About Harper Edison
Harper Edison was created to remove the friction and frustration of ad-hoc, everyday phone calls. Despite decades of digital transformation, many essential personal and administrative tasks, such as scheduling appointments, resolving issues, and following up with businesses, still depend on outbound calls. These calls are often time-consuming and unpredictable, requiring long wait times, repeated explanations, and attention during business hours that rarely align with busy schedules. Individually, these tasks seem small, but together they impose a persistent tax on focus, productivity, and mental energy. We believe this is a problem worth solving well and Harper Edison exists to remove that tax.
Our philosophy
Human-centered assistance
Harper is designed to work in the real world, where situations are nuanced, requirements change, and judgment matters.
Delegation over interaction
We believe the most useful role for AI is not constant interaction, but responsibility. Harper Edison lets users hand off real-world tasks entirely, rather than manage tools, prompts, or workflows themselves.
Control and accountability
Harper operates with clear intent and visibility. Calls are documented, outcomes are summarized, and users remain in control of next steps.
Built for trust and transparency
Harper identifies itself when making calls.
Users approve the context and instructions before calls are placed.
Every call produces a status update with a structured clear summary.
Users can review, correct, or revoke information at any time.
We believe trust is earned through transparency, auditability, and consistent behavior.
Harper's origin
In August 2025, Harper Edison’s co-founders, Brett Paden and Michael Gersh, were on a daily zoom for another project when Brett described a week-long struggle to resolve a denied health insurance claim. The process was painfully familiar: calling the insurance company, waiting on hold, getting no resolution, calling doctors, following up, and repeating the cycle over and over. The mental strain of navigating the system had become more exhausting than the slipped disc that required the medical care in the first place.
Half-jokingly, Michael said, “We should spin up an AI to handle those calls.” Almost immediately, it became clear that it wasn’t a joke at all.
As they began digging into the space, a gap quickly emerged. There were plenty of AI telephony products designed for call centers or for answering inbound calls on behalf of a business. There were also tools optimized for scripted, repetitive outbound sales or lead-generation calls. But there was nothing capable of reliably handling ad-hoc calls — calls initiated by a random user, to a random phone number, with a unique objective each time.
When they started speaking with potential users, the response was immediate and universal. Everyone had a story. People dread calling insurance companies, banks, airlines, contractors, and utilities. Simple calls are endlessly postponed because they’re frustrating, time-consuming, and unpredictable. The pain was obvious and widespread.
Harper Edison was incorporated in September 2025 to build a solution to that problem. We are starting with a service because we believe real-world delegation cannot be designed in theory. By working directly with users, we learn what people actually delegate, where automation fails, where judgment is required, and how trust is built and lost. This service-first approach allows us to build systems that are grounded in real behavior, not idealized workflows. It also allows us to define a new category: ad-hoc telephony — outbound calls for unpredictable, one-off, real-world tasks.
Our plan
We are starting with a service because we believe real-world delegation cannot be designed in theory. By working directly with users, we learn what people actually want to delegate, where automation breaks down, where judgment is required, and how trust is built and lost. This service-first approach allows us to both define a new category- Ad-hoc telephony - outbound calls for unpredictable, one-off, real-world tasks, and to build a platform that others can rely on. Over time, we plan to make this capability available to companies and organizations that need to initiate phone calls on behalf of their users or customers, including AI agents, consumer applications, enterprise software, and service providers that cannot justify building or operating their own voice workflows. Our long-term goal is to become the default ad hoc voice telephony layer.
We are building Harper Edison to help people get more of their time back, while defining a new, responsible model for real-world AI assistance.