AI certifications for non-technical roles

The AI certifications that suit non-technical roles are AB-730 if you use Microsoft 365 Copilot in your daily work, AB-731 if you decide how your organisation adopts AI, and AIF-C01 if your company works on AWS or you want a vendor view of AI concepts. None of the three needs code. The better-known AI-901 is no longer a safe choice for non-programmers, because it now assumes Python.

Example profiles, not real people. The scenarios below are illustrative planning examples built from the official exam outline. They are not reviews or reports of anyone’s result.

The decision rule

Ask two questions. Which tools does your employer actually run, and do you use AI, decide about it, or administer it? The answer picks one exam.

The Microsoft pair: AB-730 and AB-731

Both are 45-minute exams with a passing score of 700 out of 1000, a Beginner level and thirteen languages. Microsoft does not publish a question count for either.

AB-730 is the individual user’s exam. From October 20, 2026 its four skill areas are generative AI fundamentals (25 to 30 percent), managing prompts and chats (20 to 25), managing business content and collaboration (20 to 25) and driving business outcomes with agents (20 to 25). It describes someone comfortable with Outlook, Word, Teams, PowerPoint and Excel who has already used generative AI tools.

AB-731 is the decision-maker’s exam: business value of generative AI (35 to 40 percent), Microsoft’s AI apps and services (35 to 40) and implementation and adoption strategy (20 to 25). Microsoft’s profile expects experience leading adoption or change in a business context. The AB-730 vs AB-731 guide explains the split in detail, including where they overlap.

Neither exam page states whether the credential expires. The site’s AB-730 and AB-731 renewal pages say to check your Microsoft Learn profile after passing, because the expiry term is not confirmed.

The AWS option: AIF-C01

AWS describes the AI Practitioner candidate as someone who uses, but does not necessarily build, AI and ML solutions. The exam has 65 questions (50 scored, 15 unscored) in 90 minutes, a passing score of 700 on a 100 to 1000 scale and a price of 100 USD. Its five domains are fundamentals of AI and ML (20 percent), generative AI fundamentals (24), applications of foundation models (28), responsible AI (14) and security, compliance and governance (14). The credential is valid for three years.

AWS’s exam guide lists tasks the candidate is not expected to do: coding models, feature engineering, tuning, building pipelines and writing governance frameworks. For a product manager or consultant, that scope is the point. For the choice between AIF-C01 and Microsoft’s fundamentals exam, see AIF-C01 vs AI-901.

Exams that look suitable but are not

AI-901. Microsoft’s audience profile asks for Python coding syntax and familiarity with REST APIs and SDKs, and 55 to 60 percent of the weight is implementing solutions in Foundry. That is a technical fundamentals exam.

AB-900. It tests Microsoft 365 services, data protection and governance, and Copilot and agent administration. It suits an administrator, not an end user.

ISACA AAIA, AAIR and AAISM. The site’s ISACA path says none is entry-level and all assume you already work in audit, risk or security management. The exam pages add that AAIA requires CISA or a qualifying audit or accounting designation and AAISM an active CISM or CISSP. They are non-coding but not beginner exams.

Salesforce Agentforce Specialist. It is a platform exam for people who configure Salesforce agents, so it only fits if you already work in that ecosystem.

Realistic constraints

Time. The AB-730 plan is four weeks at three to five hours a week, or two weeks if you use Copilot daily. The AB-731 plan is four to five hours a week for four weeks. The AIF-C01 plan is five to seven hours a week for four weeks.

Hands-on access. AB-730 questions often name a button or feature in the product, so a Copilot licence helps; the free practice assessment and exam sandbox are the fallback.

Language. The Microsoft pair is offered in thirteen languages. AIF-C01 is offered in English and several others, but German and Italian end after October 15, 2026. See the language guide.

Retakes. Microsoft allows a retake after 24 hours, then 14 days between later attempts, with at most five attempts in 12 months. AWS requires 14 calendar days and the full fee each time, with no attempt limit.

Scenario: a team lead coordinating a Copilot rollout

This is an illustrative planning example. The candidate coordinates a Copilot rollout, does not administer the tenant and has never trained a model.

Where the time goes: AB-731 puts 35 to 40 percent each on business value and on Microsoft’s AI apps and services, so the work is learning which product fits which need, then adoption and governance at 20 to 25 percent. The trap is treating the exam as a Copilot features test and skimming cost drivers, licence models and responsible AI governance, which the site’s study plan lists as exam material.

A reasonable planning assumption is four weeks at four to five hours a week, 16 to 20 hours, in line with the site’s 15 to 20 hour planning figure for a leader with a relevant background. This is a planning figure, not a guarantee. If the same person mostly uses Copilot rather than directs it, AB-730 and its four-week plan (two weeks for a daily Copilot user) fits better.

Where to go next

Read the Azure AI certification path or the AWS AI path for what follows the fundamentals. The AB-730 prerequisites and AB-731 prerequisites pages include a self-check, and the exam changes tracker shows the October 2026 outline dates.