Where Apres came from

When I talked to candidates in late 2018, I used a call script. Partway through, after asking about them, it gave my own background in a few lines. The part that matters is this:

In that time, scaling the operational arm of the company became the bottleneck to the business. It surfaced a larger industry problem that led me to transition to Apres at the end of last year.

My recruiting script, November 2018

“That time” was Fluid Financial. I co-founded it and ran product as CPO for three and a half years, from Oakland. The bio I kept in January 2019 says the software handled millions of dollars in monthly transactions for hundreds of business clients. What didn’t keep up, by my own account a year later, was the operational work around it.

Before Fluid

That same bio lists the jobs before it in one sentence: design for a product agency, FTW Group; sales processes for Uversity; and production budgeting software for filmmakers at Tentpole, which I founded. Before that, in 2011, I started Short Cinema and produced the pilot of a mobile series, “The Mondays,” with more than 25 independent creatives, paid for with brand partnerships.

The first Apres

The earliest Apres deck in my notes was last edited in May 2018. Its tagline was “API For Human Automation.” Under “What are you trying to do?” it said: “A simple way to automate repeated tasks using humans and machines.” The pitch draft from the same month said: “We believe the future of work is autonomous and are building Apres to remove the need to run your own operations team.”

The deck also had a line for risks. There was one entry: “Getting the right experts in the app.”

That risk runs through the next year of notes. By January 2019 my notes include a survey for remote workers and a list of places to find them: job sites, paid-survey sites, data labeling firms. The survey asked what people liked and disliked about remote work, what their expertise was (web research, image annotation, personal assistance, finance, tech) and what they earned an hour, in two buckets: 1 to 10, or 10 and up.

Borrowing a sentence

My January 2019 pitch note includes a pitch marked “Example,” about operations teams at fast-growing companies. Its key line was that 100% growth in headcount a year means half your operations team is new to the job, “and this fact never changes.”

Under it, I wrote the Apres version: “Integrating AI technology means that some percentage of data needs to be transformed and this fact never changes.” Same sentence, new noun. By then Apres was about the data AI systems need, and the human work behind it.

What I take from it

Start from the bottleneck you have lived. I didn’t need a market report to describe this one. My own line from 2018 says it: the operational arm became the bottleneck.

And notice what the product keeps asking, even when the pitch changes. In 2018 the question was how to get repeated work done well by people you can’t watch. In 2019 it was how to know the training data was right. At Apres in 2021 it was how to explain what a model did. At Excellent it is how to know an agent did what it said. Four pitches over eight years, and one question underneath: how do you know the work is right?