Augment Now, Augment Soon, Keep Human: Putting AI to Work in a Qatari SME

Nearly one in three small businesses in advanced economies already uses generative AI. An OECD survey of more than 5,000 SMEs in seven countries, published in November 2025, found generative AI in use at 31% of them. Those firms report better performance and say the tools help them cover skill gaps and labour shortages. The harder question for a Qatari business owner is not whether to use AI but where. Which tasks should a 20-person trading company, clinic or agency hand to a tool this quarter, which should wait, and which should stay firmly with people? In June 2026 Qatar's Ministry of Communications and Information Technology (MCIT) published a method that answers exactly that. This article turns it into an exercise a small leadership team can complete in an afternoon.

What the evidence says about SMEs and AI

The OECD study, Generative AI and the SME Workforce: New Survey Evidence, draws on a representative 2024 survey of SMEs in Austria, Canada, Germany, Ireland, Japan, Korea and the United Kingdom. Its headline findings are worth noting:

• Adoption is mainstream, not fringe. 31% of surveyed SMEs use generative AI.
• Users report gains. SMEs say the tools improve performance and help offset skill gaps and labour shortages.
• Skills demand rises, not falls. Use of generative AI is linked to greater demand for highly skilled workers.
• The worries are legal, not existential. The main concerns are copyright and legal or regulatory issues, and negative attitudes are rare.

Qatar is not in the sample, so treat these as international benchmarks rather than local data. The pattern is still useful: the businesses getting value are the ones that redesign specific tasks and train their people, not the ones that simply buy licences.

Qatar's playbook: the MCIT AI Workplace Guidelines

MCIT's AI Workplace and Job Transformation Guidelines (June 2026, version 1.0.0) are non-binding recommendations for both private-sector organisations and public bodies. They are written for organisations generally, not SMEs specifically, but their logic scales down well. The method has two phases:

• Phase 1 – Redesign the work: diagnose current work, assess AI augmentation, define human controls, and recombine tasks into future roles.
• Phase 2 – Empower and upskill: translate the redesigned roles into skills, and embed learning in daily work.

The core of the method is a simple two-question test applied to each task, not each job.

The Can/Should test, step by step

1. List the tasks. The guidelines ask you to write each role as a plain list of tasks in an "action, object, purpose" format, using job descriptions, procedures and short conversations with the people who do the work, then check the list with them. In an SME, "prepare monthly supplier statements to reconcile payables" is a task; "accounts" is not.

2. Score "Can" from 1 to 3. This asks whether a tool could realistically help without months of preparation. A 1 means data is scattered or missing, rules are unclear and systems are not connected. A 3 means the data and rules are clear, access is in place, and the task is repetitive with steady volume.

3. Score "Should" from 1 to 3. This asks whether AI fits your duty of care: what happens if the output is wrong, and how easily would you catch it? A 1 means a high risk to rights, safety or trust, or heavy reliance on judgment and empathy. A 3 means low risk and output that is easy to review.

4. Sort, never average. The two scores stay separate:

• Augment now: both scores are 3.
• Augment soon: both are at least 2, but not both 3.
• Keep human-led: either score is 1.

Where the team is split between 2 and 3 on "Should", MCIT recommends taking the lower score unless clear human review and audit trails already exist. For a small firm, typical results might look like this (illustrative only): drafting first-pass product descriptions or meeting summaries lands in "augment now"; reconciling invoices across two disconnected systems sits in "augment soon" until the data is cleaned up; and decisions on staff performance, credit to customers or anything clinical stay human-led.

Build the guardrails before you scale

The guidelines are explicit that people remain accountable. For every task you augment, decide in writing:

• Decision boundaries: what the AI may propose, what only a person may decide, and who is accountable.
• Review triggers: low confidence, missing or conflicting data, and any action that could deny or restrict a service should go to a person.
• Audit logs: keep a record of AI recommendations, human approvals, overrides and final decisions.
• Transparency with staff: tell employees how AI is used, which tasks it supports and what data it relies on. MCIT states that AI "should not be used for intrusive surveillance", that AI insights in performance reviews are only one input for a human manager, and that staff should be able to challenge AI-supported assessments that affect them.

Security belongs in the same conversation. MCIT points organisations to the National Cyber Security Agency's guidelines on the secure adoption and use of AI, and asks you to weigh data sensitivity, how autonomous the tool is, and what systems it can reach before switching it on. As general guidance, set a simple rule on which customer, employee and financial data may never be pasted into public AI tools. Our earlier piece on cyber readiness for Qatar startups covers the wider controls.

Measure quality, not just speed

The easiest metric to celebrate is time saved. MCIT asks for more. Track quality and outcome measures (including error rates, rework and customer satisfaction), control measures (including human override rates and escalation rates) and employee confidence, and read productivity figures alongside them. A rising override rate is an early warning that a task was put in the wrong bucket. A falling one, with stable error rates, is your evidence to move an "augment soon" task forward.

How this connects to Qatar Vision 2030 and NDS3

The MCIT guidelines anchor themselves in national strategy. They cite Qatar National Vision 2030's aim to "foster innovation in a knowledge-based economy", the Digital Agenda 2030 goal to "build a digitally savvy population", and the Third National Development Strategy's treatment of IT and digital capacity as "a core enabling sector for economic diversification". Qatar's National AI Strategy rests on six pillars, among them business, employment and ethics. SMEs that redesign work carefully and upskill their teams are where those national goals become day-to-day productivity.

Practical support exists too. MCIT's SMEs Go Digital programme offers digital maturity assessments, transformation roadmaps, training with Microsoft and Google Cloud (including AI-focused sessions), matchmaking with approved technology providers and possible financing through Qatar Development Bank. According to QNA, eligible businesses are registered in Qatar with up to 250 employees and annual revenue of up to QAR 100 million. If you are building AI products rather than using them, see our piece on compute capacity in Doha.

What to do next: a one-afternoon exercise

• Pick two or three pilot roles with high-volume, repetitive work, such as customer service, finance administration or marketing.
• List 10–20 tasks per role in "action, object, purpose" form, and check the list with the people who do the work.
• Score each task for Can and Should separately, and sort into augment now, augment soon or keep human-led.
• Start with no more than three "augment now" tasks, each with a named human reviewer and a short written rule on what data the tool may see.
• Tell your team what is changing and why, and how they can challenge AI-supported decisions.
• Review after 30 days using error, rework and override rates alongside time saved, then decide what moves up a bucket.
• Check eligibility for SMEs Go Digital if you need help with assessment, tools or training.

This article is general guidance. The MCIT guidelines are non-binding, and the OECD figures come from countries outside Qatar.

Sources

• Generative AI and the SME Workforce: New Survey Evidence – OECD (4 Nov 2025) – oecd.org
• Generative AI and the SME Workforce – Full report, summary chapter – OECD – oecd.org
• AI Workplace and Job Transformation Guidelines (DIA-GUI-004, v1.0.0, June 2026) – Ministry of Communications and Information Technology – mcit.gov.qa
• Qatar's National AI Strategy – Ministry of Communications and Information Technology – mcit.gov.qa
• MCIT advances digital transformation in small and medium-sized enterprises – Qatar News Agency (19 Feb 2026) – qna.org.qa
• The National Cyber Security Agency launches guidelines for safe artificial intelligence use – Mondaq (2024) – mondaq.com

Photo: Nataliya Vaitkevich / Pexels