The SIAA Standard
The AI skills we examine and the six areas we rate. How each exam is built and marked, and how sure each score is.
Learn the skills any way you like; no book or course is required. Every credential we issue is a claim about one skill here, or about a level built from several.
Edition 2026-09. The Standard includes 102 skills.
The Academic Board of the Straits Institute for Applied AI reviews the SIAA Standard. Meet the Academic Board
What we measure
Every exam rates the same six areas, three on each of two axes. Value is what you get from the assistant; Control is what you keep in hand. Each area gets one of 4 ratings: Emerging, Developing, Proficient, Accomplished.
Each area states what it includes, what it leaves out and what counts as evidence, so two examiners reading it look for the same thing. When a task needs behaviour from two areas, it is scored as two separate events, one in each.
Value
What you get from the assistant: how far it takes the work, how well you brief it, and the piece you hand in.
Reaching further
Using the assistant to produce work beyond what you could have produced alone, rather than the same work faster.
- Includes
- Using the assistant to cover more of what the reader needs: questions to anticipate, alternatives to compare, gaps to name.
- Leaves out
- Doing the same work faster or longer, the number of messages, and work the assistant was not used for.
- Evidence
- Turns that ask the assistant to extend the work, and the reader needs or alternatives that reach the finished piece from them.
Briefing the assistant
Giving the assistant the reader, purpose and source facts the work depends on, and the form, length and contents you need back.
- Includes
- Context (the reader, the purpose and the source facts) and specification (the form, length and required contents), scored as two separate events.
- Leaves out
- What you keep out of the tool (Safety and confidentiality), who approves the result (Knowing what to delegate), and wording style.
- Evidence
- Your turns before you rely on the output. Saying afterwards that you briefed it fully is not evidence.
The finished piece
Turning the assistant's output into something a colleague could use: editing it, combining it and adding original thinking.
- Includes
- Editing, combining and adapting the output to its audience, with your own reasoning added.
- Leaves out
- Exact wording and length, and the briefing that produced the draft.
- Evidence
- The work you hand in, and your answer when the scenario changes.
Control
What you keep in hand: checking its work, keeping information safe, and deciding what stays with a person.
Checking its work
Checking the assistant's claims, figures and assumptions against the source before the work is used.
- Includes
- Checking claims, figures and assumptions against the supplied source, and correcting what is wrong.
- Leaves out
- Saying afterwards that you checked, and disagreeing with a correct answer for its own sake.
- Evidence
- A check in your turns or your work that names the claim and the source. It is rated only when the assistant's planted mistake reached you.
Safety and confidentiality
Handling confidential and personal data, seeing compliance, ethical and reputational risk, and declining unsafe requests.
- Includes
- Keeping restricted and personal information out of the tool and the output, and declining unsafe requests.
- Leaves out
- Whether the content is accurate, and who has authority to approve it.
- Evidence
- What you send to the tool, what your work contains, and your account of what you withheld. A cited 0 on a critical Safety event fails the result.
Knowing what to delegate
Deciding what to hand to the assistant and what to keep, and at what stakes, before any output exists.
- Includes
- Deciding what to hand to the assistant and what to keep, saying only what is confirmed, and naming who must approve an open decision.
- Leaves out
- The quality of the briefing, and the arithmetic itself.
- Evidence
- Your recommendation or notice, and the reason you give for keeping a decision with a person.
How a brief to the assistant is scored
We never score a prompt for style. Two events read your turns before you rely on the output. Context: did you give the assistant the reader, the purpose and the facts the work depends on? Specification: did you say the form, length and contents you need back? What you keep out of the tool counts under Safety, and whether you checked the result counts under Checking its work. A polished prompt with no check loses the checking points.
Results issued before September 2026 were rated on the five areas of the 2026-05 edition, and they keep them. Applied AI Specialist and Applied AI Strategist are still marked on that edition until the Examination Board approves their 2026-09 version.
Find a skill
Find a skill you use at work. See what a strong performance looks like in each area, practise it, or choose an available exam.
102 skills
Working with AI
Knows what today's AI can and cannot do, and tells a real capability from a demo.
