Hone raises from Benchmark and Index to build AI 'engines' that own business outcomes, not just automate tasks

Hone is a seed-stage AI company arguing that the real ROI problem in enterprise AI isn't engineering — it's everything else. Founder Moritz Stephan says AI adoption outside of software teams has been largely disappointing, and Hone is built around a specific diagnosis: most organizations are using agents to hand off individual tasks, micromanage outputs, and stitch together workflows that don't clearly map to business outcomes. Hone's answer is to put AI in charge of outcomes directly.

The company calls its products "engines" — autonomous systems configured to own a specific business result rather than execute a sequence of steps. Current focus areas include go-to-market, recruiting, financial risk, and procurement. Stephan says the engines are designed to self-adapt to an organization's context, allowing non-technical staff to provide feedback and refine them without dedicated AI implementation headcount.

“Moritz Stephan: 'The ROI of AI outside of engineering today is quite disappointing. That's what we're here to solve.' On the model: 'We build engines — AI that owns business outcomes. Put AI in charge of outcomes, dial up the autonomy.' The round was co-led by Benchmark and Index Ventures, with GIL, Definition, Diffusion, and others participating.”

Early commercial focus

The clearest beachhead Stephan describes is revenue — specifically account management. The pattern he highlights is staying on top of signals across a company's customer relationships and acting on them automatically. He says this shape of use case is common across traditional GPM, banking, and insurance, making it a practical landing zone for early deployments.

Deployment model

Rather than building bespoke solutions for each customer, Hone is developing template engines that can be deployed across organizations with minimal setup. Stephan frames the forward-deployed engineering work as a source of learning — the hands-on time with complex use cases feeds back into the templates. Over time, he expects the outer loop of that process (gathering requirements, scoping use cases, defining success) to be largely automatable, with human effort concentrated on relationships and prioritization.

Pricing

Hone is exploring a budget-based pricing model: rather than billing by task or seat, give the engine a fixed budget and let it determine how to achieve the outcome within those constraints. The engine would make its own trade-offs — how much context to pull in, how thoroughly to branch — based on what the budget allows. Stephan cites Sierra as a reference point for outcome-based pricing but acknowledges the heterogeneity of Hone's use cases makes direct replication difficult.

Build vs. buy on models

Stephan takes a pragmatic line on model strategy: if a capability gap is likely to be solved by frontier labs within a few months, investing heavily in custom training now is probably a mistake. Hone has researchers and engineers on the team but says it doesn't have a strong prior on whether to fine-tune or operate purely at the harness level — the goal is driving the outcome, and the method follows from that.

Funding

The seed round was co-led by Benchmark and Index Ventures, with GIL, Definition, Diffusion, and others participating. No round size was disclosed.

Stephan previously worked at Cognition, where migrations and refactors became a durable commercial anchor. His bet at Hone is that account management and revenue growth can play a similar anchoring role — universal enough in shape to scale, complex enough in execution to resist commoditization.