This fortnight, agentic AI stopped being a pilot and became a payroll question. The world's largest firms are rolling agents out across entire workforces, and the Gulf is wiring them into government and 295,000 private companies. The technology is no longer the hard part — your people are. So the question every leader should be asking has changed from 'should we adopt agents?' to 'how ready are we?' This edition pairs the news with a practical AI Readiness Scorecard to show you exactly where your organisation stands, and where to act first. As always: the development, why it matters to leaders, and what to do about it.
What's moving in AI worldwide
AI agents go all-in across the enterprise
The agent question has flipped from 'do they work?' to 'can you run them across the whole company?' — and the biggest firms just answered yes.
$1.3tn — projected global AI spending by 2029, led by agentic AI (IDC)
For two years the enterprise debate was whether AI agents could do real work. In June 2026 it moved on. KPMG — 276,000 people across 138 countries — announced it is putting Microsoft 365 Copilot in every employee's hands and standing up Microsoft Agent 365 as the control layer that manages, monitors and secures the agents it and its clients deploy. The stated goal is to move 'from isolated pilots to trusted enterprise-scale AI deployment', with governance, visibility and accountability as the foundation.
The money backs the story. IDC expects worldwide AI spending to grow about 32% a year and reach $1.3 trillion by 2029, with agentic AI the dominant driver — and warns of a tenfold rise in the number and complexity of enterprise AI agents over five years. This is no longer a discretionary line in the innovation budget; it is becoming a structural shift in where IT money goes.
For leaders the competitive question has changed. It is not 'should we pilot agents?' but 'can we deploy and govern them across the whole organisation?' The firms answering yes are pairing scale with control — treating AI as a managed enterprise platform, not a series of demos.
Why it matters: AI has shifted from departmental experiment to a board-level budget and governance decision. Treat it as a governed platform, not a pilot.
Source: Microsoft — KPMG and Microsoft scale enterprise AI agents globally
Big tech turns the model into a commodity
With capable models now everywhere, the winners are building the layer that runs all of them — and driving the cost of intelligence down.
10x — expected rise in enterprise AI agents over five years (IDC)
The headline from Microsoft Build 2026 was not a single model but a posture. Microsoft positioned Azure AI Foundry as a multi-model 'control plane', offering OpenAI, Anthropic, Mistral, DeepSeek and its own MAI family on one platform with built-in compliance logging, access controls and usage metering. Alongside it, Microsoft launched its own MAI models explicitly to reduce dependence on OpenAI and lower costs for developers.
The strategic read is that frontier-model performance is converging and commoditising. When several providers are 'good enough', value moves to the layer that lets an enterprise route work to the cheapest or most compliant model, switch when prices or capabilities change, and prove governance across all of them.
For buyers the takeaway is concrete: evaluate AI on the strength of the orchestration and governance layer, negotiate on cost-per-token across providers, and design for multi-model from the start. Betting the business on one model — or one vendor — is now the riskier choice.
Why it matters: Raw model performance is commoditising. Procurement leverage and avoiding lock-in now matter more than picking a single 'best' model.
Source: CNBC — Microsoft unveils new AI models to lessen reliance on OpenAI
AI in the region: UAE & KSA
Dubai brings agentic AI to 295,000 companies
Dubai isn't just automating government — it's pushing agentic AI into every private company, and the clock is two years.
295,000 — Dubai companies to be empowered with agentic AI within two years (Dubai Chamber)
In June 2026 Dubai extended its agentic-AI ambition from government to the wider economy. Under a programme directed by Crown Prince Sheikh Hamdan bin Mohammed and run through the Dubai Chamber of Commerce, the emirate aims to empower 295,000 companies with agentic AI over two years, deliver 100 specialised AI agents, and seed 50 new agentic-AI companies — supported by training tracks, incubators and dedicated funds.
It sits on top of the federal move announced in April: a directive to shift 50% of UAE government services to autonomous, agentic AI within two years, with specialised AI training for all federal employees. Together they make the Gulf's intent unambiguous — agentic AI as the default operating model, public and private.
