Tony Holdstock-Brown, CEO of Inngest: orchestration, traction and not using LinkedIn
Tony Holdstock-Brown has largely rejected social media even though founder orthodoxy says to build an audience. The Inngest CEO explains the distribution experiments that work instead, why positioning must change with the customer, and how higher-level infrastructure primitives become more valuable in an AI-driven world.
Tony Holdstock-Brown is the co-founder and CEO of Inngest, a platform for running durable backend and AI workflows without making developers manage queues, state and orchestration infrastructure.
We talk about applying engineering sprints to marketing, borrowing other people's distribution, earning trust for infrastructure products, moving from product-led growth to enterprise customers, choosing the right level of abstraction, and keeping cool technology subordinate to a real user problem.
The Q&A below is experimental — extracted from the episode transcript with AI. If you spot a mistake, let us know at jack@scalingdevtools.com.
Why have you opted out of most social media as a founder?
Social media destroys my focus, and there is already more than enough inside the company that needs attention. I do not actively use Twitter, Instagram or LinkedIn; my co-founder and other teammates are better suited to that work. It is atypical and may change, but dividing the role according to people's strengths lets me spend more time on product, strategy and execution.
How can a company gain reach without the founder building a personal audience?
We partner with people who have already built relevant audiences. Their overlap with our existing community may be small, so they can introduce Inngest to net-new groups. It does not create the same long-term ownership as a founder's audience, but it produces useful reach now and lets the company learn which messages and examples resonate.
How do you apply the Traction framework to marketing?
We translate an unfamiliar discipline into an engineering cadence. Pick a couple of traction channels, write a marketing spec, define the test and how it will be measured, run it for about two weeks, then decide whether it moved the needle. Repeat across channels such as content, education, YouTube, partnerships, marketplaces, conferences and sponsorships. The winning channels change as the audience and company mature.
What distribution experiments worked early for Inngest?
Technical content helped establish trust because Inngest runs the orchestration and state around a customer's application, which is a meaningful responsibility. YouTube partnerships and code examples also worked because showing a complete thing people can build is more inspiring than listing API methods. An example gives developers both the mental model and the unknown details they would not know to search for.
How does the message change between a startup and a public company?
Both care about developer experience, speed and reliable APIs. A young company can tolerate more risk because speed is part of its advantage. A public company puts much more weight on security, scalability, SLAs, global infrastructure and reducing operational risk. Product-led-growth needs form the base; enterprise requirements are largely a superset added on top.
Why is positioning especially difficult for an infrastructure product?
A high-level API can make an extremely complex system look deceptively simple. Developers who have not yet lived through broken queues, retries and distributed workflows may not recognize the value, while experienced teams immediately see months of work they can avoid. Sometimes the answer is not a different product but better education or a small UX change that explains what failed and why. We made our language more technical so the right users could self-select.
What problem did your own healthcare company reveal?
Patient and doctor processes contain long waits, conditions and obligations: if something has not happened after several days, another action must occur. We implemented those flows with queues and state spread across services, which made them slow to build and painful to debug. Inngest turns those facts into events and lets code wait for another event or a timeout, while preserving an audit trail of what happened.
Why do higher-level infrastructure abstractions become more important with AI coding?
An agent can generate a few Inngest functions with durable steps and deploy them alongside the application without first creating Redis, queues, workers and Terraform. Lower-level infrastructure demands knowledge the model and developer may not have. A good abstraction removes that setup while preserving the behavior the product needs, much as React provides a useful layer over UI state and the DOM.
How do you choose the right level of abstraction for a platform?
Cloudflare's security wedge funded a global network that later supported Workers and other products. Inngest sits above static queue and event-stream infrastructure: it understands typed events, function steps and execution state while still letting code run on different hosting providers. That awareness creates observability and future products, but the first wedge — orchestration — must become excellent before the platform expands.
What is your most important advice for DevTools founders?
Think about distribution from first principles: who the customer is, where they spend time, how you reach them, what story you tell and which part of the funnel is slowing growth. Then keep returning to the user's problem. Engineers love technology for its own sake, but impressive implementation is a trap if it does not remove real friction for the person using it.
Links & references
This episode is brought to you by WorkOS.