Decides which tasks to hand to AI, which to keep human, and why.
Breaks a working week into tasks and spots where AI saves time without adding risk.
Gives an assistant the context, constraints and format that get usable output first time, and iterates when it does not.
Catches fabricated facts, missing steps and confident errors against the source before anything goes out.
Keeps confidential and personal data out of the tool, declines unsafe requests and stays accountable for the result.
Recognises the compliance, ethical and reputational risks of AI in a specific profession and acts on them.
Working with agents
Tells assistance from automation from autonomy, and matches the level to the stakes of the task.
Sets up a purpose-built assistant with the right instructions, knowledge and guardrails for a team's work.
Chains AI with the documents, data and software a job already runs on, safely.
Hands multi-step work to AI agents with clear goals, checkpoints and a human decision where it matters.
Sets the rules, oversight and accountability for AI that acts on a team's behalf.
Research and Insight · 14 skills
Everything in Research and Insight, including credentials for this work area.
Conducts academic research with AI and verifies every source and claim.
Runs market research with AI and separates evidence from confident guesswork.
Researches policy with AI and confirms every citation against the primary text.
Writes academically with AI's help, with sources real, cited and read.
Evaluates claims and evidence critically with and about AI.
Gathers customer insight with AI and checks that the synthesis reflects what customers said.
Analyses competitors with AI and verifies the facts before they shape strategy.
Applies qualitative methods with AI support and keeps interpretation rigorous and human.
Runs systematic literature reviews with AI and confirms every included study exists and says what is claimed.
Runs quantitative analysis with AI and checks the statistics and the assumptions.
Designs experiments and analyses lab data with AI, with every result reproducible.
Uses AI to organise and analyse fieldwork while protecting participants and context.
Analyses text and discourse with AI and validates the coding against the source.
Applies evidence-based practice with AI and verifies every clinical claim against the literature.
Marketing and Content · 5 skills
Everything in Marketing and Content, including credentials for this work area.
Uses AI as a creative collaborator and delivers work that is distinctly theirs.
Writes scripts with AI, then edits for voice, accuracy and rights.
Shapes a story with AI and keeps the facts, the audience and the message straight.
Builds a personal brand with AI-assisted content that is authentic and true.
Produces social and professional content with AI that is accurate, on brand and safe to publish.
Communications · 10 skills
Everything in Communications, including credentials for this work area.
- Business Writing
Free practice available
Produces clear, correct business documents with AI and owns every claim in them.
- Report Writing
Free practice available
Turns data and notes into a report with AI, with every figure traced to its source.
Builds a presentation with AI that says one thing per slide and nothing the speaker cannot defend.
Prepares and sharpens a pitch or talk with AI, then delivers it as their own.
Designs and runs a session with AI's help and keeps the room, not the tool, in charge.
Communicates across channels with AI-assisted drafts that stay accurate, on tone and on record.
Prepares for a hard conversation with AI, without letting it script away the human judgement.
- Email & Correspondence
Free practice available
Drafts, checks and sends professional correspondence with AI, and owns every line of it.
Adapts messages across cultures and languages with AI, and catches what the tool gets wrong.
Drafts fast, accurate statements under pressure with AI, with nothing unverified released.
Legal, Risk and Compliance · 8 skills
Everything in Legal, Risk and Compliance, including credentials for this work area.
- Document Review
Free practice available
Reviews documents with AI at scale and catches what the tool misses or invents.
Identifies and weighs risks with AI's help and owns the call on each one.
Applies AI ethically in practice: consent, fairness, transparency and knowing when not to use it.
Tracks and interprets regulation with AI and verifies every citation before relying on it.
Reviews contracts with AI, confirms every clause against the text and escalates what matters.
Prepares ESG and sustainability reporting with AI, with every metric evidenced.
Uses AI in quality assurance to find defects and never to sign off on its own.
Designs monitoring and evaluation with AI and validates indicators against real data.
Customer Service · 6 skills
Everything in Customer Service, including credentials for this work area.
Answers customer enquiries with AI at speed while keeping every answer true to policy.
