Declared conflict of interest: we are an AWS consultancy, so this article is written by a player, not a referee. We are publishing it anyway, because we sit through enough first meetings where a company has just left a bad engagement — overbilled, under-delivered, locked in — and the pattern of what went wrong is always visible in hindsight, usually in the proposal itself. Here is how to evaluate an AWS partner before you sign, including the questions we would want to be asked.

First: what partner badges actually mean

AWS partner tiers (Select, Advanced, Premier) verify real things — certified engineers, documented customer wins, revenue thresholds. They are a useful floor: a Premier partner is an established business, not two people and a landing page. But badges measure scale more than fit. A Premier partner optimized for 500-person enterprise migrations may serve a 50-person SaaS company worse than a smaller senior team — you may be staffed with their most junior bench. Treat the badge as a filter, never as the decision.

The 7 questions that separate senior teams from body shops

  • 1. "Who exactly will work on our account?" Not "our team of 200 engineers" — names. Ask for the CVs of the actual engineers and whether they are the same people who show up after the sales cycle. The classic failure mode: senior people in the pitch, juniors in the delivery.
  • 2. "Show us something you've published." Teams with real expertise leave a public trail — engineering write-ups, incident postmortems, open checklists. It is the hardest signal to fake. (You are reading our version of it; our war stories are the deeper cut.)
  • 3. "What does it cost?" — and watch how they answer. "It depends, let's scope a workshop" is sometimes legitimate and sometimes a stall. Firms confident in their delivery can publish prices or give fixed quotes after discovery. If every number requires three meetings, the pricing model is the product.
  • 4. "Fixed price or time-and-materials — and who carries overrun risk?" On migrations and audits, T&M puts all uncertainty on you. A partner who offers fixed-price with defined scope is betting on their own competence; that is the bet you want them making.
  • 5. "Where do our workloads live, and what happens if we leave?" The right answer: in your AWS accounts, with role-based, logged access for the partner, and a documented offboarding path. Workloads hosted in the provider's own account are how consulting relationships turn into hostage situations.
  • 6. "Walk us through a production incident you handled." Not the marketing version — the timeline. Teams that have actually carried a 3am outage talk differently about it: specifics, mistakes included. Teams that haven't, generalize.
  • 7. "How do you use AI in operations?" A 2026 question that sorts the field fast. Monitoring, log triage, and first response are being transformed; a partner with no answer is selling you yesterday's cost structure at yesterday's prices.

5 red flags in the proposal

  • Savings promises before seeing your bill. "We'll cut your AWS costs 40%" without access to your Cost Explorer is a sales number, not an engineering estimate. Real firms audit first, then commit — ideally taking payment from realized savings.
  • No named engineers, only roles. "1x Solution Architect, 2x DevOps Engineer" staffing tables with no humans attached means you're buying from the bench.
  • Vague deliverables. "Cloud optimization support" is not a deliverable. "A per-finding cost report with € impact, delivered in two weeks" is. Insist on artifacts, not activities.
  • Everything is a retainer. Some needs are projects with an end. A partner who converts every conversation into an open-ended monthly engagement is optimizing their revenue, not your outcome. (We sell retainers too — for the genuinely continuous work: 24/7 operations, monitoring, incident response.)
  • They discourage your team from learning. A good partner documents, trains, and makes themselves replaceable. Gatekept knowledge and undocumented environments are lock-in by another name.

Partner, freelancer, or in-house?

A senior freelancer beats a mediocre agency for well-scoped project work — but one person cannot cover 24/7 operations or the full breadth of networking, security, cost, and databases, and continuity ends at their next engagement. Building in-house is right when infrastructure is your product or your scale genuinely fills a platform team; the honest math is in our in-house vs managed services comparison. For most companies between those poles, a small senior partner team — with named engineers and published prices — is the efficient middle.

Frequently asked questions

Do I need an official AWS partner?
Not necessarily — the badge verifies scale and certifications, not fit. Use it as a floor filter, then evaluate on named engineers, published work, pricing transparency, and incident experience.

What does AWS consulting cost?
Wide range: senior freelancers bill by the day; agencies run monthly retainers (ours start at €3,000) and fixed-price projects (our audits: €5,000-8,000; migrations: €25,000-150,000). Beware of any number quoted before anyone has seen your environment.

How do I avoid vendor lock-in with a partner?
Keep workloads in your own AWS accounts, require role-based logged access, insist on documentation as a deliverable, and agree the offboarding path in the contract — before signing.

Should the partner also resell our AWS billing?
Reseller arrangements can bring discounts but add a dependency: your bill flows through the partner. If you take one, make sure account ownership and the exit path stay clean.

Bottom line

Choose the team, not the logo: named senior engineers, a public engineering trail, transparent prices, fixed-price risk-sharing, your accounts, and a documented way out. Any partner worth hiring will pass those checks without flinching — and if this article's checklist disqualifies us for your case, it did its job.

Evaluating partners right now? Our services and pricing are published — assessments from €5,000, managed operations from €3,000/month — and a free AWS health scan is a low-risk way to see how we work before any commitment.