This month of October, AWS dropped a signal moment in its certification portfolio: the upcoming AWS Certified Generative AI Developer, Professional (beta opens November 18 2025) and the planned retirement of the AWS Certified Machine Learning – Specialty exam (last opportunity March 31 2026). That’s a clear pivot from a machine learning approach to something more focused on generative AI models.
Why this shift?
I think there are two major factors that explain this move:
- Technology shift – Generative AI (foundation models, RAG, vector DBs, embeddings) has moved from research labs to production workflows. The new cert explicitly calls out integration of foundation models into applications and business workflows and RAG architectures and vector databases.
In practice: the skills demanded of developers and architects are evolving. AWS is aligning the credentialing to that evolution.
- Market demand – From a strategic lens, designing certifications that map to business outcomes is key and when the business world says: “we need generative AI in production now”, the certification world should respond. The new Professional GenAI cert fills a gap.
Entry-level GenAI certification
While we wait for the new Pro level cert, there is a foundational GenAI certification, the AWS Certified AI Practitioner. Unlike the associate and professional AWS certifications, the "practitioner" category are targeted toward technical and non-technical folks alike. At Metal Toad we've embraced this with 100% of our team taking the cert. This GenAI practitioner cert lives alongside only one foundational certification, the AWS Cloud Practitioner, marking how important AWS believes this technology is.
Given the current business environment if you only have time for one, I would recommend the AI practitioner, as I believe we have moved from the cloud epoc to the AI epoc.
Both foundational certs are for non-builders or early entrants: business analysts, IT support, product managers, sales professionals — folks who may not code models but need fluency in the concepts. The exam is 65 questions, 90 minutes.
By comparison the pending “Generative AI Developer – Professional” is for the builders: developers, AI engineers working with foundation models, vector stores, RAG, deploying at scale.
The AWS certification ecosystem
The AWS Certification site positions the portfolio as validating technical skills and cloud expertise to grow your career and business. There are four categories, but can be thought of in 3 tiers:
- Foundational
- Associate
- Professional or Specialty
These are not easy credentials. Especially at Associate / Professional / Specialty level, you're expected to bring hands-on experience, choose architectures, justify trade-offs, integrate services—not just regurgitate trivia. Because of this, they carry portfolio value for teams. AWS themselves claim that organizations with AWS Certified staff report faster troubleshooting and increase innovation from certified individuals. Additionally, for an organization, having team members with certifications sends a signal of rigor and capability to clients and prospects.
When we talk about digital maturity, certifications are one of the gears in the machine. They help standardize capability, reduce risk (you know folks have been vetted), support talent mobility, and provide a shared vocabulary across teams.
Practical implications
- For individuals: If you hold ML-Specialty, don’t panic, but plan ahead. If you are building GenAI solutions (or even if you haven't yet), you should consider adding the new GenAI Professional cert to your badge list. And the ML training can still be useful. In our professional AI engagements, GenAI is the right answer approximately 75% of the time, with 25% needing an "old fashioned" ML solution.
- For teams and leaders: Map your learning path now. Use AI Practitioner for broader team literacy, then Generative AI Developer for those leading GenAI production efforts. If you are or have a subject matter expert in artificial intelligence more broadly, the ML specialty cert may still be worth doing, even if it is being taken off the roster next year.
A note of caution & reflection
While certifications matter, they are not the whole answer. Holding a certificate doesn’t guarantee a project succeeds. What matters is how you apply the credential to real business problems, how you embed practices, governance, sustainable models of operation.
The AWS announcement signals the direction, but keep an eye on how quickly content and exam domains evolve. Foundation models and GenAI move fast—so certifications may lag, and you must keep your practice current beyond the certificate.