← All projects
Concept for Abundant · by Anirudh Devella

Route every task to the right expert

Abundant puts human experts on AI lab tasks. The hardest tasks should go to the people who are best at that exact kind of work, and there should be enough of them before the demand shows up. Pick a task and see who gets it, and why.

Fictional example. The people, numbers and events are made up to show the method. This is an outside concept, not Abundant's product or data, and not affiliated with Abundant.

Built for Abundant · the short version

Labs want the right expert on every task, and enough experts lined up before the demand arrives.

In this sample, two domains run short within three weeks, and the first available expert is rarely the best fit.

The data is made up. To make it real, I'd need task history with reviewer acceptance and rework, expert availability, and lab order forecasts.

Who built it

Anirudh Devella, data scientist. Before this, in production:

  • 8.3% forecast error at Bluelight (YC)
  • $14M+ in invoices benchmarked at Bluelight
  • AI Coach for 10,000+ athletes at Brandlete
  • 150K leads scored at CoverRight

“I’d hire him again.” Alex Dixon, Co-founder, Bluelight (YC), managed Anirudh directly. Read it on LinkedIn

01 · Route a task

Pick an incoming task. See who gets it, and why.

Experts are ranked inside their real specialty, using how their past work on similar tasks was accepted, how often it needed rework, and whether they have time this week. The first available person is rarely the best one.

02 · Specialty, not job title

Clusters come from the work, not the résumé

Two experts with the same title can be strong at very different things. Grouping experts by the tasks they actually do well splits each domain into specialties, and ranking happens inside each one.

03 · Enough experts before the demand arrives

The next six weeks, by domain

Forecast demand divided by available expert hours. Above 100% means labs wait or tasks go to a weaker fit. Vetting new experts takes weeks, so the call has to be made now.

Next step

My first week with your real data

  1. Score every expert inside their specialty on accepted work and rework.
  2. Forecast expert hours by domain six weeks out.
  3. Route one domain's tasks by score for a week and compare acceptance.