- What You Are Actually Sitting: Exam Facts That Shape Your Plan
- The Two-Outline Trap: Which Objectives Are Current
- Working Through the Six Domains
- Question Style, Timing, and the 700 Scaled Score
- Sequencing the Domains Across Your Prep Weeks
- Registration, Vouchers, and Test-Day Channels
- Where This Credential Fits in the Job Market
- Frequently Asked Questions
- Study the revised six-domain objectives (copyright 2026); the legacy five-domain outline is not complete current English coverage.
- Passing means a scaled 700 on a 1-1000 scale, not a raw 70 percent.
- Expect roughly 45-50 questions in a 50-minute timed limit, so pace around one minute per item.
- Domain 5 is titled Application Integrity and Deployment; integration, CI/CD and security all live there.
What You Are Actually Sitting: Exam Facts That Shape Your Plan
The Information Technology Specialist - Artificial Intelligence certification is issued by Pearson and delivered through Certiport, a Pearson VUE business. It is a computer-based, proctored knowledge assessment. Before you open a single study resource, it helps to understand that this is an objectives-driven exam: Pearson publishes a list of exam objectives, and the test measures your command of those objectives. Those objectives are not a curriculum, so no single course is "the" syllabus. Your job is to map every objective to something you can explain and apply.
If you are still orienting yourself, our explainers on what ITS-AI certification is and what ITS-AI stands for cover the basics. This guide assumes you have decided to sit the exam and want a plan that holds up on the first attempt.
The Two-Outline Trap: Which Objectives Are Current
Here is the single most important thing most candidates miss. The English exam was updated on June 30, 2026, following an announcement Pearson published on April 29, 2026. The revised objectives document (copyright 2026) is organized into six domains. However, Pearson's general IT Specialist resources page has continued to link a legacy copyright-2025 outline with five domains. If you study from that older document, you will be missing material.
The revised coverage separately lists feature engineering, natural language processing, explainability, human interaction, feasibility, model-evaluation metrics, CI/CD, security, drift and decommissioning. Several of those topics either did not appear or were folded into broader headings in the legacy outline. Other-language rollout dates differ, so if you are testing in a language other than English, check Certiport's exam content updates archive for your language's date.
| Check | What to do |
|---|---|
| Which outline am I using? | Confirm it is the 2026 six-domain objectives, ending at objective 6.5. |
| Does my practice material match? | Some retail products retain legacy headings; verify against the revised objectives. |
| Are item counts per domain official? | No. Practice-bank counts are not official exam weights. |
| Is my language updated yet? | Rollout dates differ by language; the June 30, 2026 date concerns English. |
A deeper domain-by-domain treatment lives in our complete guide to all six content areas.
Working Through the Six Domains
Pearson does not publish percentage weights for these domains, so do not trust any site claiming exact weightings. The header simply counts six domain lines. That means you should prepare for balanced coverage and use practice results, not rumors, to find your weak spots.
Domain 1: AI Problem Definition
This is where you decide whether AI is even the right tool. Expect scenario questions about framing a business need as a machine learning task.
- Translate a vague goal into a well-posed problem type (classification, regression, clustering, generation)
- Judge feasibility: data availability, cost, risk and expected value
- Recognize when a rules-based or non-AI solution is the better answer
- Consider human interaction and how users will work with AI output
Domain 2: Data Collection and Transformation
Data quality drives everything downstream. Be ready to reason about sources, cleaning and preparation choices.
- Identify sources, collection methods and common quality problems such as missing values, duplicates and outliers
- Choose sensible transformations: normalization, encoding, splitting into training, validation and test sets
- Spot data leakage and sampling bias in a described scenario
Domain 3: Feature Engineering
Now a standalone domain, so do not treat it as a footnote to data prep.
- Create, select and transform features, and explain why each choice helps or hurts a model
- Understand dimensionality concerns and the tradeoff between richer features and overfitting
- Recognize how text and other unstructured inputs become usable features, which connects to NLP
Domain 4: AI Algorithms and Models
The conceptual core. You need to match algorithm families to problems and interpret how well they perform.
- Distinguish supervised, unsupervised and reinforcement approaches, plus neural networks and natural language processing concepts
- Know model-evaluation metrics and when each is appropriate (for example, why accuracy misleads on imbalanced data)
- Understand explainability: why stakeholders need to trust and interpret model outputs
- Diagnose overfitting and underfitting from a described training result
Domain 5: Application Integrity and Deployment
Use this exact wording when you search for materials. Application integration is objective 5.1, and the domain extends well beyond simply "shipping a model."
- Integrate a model into an application and its surrounding systems
- Understand CI/CD concepts as they apply to AI workflows
- Address security and the integrity of the deployed application
- Consider responsible use, fairness and compliance concerns at deployment time
Domain 6: Maintaining and Monitoring AI in Production
The lifecycle does not end at launch. This domain runs through objective 6.5.
- Monitor performance and detect drift when real-world data diverges from training data
- Decide when to retrain, update or roll back a model
- Plan decommissioning when a model or application is retired
Because the coverage is end-to-end, scenario questions often span two domains at once. A question about a degrading model might require both Domain 4 metric knowledge and Domain 6 drift reasoning.
