Working with an AWS Partner in 2026: The Guide
Schedule a free consult
Back to Blog
Working with an AWS Partner in 2026 - allOps

Working with an AWS Partner in 2026 – What They Do, What AWS Co-Funds and How an Engagement Runs

Danijel Milosevic

An AWS partner is a company AWS has audited to deliver specific kinds of work on its platform: moving workloads onto AWS, reviewing and fixing the architecture of what already runs there, taking an AI use case from idea to production, and running the estate afterwards.

You bring in a partner to deliver that work and to lower the risk of running on AWS while it gets done. AWS co-funds part of that delivery, as services it pays the partner for or as credits on your bill, when the project is scoped and approved in advance.

This guide explains what a partner actually does, when you need one, how the four kinds of co-funded work look from your side, how an engagement runs from the first conversation to handover, and what to ask before you sign.

What does an AWS partner actually do?

The AWS Partner Network (APN) is AWS’s programme for companies that build, migrate and operate on AWS for customers. AWS audits partners per area of work and calls the result a competency: a review of delivered projects, customer references and certified people in one field, such as migration or AI services.

In practice a partner does four things for you.

  • Assesses – looks at what you run, what it costs, what is at risk, and what a move or a change would take.

  • Builds and migrates – designs the landing zone, moves the workloads, modernises what the business case supports, ships the AI use case.

  • Reviews and fixes – holds a production workload against the AWS Well-Architected Framework and closes the high-risk issues.

  • Runs – managed services – takes over the day-to-day care of your AWS account under a managed-services agreement: operations, cost reviews, incident response, and the architecture decisions that come up as the business changes, with a named team that knows your estate.

For companies that don’t want to run AWS in-house at all, this is the whole relationship.

The partner also does one thing you cannot do yourself: it takes the project to AWS for co-funding through channels only partners have.

When do you need an AWS partner?

You need one when the work is bigger than your team’s spare capacity, when it is a kind of work your team has not done before (e.g. a first migration, a first production AI system, a compliance-driven review), or when the timeline is set by something outside engineering, for example, a data-centre lease, a licence renewal, an audit date or a board decision.

What if your team already runs AWS well? A strong in-house team can absolutely keep tuning a workload it knows, like a self-run Well-Architected Review in the free AWS tool or a monthly pass through AWS Cost Explorer. Where a partner still earns its keep is everything around that one workload.

A partner watches the whole account – cost drift, security posture, the services AWS ships between two of your sprints – and it carries expertise an in-house DevOps team rarely gets to build, because an in-house team sees one estate while a partner sees dozens of them, across industries, with AWS auditing the work.

Also, it often happens that a team could do the work, but slower, without the funding, while everything else waits. It’s a capacity decision before it is a skills decision.

Which kinds of work does AWS co-fund through a partner?

Four kinds of partner work come with AWS co-funding when the project is scoped and approved in advance. Each has its own guide in this series.

Migration to AWS

Moving servers, databases and applications from a data centre or another cloud onto AWS, in the three phases of the AWS Migration Acceleration Program (MAP): Assess, Mobilize, and Migrate and Modernize.

The partner runs the readiness assessment, builds the landing zone, moves the workloads in waves and reports milestones to AWS. From your side the decisive detail is a tagging rule. Credits are calculated from resources that carry a specific tag, and a resource nobody tagged earns nothing.

Architecture review of a running workload

A structured review of one production workload against the six pillars of the AWS Well-Architected Framework, run in the AWS Well-Architected Tool, ending in a list of high-risk issues and their remediation. AWS funds partner-led reviews in credits on your bill, tied to the usage the remediation generates – the review only changes anything once the issues are fixed.

AWS co-funds AI work well before anything gets built. It starts with the assessment. They check one or two candidate use cases, whether the data supports them, which AWS services fit (Amazon Bedrock, Amazon Q, Amazon SageMaker), what governance they need, and what they return. If the answer is yes, a co-funded proof of concept follows, and the same track carries the workload into production, with integration, monitoring and cost at scale planned from the start.

Whether you know what to build or you’re in that thinking loop “we think AI could help, we’re not sure where.”, the assessment covers that.

New workloads on AWS

When a partner brings a new workload to AWS with you, AWS can grant credits against the bill for that workload as it goes live. This is the simplest of the four from your side. The partner scopes it, AWS approves it, and the credits land on your account.

How does AWS co-fund the work?

AWS funds partner work per project, against a scope the partner submits and you sign off.

Funded services: AWS pays the partner for a defined piece of work, such as an assessment, so you receive that work without paying for it. Credits: AWS applies an amount against your AWS bill, released as the project hits milestones AWS can verify.

Three rules decide whether the funding arrives.

  1. Nothing is retroactive


    Work delivered before AWS approves the project is your cost, and it cannot be added afterwards.

  2. The scope is the contract


    If the project grows or changes, the change goes back to AWS before the work changes, or the changed part is uncovered.

  3. Milestones have to be visible


    Credits are released against milestones the partner reports and AWS can verify in your account, such as tags, billing reports, saved review milestones.

How much AWS contributes depends on the project and on AWS’s decision at the time.

Eligible projects may qualify for AWS funding, subject to AWS approval.

How does a partner engagement run, start to finish?

