KnownforAI builds its AI services without building the workflow infrastructure underneath them

At a Glance

• Company: KnownforAI, a team of experienced operators that helps revenue and marketing teams put AI to work
• Industry: AI services for revenue and marketing
• Product: Diagrid Catalyst
• Why they moved: keep the team's time on client work while Catalyst handles workflow reliability

AI Analysis That Runs for Months

KnownforAI brings together operators with backgrounds in revenue, marketing, finance, and data. It trains teams on practical AI, implements AI workflows for revenue operations, and runs AEO analysis for companies that want to show up when buyers ask AI tools for help. AEO, or answer engine optimization, is the work of making sure AI tools like ChatGPT can find and cite a company when a buyer asks a question it should answer.

The AEO analysis is a long-running workflow. KnownforAI maps the buyer questions a client should answer, reviews the client's public pages, and tests those questions against AI tools to set a baseline. The client gets a five-page analysis and a 90-day plan. Then, once a month for six months, the workflow repeats the priority questions against the baseline and reviews AI-referred traffic to measure what changed. One client engagement runs for the better part of a year. KnownforAI runs it on Catalyst.

The Challenge: Production Reliability Was Becoming Its Own Engineering Project

A workflow that runs for months fails in more places than one that runs for seconds. A model call can time out during prompt testing, or an AI tool can rate-limit a batch of questions. The baseline has to survive every restart for six months, because every monthly check-in compares against it.

Building that in-house means writing retry and recovery code, storing state somewhere durable, and running a scheduler to fire each monthly check-in on time. For a small team whose value is operating expertise, every hour spent on that infrastructure comes out of client work.

KnownforAI needed production-grade workflow infrastructure without turning its application code into an orchestration system, the code that coordinates a workflow's steps and decides what happens when one fails.

Why Diagrid Catalyst

KnownforAI needed durable execution it did not have to build. Durable execution saves each completed step of a workflow so the workflow can resume from that point after a failure or a long wait. Diagrid Catalyst provides it as a platform. KnownforAI's code describes the workflow, and Catalyst keeps it running.

Durable Execution Without Building It In-House

Catalyst saves workflow state as each step completes, including the baseline every check-in depends on. After a failure or restart, the workflow continues from the last completed step, and KnownforAI does not write its own code to save and restore progress.

Retries and Recovery Handled by the Platform

When a model call fails or an AI tool rate-limits a request, Catalyst retries the step and keeps the answers already collected. A failed prompt test does not rerun the whole analysis.

Less Infrastructure to Own

Catalyst timers fire each monthly check-in, so KnownforAI does not run cron jobs or a separate scheduler. A cron job is a scheduled task that runs on a fixed timetable, and it has no memory of what happened in the last run. A Catalyst workflow does.

Engineering Focus on Client Work

With workflow reliability handled by Catalyst, KnownforAI's operators spend their time on buyer questions, client plans, and revenue operations work.

Results

• Six months of monthly check-ins run on Catalyst timers with no scheduler to maintain
• Failed prompt tests retry on their own without rerunning the analysis
• Fewer custom components for retries, state, queues, and recovery
• More of a small team's time spent on client work

Additional Resources

• Diagrid Catalyst: https://www.diagrid.io/catalyst
• Dapr Workflow overview: https://docs.dapr.io/developing-applications/building-blocks/workflow/workflow-overview/