- Customer & Client Communication
Free practice available
Handles customer and client communication with AI without leaking data or promising what cannot be delivered.
Onboards clients with AI-assisted materials and checks that every step and figure is right.
Resolves complaints with AI's help while keeping judgement, empathy and accountability human.
Runs case management with AI summaries and drafts, verified against the record.
Uses AI to spot churn signals and prepare retention outreach that is accurate and honest.
Operations and Administration · 12 skills
Everything in Operations and Administration, including credentials for this work area.
Maps and improves a process with AI and tests the change before rolling it out.
- Meeting Minutes & Action Tracking
Free practice available
Turns meetings into minutes and actions with AI, checked against what was said.
Writes technical documentation with AI and verifies every step works as written.
Runs a remote team with AI-assisted coordination and keeps people, not tools, informed and accountable.
Organises their own work with AI and knows what to automate and what to keep by hand.
- Meetings
Free practice available
Prepares, runs and follows up meetings with AI, from agenda to verified actions.
Builds and maintains a knowledge base with AI and keeps it accurate and current.
Manages records with AI-assisted classification and retention, correctly and lawfully.
Writes standard operating procedures with AI and checks them against how the work is done.
Plans inventory and resources with AI forecasts and sanity-checks the numbers.
Automates scheduling with AI without losing control of priorities and confidentiality.
Catalogues and tags information with AI and audits the results.
Sales and Business Development · 8 skills
Everything in Sales and Business Development, including credentials for this work area.
Writes grant applications with AI that meet the call exactly and make no claim the evidence cannot back.
Prepares negotiations with AI and keeps the judgement, and the relationship, human.
Researches and approaches prospects with AI and keeps every claim about them accurate.
Manages accounts with AI-assisted insight and communication that stays honest and on record.
Writes outreach with AI that is personal, accurate and compliant.
Prepares sales conversations with AI and never lets it promise what the product cannot do.
Builds relationships with AI-assisted research and follow-up that stays genuine.
Writes tenders and bids with AI that answer the specification and evidence every claim.
People, HR and L&D · 10 skills
Everything in People, HR and L&D, including credentials for this work area.
Plans lessons with AI, aligned to objectives and checked for accuracy and level.
Designs assessments with AI and validates that they measure what they claim.
Designs learning experiences with AI and keeps pedagogy, not the tool, in the lead.
Builds online courses with AI-generated material that is verified before learners see it.
Coaches with AI as preparation and reflection, never as a substitute for the conversation.
Mentors with AI-assisted resources and judgement about when advice must be human.
Plans career development with AI and tests its suggestions against real evidence.
Manages people with AI-assisted planning and communication while keeping decisions and fairness human.
Gives feedback and runs performance conversations with AI preparation and human accountability.
Prepares and conducts interviews with AI and guards against bias in the process.
Finance and Analysis · 8 skills
Everything in Finance and Analysis, including credentials for this work area.
- Data Analysis
Free practice available
Analyses data with AI and verifies every result before it drives a decision.
Designs surveys and interprets responses with AI, checking bias and sample before reporting.
Builds charts with AI that show the data faithfully, labelled and sourced.
Manages a budget with AI-assisted analysis and checks the arithmetic and the assumptions.
Reads financial statements and terms with AI's help and knows when the explanation is wrong.
- Spreadsheets & No-Code Data
Free practice available
Builds spreadsheets and no-code data flows with AI and tests them before trusting them.
Runs financial analysis with AI and reconciles every number to its source.
Turns numbers into a narrative with AI that stays true to what the data shows.
Strategy and Leadership · 9 skills
Everything in Strategy and Leadership, including credentials for this work area.
- Project Planning
Free practice available
Plans projects with AI and stress-tests the schedule, dependencies and risks.
Runs agile work with AI in the loop and the team in charge.
Leads change with AI-assisted analysis and communication and keeps the human case central.
Plans events with AI and verifies logistics, budgets and commitments.
Thinks strategically with AI as a challenger and reaches conclusions they can defend.
Uses AI to generate and test ideas and judges which are worth pursuing.
Makes decisions with AI-assisted analysis and knows when the analysis cannot be trusted.
