How a Tiny Team Ships Like a Big One · The Product Track

We automated the AWS partner-funding maze

AWS has 80+ partner funding programs, and navigating them is a part-time job nobody has. We wired the APIs so every CRM deal checks its own eligibility - and found ~$3.2M for one customer.

5 min readawspartnerautomationproduct

Somewhere inside every AWS partner is a spreadsheet, lovingly maintained by whoever drew the short straw, tracking which deals might qualify for which of AWS's partner funding programs. There are more than eighty of these programs. Their eligibility rules differ by partner tier, workload, customer segment, and deal stage; they change; and the money is real - real enough that leaving it unclaimed is leaving margin on the table, and real enough that AWS genuinely wants partners to use it.

And still it goes unclaimed, because the job of checking is miserable: for every deal, someone has to know the current program landscape, map the deal's shape onto each program's criteria, and start a process in a portal. Multiply by every deal in a pipeline, and you get the universal answer: nobody checks, except for the biggest deals, sometimes, late.

For a while we were that spreadsheet person for our own business. Then we noticed the whole problem is API-shaped.

Eligibility is a batch job, not a job title

AWS exposes its partner machinery programmatically - Partner Central and the ACE co-sell APIs. So we built the check into the product: every deal in a customer's CRM is automatically evaluated against the funding programs we support, continuously, the way signals get checked against accounts. No one asks "could this deal get funding?" - the deal already knows.

When a deal qualifies, it surfaces with the program, the eligibility reasoning, and the AWS account-team context attached - and the workflow to act on it: application forms pre-filled from deal data, supporting documents attached, submission routed through the humans who are allowed to submit. The maze doesn't disappear; the walking of it does.

The first serious run of this for one customer found 39 co-sell-eligible opportunities representing roughly $3.2 million in potential AWS funding - sitting in a pipeline that a capable, busy sales team had been working all along. Nothing was hidden. Checking was just nobody's job.

Every deal checks its own funding eligibility · click to enlarge

What made it harder than it sounds

The APIs give you the plumbing; they don't give you the mapping. A CRM opportunity and an AWS program speak different languages - one describes a customer and a dollar amount, the other describes workload types, migration stages, and partner qualifications. The mapping layer - how a deal's shape translates onto a program's criteria - is where the effort went, and where we'll stay vague, because it's also where the value is. What we will say: precompute eligibility natively rather than round-tripping per deal per program - the check must be cheap enough to run on everything, or you're back to spot-checking big deals, which is the spreadsheet with more steps.

Real numbers

  1. 39 co-sell-eligible opportunities ≈ $3.2M in potential AWS funding, surfaced automatically from one customer's existing pipeline.
  2. 80+ AWS partner programs in the wild; 4 supported with enriched eligibility detail so far, chosen by expected value.
  3. Every deal, continuously - not big deals, sometimes, late.
  4. Zero new headcount - which is the entire point.

Where the humans sit

Submissions are role-gated: only designated people can send an application to AWS, and the AWS relationship stays a human relationship - the system prepares, people submit and negotiate. Eligibility findings are recommendations with reasoning attached, and when the reasoning is wrong, that feedback tunes the mapping. Money conversations don't get automated endings here; they get automated beginnings.

Steal this

If you're an AWS partner: the co-sell and funding checks you're doing manually (or skipping) are Partner Central/ACE API calls away from being a batch job - start with the two programs your deal profile hits most, precompute eligibility across the whole pipeline, and only then invest in more programs. And the general shape travels beyond AWS: any partner ecosystem with published criteria and an API is a place where "knowing the programs" should stop being a person's job and start being a property every deal simply has.

On Thursday the partner track publishes this post's companion for humans, a map of what AWS actually funds. Next week the build track turns the agents on ourselves. Then back here: what a bank-grade AI governance review actually asks, and how six builders answer it.

This post is part of the product track of How a Tiny Team Ships Like a Big One, a series on how six builders run a production AI company. Building at Aithon - if this is how you want to work, talk to us.