PitchBook analyst note titled "SaaS is Dead, Long Live SaS" described a shift toward outcome-based, agentic AI systems. An April analysis argued that Excel and Google Sheets might remain the dominant enterprise AI interfaces—ironic backdrop for a company promising to eliminate them. Maybe there's something to that irony. Or maybe Pit's bet is that companies are finally ready to move on.
The Path Ahead
Pit hasn't disclosed pricing or its commercial model beyond the "AI product team as a service" framing. The target buyers are operations, finance, and customer workflow teams at mid-to-large enterprises—companies with enough manual process volume to justify custom software, but perhaps not enough engineering bandwidth to build and maintain it in-house.
The company's emphasis on governance and security likely anticipates procurement conversations with risk-averse buyers. The EU compliance posture and data residency options signal awareness that regulated industries move slowly and care deeply about where data lives and who touches it.
Whether Pit can scale this approach—moving from a 13-person team to something resembling a repeatable product company—depends on how much of the "learning" process can be standardized and how much remains bespoke, human-intensive consulting work. The "AI builds your software" promise is compelling. Delivering it repeatedly, reliably, and profitably is considerably harder.
For now, the company has customer logos, tangible metrics, and a16z's backing. That buys time—and perhaps more importantly, credibility—to prove the category isn't just hype.
Whether enterprises are truly ready to let AI rebuild their operations infrastructure remains an open question. But if anyone's tired enough of spreadsheets to find out, Jafer seems to be counting on it.
