How to Fill Your ACE Pipeline When You Have Nothing to Register

There is a question AWS partners ask each other privately and almost never ask a vendor, because every vendor in the co-sell market answers a different question. It goes something like this: everyone keeps telling me to register opportunities in ACE. I don’t have any opportunities to register. What am I supposed to do?

It is worth sitting with how strange the tooling landscape looks from that position. There is software to sync your CRM to Partner Central. Software to route an opportunity to the right co-sell motion. Software to map your accounts against a partner’s accounts. Software to transact a private offer and meter the usage. Every one of those products is genuinely good at its job, and every one of them begins at the same place: an opportunity that already exists.

Nobody sells the part before that. Which is unfortunate, because for most partners that is the only part that is actually broken.

An empty ACE pipeline is a sourcing problem wearing a co-sell costume

The co-sell motion gets blamed for a failure that happens upstream of it. A partner submits four opportunities in a quarter, two get rejected, one stalls, one closes. The conclusion drawn is that co-sell is not working, or that the PDM relationship needs attention, or that the program requires a specialization the partner does not yet hold.

Sometimes that is true. Far more often, the honest accounting looks like this: the partner submitted four opportunities because four opportunities is what existed. Three came from inbound. One came from an existing customer expanding. None came from a deliberate act of creation. The co-sell motion performed exactly as well as the pipeline feeding it, which is the only thing a co-sell motion can ever do.

Co-sell is an amplifier. Amplifiers do not generate signal. If you put four accidental opportunities into ACE, a perfect co-sell motion returns four amplified accidental opportunities, and the partner concludes that the amplifier is defective.

Why the obvious fixes do not move the number

Faced with an empty pipeline, most partners reach for one of four levers. Each fails in a specific and predictable way, and it is worth naming them because a partner who has tried three of them tends to conclude the problem is unsolvable rather than mis-diagnosed.

Ask the PDM for referrals.This produces AWS-Originated opportunities, which is real pipeline. But AO flow is a selection process the partner does not control, sized by the PDM’s book and the partner’s capability profile within it. A partner who asks for more AO flow is asking a busy person to redistribute a fixed quantity in their favour. It works occasionally. It does not compound, and it resets when the PDM changes role.

Add a competency or specialization. This is the right lever aimed at the wrong track. Competencies improve AO routing — they tell AWS which referrals you should receive. They do very little for Partner-Originated flow, because a PDM evaluating a PO submission is not reading your badge wall. She is asking whether you actually know this account.

Run an ICP filter and submit what comes out. This produces volume, and it is the most expensive failure of the four. A list of companies matching firmographic criteria is not a list of opportunities. Submitting it to ACE converts a pipeline problem into a credibility problem, and credibility with a PDM takes quarters to rebuild.

Hire an SDR, or buy an AI SDR. Closer to correct — at least it accepts that pipeline has to be created rather than requested. But generic outbound into an ecosystem it does not understand produces generic replies, and generic replies do not become opportunities a PDM wants. The mechanism is sound; the input is missing.

What a PDM is actually evaluating when you submit

Everything downstream gets easier once you understand what happens on the other side of a submission. A PDM receiving a Partner-Originated opportunity is making a resource-allocation decision under uncertainty. She has limited hours, a large book of partners, and a quota that depends on deals that actually close.

So she is asking three things, roughly in this order. Is this real? Does this partner know this account, or did it come out of a filter? And is there a reason it is happening now, rather than being an account that has looked like a fit for two years and always will?

Notice that none of those questions are about your capability. Your competencies were relevant when AWS was deciding whether to route you a referral. They are nearly irrelevant here. What a PDM needs from a PO submission is evidence: a named stakeholder who has responded, a specific account situation, and a why-now that is a fact about the account rather than a fact about your quarter.

The thing nobody automates

So the shape of the actual problem becomes clear. A partner does not need help moving opportunities. They need opportunities to exist — and not just any opportunities, but ones carrying enough evidence that a PDM can champion them without doing the partner’s homework.

That is a manufacturing problem, and it has four stages. Each one is skippable, and skipping any of them is what produces the empty pipeline or the rejected submission.

Find the account before it is in anyone’s pipeline. Not a firmographic match — a company whose behaviour in the ecosystem says something changed. Engineering roles posted against a specific workload. A migration disclosed in a job description or a conference talk. A vertical expansion that implies infrastructure the partner happens to be excellent at. This is where a why-now comes from; it cannot be manufactured later.

Open the conversation in the ecosystem’s language. A message that references a MAP-funded assessment, a Marketplace private offer, or a specific migration pattern reads as coming from someone who lives in the same world. A message that references “digital transformation” reads as coming from a sequence. Both arrive in the same inbox. Only one gets answered.

Qualify in writing before it goes anywhere.This is the stage that is almost always missing, and it is the one that determines whether the submission survives a PDM’s review. A written assessment of who the account is, why the offering fits, and what makes now the moment — produced before outreach, not reconstructed afterwards from a call note. Equally important: the accounts that don’t fit should carry the written reason they were excluded, because a qualification system that never refuses anything is not qualifying.

Carry the evidence into the submission.The written qualification, the stakeholder, the why-now, and the conversation history travel with the opportunity into ACE. Nothing is re-keyed by hand at the end of the quarter, and the PDM is not asked to take the partner’s word for it.

Volume was never the constraint

There is a reflex to treat this as a numbers problem — if four submissions produced one close, then forty should produce ten. It does not work that way, and the reason is that ACE submissions are not free.

Every submission is a withdrawal from or a deposit into a credibility account with your PDM. Forty thin submissions do not produce ten closes; they produce a PDM who reads your name in her queue and allocates less attention than she did last quarter. The partner has converted a solvable sourcing problem into a relationship problem that takes far longer to fix.

The partners whose co-sell motion compounds are almost never the ones submitting the most. They are the ones whose submissions the PDM has learned to trust on sight — which is a reputation built one well-evidenced opportunity at a time, and which survives the PDM rotating out, because the next PDM reaches the same conclusion from the same evidence.

This is not only an AWS problem

ACE is the sharpest version of the problem because AWS’s program mechanics are the most explicit, but the structure repeats everywhere. A Microsoft partner submitting a referral through Partner Center faces the same question about whether the deal is real. A Google Cloud partner is measured on contribution to validated closed-won work. An SAP partner in an increasingly indirect-by-default motion is expected to arrive with sourced pipeline rather than wait to be handed some.

Every ecosystem grades the pipeline its partners bring it. None of them supply it. The partner who can manufacture well-evidenced opportunities on demand is advantaged in all of them at once, and the capability transfers even when the program vocabulary does not.

Where to start if your pipeline is empty right now

Not with tooling. Start by separating your two tracks and being honest about which one is broken: count last quarter’s opportunities and mark each as AWS-originated, inbound, existing-customer expansion, or deliberately created. Most partners discover the fourth bucket is empty or nearly so, and that everything they have been doing to fix co-sell was aimed at the first three.

Then look at the ones that were rejected or stalled, and ask what the PDM would have needed to champion them. Usually the answer is not a competency. It is a named stakeholder who had actually responded, and a why-now that was about the account.

That gap — between accounts you could plausibly serve and accounts you have earned the right to submit — is the whole problem. Closing it is not co-sell automation. It happens before any of that, and it is the part of the motion that nobody sells you.