  1. First conversation


    What you run, what you want to change, what is forcing the timeline. Half an hour, no slides.

  2. Assessment


    A short, structured look at the estate or the use case that reveals the workloads in play, risks and estimated effort. At allOps this is the Foundation Sprint. It takes ten days and it’s free. Its output is the scope everything else rests on.

  3. Scope and staging


    The work is cut into stages you can stop between: for a migration, Assess, Mobilize, then waves; for a review, the review, then remediation; for AI, the assessment, then a proof of concept, then a build.

  4. Submission to AWS


    Where the stage qualifies for co-funding, the partner submits the scope through partner channels and AWS approves, asks for changes or declines. You wait for the answer before the stage starts.

  5. Delivery


    The partner delivers the stage against the approved scope and reports milestones. You confirm them.

  6. Handover or run


    Either your team takes over with the documentation and the runbooks, or the partner runs the estate under a managed-services agreement.

Steps one to three cost you a few weeks and no money. Step four is where the calendar can bite: if a deadline forces a start before AWS has answered, start with the smallest piece you can and keep it out of the submitted scope.

What do you have to do on your side?

  • Name a project owner. It’s required for funding submissions and scope sign-offs.
  • Give the partner read access to the accounts for the assessment, the people who run the workloads for the question sessions.
  • Keep the AWS account structure stable during the project, including the payer account and any tagging the project requires.
  • Confirm milestones when the partner reports them.
  • Say early when the scope has to change. Late changes are the single most common reason co-funding is lost.

How do you choose an AWS partner?

Funded or not, the partner delivers the work, so the partner’s ability to deliver matters more than any line in the funding application. Six questions tell you most of what you need.

  1. Which competencies has AWS audited you for, and do they match our project?


    A migration needs the Migration Competency, an AI project needs the AI Services Competency.

  2. Who exactly will work on our project, and what have they built?


    Ask for names and for the AWS certifications behind them. Ask whether an AWS Hero or a principal-level architect is involved in the decisions, and how.

  3. Which projects like ours have you submitted to AWS for co-funding, and how many were approved?


    The answer tells you whether they know the process or are learning it on yours.

  4. What do we have to do on our side, in one page?


    A partner who has done this before has that page.

  5. What happens if AWS declines?


    Whether the project still makes sense unfunded is a question to settle before submission, with the partner in the room.

  6. Which customers can we talk to?


    Published case studies first, then a call with one of them.

It might be tempting to evaluate just a partner’s tier badge on its own. Tiers reflect a partner’s overall volume and certifications with AWS, but not specific workloads.

Where do partner projects go wrong?

  • Scope written to fit the funding instead of the work.


    The project gets approved and then does not deliver what you needed.

  • Starting before approval, to hit a date.


    The early work is never covered.

  • Tags and billing reports set up late or in the wrong account.


    In a migration this is the difference between credits and nothing.
  • A review that ends with a report.


    High-risk issues identified and never closed.

  • No handover plan.


    The partner leaves and the team that inherits the estate did not sit in the design sessions.

  • An AI proof of concept that was never scoped for production.


    It works in a demo and nobody planned the integration, the monitoring or the cost at scale.

Frequently asked questions

Is an AWS partner the same as AWS Professional Services?

No. AWS Professional Services is AWS’s own consulting arm, usually engaged on large enterprise programmes. Partners are independent companies AWS has audited. Most mid-sized projects, and most co-funded ones, run with a partner.

Do we have to buy AWS through the partner?

No. Your AWS account stays yours and you pay AWS directly. Some partners also resell AWS with added billing services; that is a separate decision from engaging them to deliver a project.

Is the co-funding a discount on our AWS prices?

No. Your AWS prices do not change. Funding is a contribution to a specific project: partner work AWS pays for, or credits on your bill released per milestone.

How long does AWS take to approve a project?

Weeks, depending on the size and on how complete the submission is. A clean scope with clear milestones moves faster than one AWS sends back.

Can we apply for the funding without a partner?

No. AWS co-funding is available exclusively through AWS Partners, and what a project can draw depends on the partner’s tier and the competencies AWS has audited them for. What you can do yourself is be ready: know which workloads are in play, name a business owner, have your account structure clear.

What if the assessment says we should not do the project?

Then you have saved a build, and the assessment report is yours to keep. A partner that never says no in the assessment is not assessing.

Does the partner have to be in our country?

No. AWS partner work is delivered remotely as a rule, with on-site sessions where the project needs them. What matters is the time zone overlap and whether the partner has delivered in your industry and under your regulation.

What to do next

The first step with allOps is the Foundation Sprint: ten days, at no cost to you, and the output is the scope any funding submission rests on. See the Foundation Sprint.

Already know the work you need? Go straight to the offer: Migration Sprint, WAFR Premium, or Managed Services.

See all offers. Or start at AWS Nexus.

About allOps

allOps is an AWS Advanced Tier Services Partner, audited by AWS for the Migration Competency and the AI Services Competency. Two of the 255 AWS Heroes worldwide work here, the team holds 50+ AWS certifications and has trained 1.000+ people on AWS.

We love cookies!
They help us make your experience on our website even better. If that is ok with you, click Accept, and if you need additional info, you can check it by clicking on Cookie policy.
Cookie policy