Question Style, Timing, and the 700 Scaled Score
Pearson MindhubPro estimates 45-50 questions with a 50-minute maximum for this exam. Treat that as a variable range rather than a fixed form count, so do not be rattled if a colleague reports a different number. The 50-minute limit is timed exam time and excludes administration and the tutorial. That works out to a little over a minute per question at the upper end, which rewards quick recognition over long deliberation.
The older IT Specialist exam tutorial (copyright 2021) describes answer-option navigation and gives a generic 36-40 question estimate. That figure predates the exam-specific listing, so use it only for interface familiarity, not for planning your pace.
Practical pacing advice specific to this format: aim to finish a first pass with several minutes in reserve, flag anything that needs a second look, and avoid sinking time into one scenario. For a candid read on difficulty, our difficulty guide breaks down where candidates tend to struggle.
Sequencing the Domains Across Your Prep Weeks
Rather than a generic study template, let the exam's own logic order your weeks. The domains follow the lifecycle of an AI project, and the later domains assume you understand the earlier ones. Given the roughly 150-hour readiness profile, an eight-week arc at around 18 to 20 hours per week fits well, though you should scale it to your background.
Problem framing and data (Domains 1-2)
- Practice turning business scenarios into problem types and feasibility judgments
- Work hands-on with cleaning, splitting and transforming a small dataset
Features and core models (Domains 3-4)
- Build features by hand, then compare algorithm families on the same data
- Memorize which metric fits which situation, and why
Deployment and operations (Domains 5-6)
- Study integration, CI/CD, security, drift and decommissioning as one connected story
- Walk through a model's whole life from release to retirement
Timed practice and gap repair
- Take full timed runs against the 50-minute limit
- Return to whichever domain your practice results show as weakest
This ordering matters because Domains 5 and 6 are new emphasis areas in the revised outline. Candidates coming from older materials often under-prepare exactly there. For a compact end-of-prep refresher, keep our one-page cheat sheet handy, and use the ITS-AI practice tests to simulate the timed conditions.
Key Takeaway
Do not let practice-bank question counts steer your time. Those counts are not official weights, and some products still carry legacy headings. Use the revised six-domain objectives as your checklist and your own practice scores as your guide.
Registration, Vouchers, and Test-Day Channels
Two purchase routes appear in current listings, and the differences matter. Check them before you pay.
| Option | Price shown | Key conditions |
|---|---|---|
| Pearson MindhubPro exam voucher | USD 127 (web/member pricing) | Country-restricted, unavailable in listed markets, expires 12 months after purchase |
| CATC voucher with retake | USD 133 | US-only; proctoring may cost extra; not usable at Pearson VUE centers or through OnVUE |
The voucher expiry is separate from credential validity, so a voucher that lapses after 12 months says nothing about how long your certification lasts once earned. A full breakdown of fees is in our certification cost guide.
Choosing how you will sit the exam
You can test at a Pearson VUE center or through OnVUE, the online proctored option, or through the separate Certiport Authorized Testing Center (CATC) channel. OnVUE requires a system check and a live webcam proctor, so run the check well before exam day and prepare a quiet, cleared space. At a CATC, expect photo and signature ID, an NDA, and parental consent for candidates under 18, with no personal paper or communication devices allowed. Because the CATC voucher is not usable at Pearson VUE centers or through OnVUE, buy the voucher that matches your intended channel.
Scheduling windows and deadlines are covered in our exam dates guide.
How long the credential lasts
Awards earned on or after June 18, 2025 are valid for five years. Earlier awards remain nonexpiring. To recertify, you need a new voucher and must pass the current exam, so keep an eye on the objectives you will face at renewal time.
Where This Credential Fits in the Job Market
This is a foundational, vendor-neutral credential from a major testing provider, which positions it as a strong signal for people entering AI-adjacent work or documenting skills for a transition. It tends to be most useful for students, career changers, IT support and junior analyst staff, and developers who want a structured baseline in the machine learning lifecycle. The six-domain scope, from problem definition through decommissioning, mirrors the responsibilities of roles that touch AI projects without necessarily building models from scratch.
We do not publish invented salary numbers here. For a qualitative look at roles and earnings context, see our ITS-AI jobs overview, the salary analysis and the ROI discussion. If you want structured preparation, our training page outlines learning paths.
Frequently Asked Questions
Pearson MindhubPro estimates 45-50 questions with a 50-minute maximum. This is a variable range, not a fixed count, and the limit applies to timed exam time, excluding administration and tutorials.
A 700 on a 1-1000 scaled score. It is not a raw 70 percent, so you cannot calculate an exact number of allowable misses. Certiport does not publish pass rates.
Use the revised six-domain objectives document with 2026 copyright, effective in English from June 30, 2026. The legacy five-domain outline still linked on some Pearson pages is not complete current English coverage.
Pearson MindhubPro lists USD 127 for the exam-only voucher, which is country-restricted and expires 12 months after purchase. A US-only CATC voucher with retake is listed at USD 133, and proctoring may cost extra.
Five years for awards on or after June 18, 2025. Earlier awards remain nonexpiring. Recertifying means buying a new voucher and passing the current exam. For more on the program itself, see our ITS-AI certification overview or return to the practice test hub to start drilling.