Applies design thinking with AI for research, ideation and prototyping, with users at the centre.
Manages stakeholders with AI-assisted mapping and communication that stays accurate and discreet.
What a score is for
A score is only as good as the decision it supports. Each exam names its decision, and the decisions its score must not carry.
- Applied AI Foundations, free
- Tells a beginner, and anyone they share the result with, whether they can use and check AI in a guided, bounded everyday task. It is not a hiring decision or a professional credential.
- An employer's first interview
- One input to decide what to explore with a candidate in a first interview. Never an automatic cutoff, a ranking of record, or a prediction of job performance. Employer exams are still marked on the five areas of the 2026-05 edition.
- Applied AI Certified
- Awards a credential stating that the holder handles general workplace tasks with AI, on their own, within stated authority.
How each exam is built
Applied AI Foundations
20 minutes. Approved provisionally by the Examination Board, before its evaluation study. Its results say so.
| Area | Axis | Share of points |
|---|---|---|
| Reaching further | Value | 20% |
| Briefing the assistant | Value | 15% |
| The finished piece | Value | 15% |
| Checking its work | Control | 20% |
| Safety and confidentiality | Control | 15% |
| Knowing what to delegate | Control | 15% |
- Pass rule
- To pass, you need Proficient overall, in Value and in Control, and at least 50 out of 100 in Reaching further. The mark is provisional until a panel of subject experts sets it.
- What the result shows
- Written observations on each task, quoting your work. Scores are not shown.
Applied AI Certified (L1)
32 minutes. Approved provisionally by the Examination Board, before its evaluation study. Its results say so.
| Area | Axis | Share of points |
|---|---|---|
| Reaching further | Value | 20% |
| Briefing the assistant | Value | 15% |
| The finished piece | Value | 15% |
| Checking its work | Control | 20% |
| Safety and confidentiality | Control | 15% |
| Knowing what to delegate | Control | 15% |
- Pass rule
- To pass, you need Proficient overall, in Value and in Control, and at least 50 out of 100 in Reaching further. The mark is provisional until a panel of subject experts sets it.
- What the result shows
- An overall score, Value and Control scores, scores for named areas, with written observations.
AI examiners score each event 0, 1 or 2, and a 1 or a 2 must quote your own words. The assistant's text alone is never evidence of your skill. How marking is checked.
How sure a score is
Every score contains some error. Reliability is the share of the spread in scores that comes from real differences between people. The standard error of measurement (SEM) turns it into points: a score of 72 with an SEM of 6 is written 72 ± 6.
Two scores less than two SEMs apart do not show that one person is stronger. Employer reports print the SEM beside each score, and treat every score within two SEMs of the top as level with it. We use Cronbach's alpha on a fixed exam, and split-half reliability with the Spearman-Brown correction on an exam drawn from the question bank. A figure appears once an exam version has 30 completed, scorable results.
| Exam version | Results | Reliability | SEM |
|---|---|---|---|
| Applied AI Foundations, version 4, form 10 | 8 | Not yet measured: 8 of 30 | |
| Applied AI Foundations, version 4, form 9 | 2 | Not yet measured: 2 of 30 | |
How these figures are made, with the fairness screen and the agreement between examiners.
Known limits
These are the gaps in the evidence today. We publish them so you can weigh a score for what it is.
- Applied AI Foundations and Applied AI Certified (L1) are approved provisionally. No evaluation study and no independent review of their structure has been done yet, so the evidence behind their scores is the design alone.
- No study yet links a score to how someone works with AI on the job. This is the weakest link for every use, and the study is planned, not done.
- The scoring guides the AI examiners are checked against were drafted with AI and reviewed by the Institute. No independent expert has scored them yet.
- The free exam samples two tasks, so it cannot show the full range of the Standard. Its result says so.
- Slow typing and less fluent written English can lower a score for reasons outside the skill. We are measuring both effects.
- Correcting the assistant on a specific point is part of briefing it, but it is not yet scored on its own.
- Item response theory needs about 200 candidates per question. No question has that many yet, so scores use classical methods.