For employers the implication is a sharp, simultaneous spike in demand for people who can deploy and govern agents. Building that capability purely in-house is slow; the firms that move fastest will combine reskilling with flexible, compliant access to specialist talent.
Why it matters: Agentic AI is now a competitiveness expectation for essentially every business in Dubai. The binding constraint is talent, not technology.
What to do: Map your highest-volume, lowest-judgement workflows for agents now, and pair the plan with a workforce plan — reskill staff and secure flexible access to specialists.
Source: Gulf Today — Sheikh Hamdan reviews Agentic AI roadmap for 295,000 firms
Saudi Arabia's 'Year of AI' scales sovereign compute
Saudi Arabia made 2026 its Year of AI — and is pouring capital into the sovereign compute to back it.
$1.2bn — HUMAIN–NIF framework to build up to 250 MW of AI data-centre capacity (Arab News)
Saudi Arabia has put AI at the centre of national policy, with the Cabinet naming 2026 the Year of Artificial Intelligence and accelerating SDAIA's national data-and-AI strategy. The most concrete expression is infrastructure: HUMAIN, backed by the Public Investment Fund, agreed a framework of up to $1.2 billion with the National Infrastructure Fund to build as much as 250 MW of AI data-centre capacity, part of a wider push to keep compute, models and data in-Kingdom.
The ambition is sovereign AI — not consuming capability built elsewhere but owning it. That requires capital, which the Kingdom is committing, and people, which is the harder constraint. Saudi AI-sector companies raised $9.1 billion across 70 deals in 2025, yet demand for AI-capable talent is outrunning local supply.
For employers and partners in the region, the message mirrors the UAE: pair every AI and infrastructure plan with a workforce plan. The compute is coming; the competitive edge will go to those who have the people ready to use it.
Why it matters: KSA is building the physical foundation for AI at national scale. As in the UAE, the gating factor shifts to the people who can run it.
What to do: Treat data-residency and sovereign-AI expectations as a design constraint now, and build talent pipelines ahead of the compute coming online.
Source: Arab News — PIF-backed Humain secures up to $1.2bn to expand AI infrastructure
AI agents & your workforce
How to keep hiring fair and fraud-proof when candidates use AI
The fix for AI in hiring isn't banning it — it's using governed AI screening with humans firmly in control. Here is how to stay fast and fair as both sides of the table adopt AI.
Five steps to trustworthy AI-assisted hiring
- Verify, don't just rank. Use AI to flag inconsistencies and check claims against evidence, not only to score and sort CVs.
- Score against explicit criteria. Define must-haves per role so an agent such as TASC's Hyrra.ai scores on clear, auditable rules rather than vague 'fit'.
- Keep a human on every decision. Let the agent rank and summarise, but a recruiter owns every shortlist and every rejection.
- Be transparent with candidates. Tell applicants where and how AI is used; openness builds the trust the data shows is missing.
- Audit who gets screened out. Review rejections for bias and error, not just who passes through.
Used with guardrails, AI makes hiring both faster and fairer — catching deception while keeping a human accountable for the call.
How to prepare your workforce for agentic AI
The Gulf's mass rollouts now assume agents will handle routine work. Here is how to get a team ready without stalling the business.
Five steps to an agentic-AI-ready team
- Map roles, not just tasks. Identify which parts of each role agents will take, and which still need human judgement.
- Run role-specific training. Teach a recruiter, an analyst and an ops lead what changes in their workflow — not generic AI awareness.
- Name an AI champion per team. Pick a respected local who can translate between the technology and the day-to-day work.
- Redeploy the time you free up. Point recovered hours at higher-value work, and say so, to turn fear into buy-in.
- Close the gap with flexible talent. Reskill where you can; bring in specialists where reskilling is too slow.
Agentic AI rewards organisations that treat it as a people change, not just a tech rollout. Prepare the workforce and the technology pays off; skip it and the pilots stall.
The AI Readiness Scorecard
Scorecard. Agents now do the work; the advantage is in your people — so the first question is how ready they are. This scorecard turns AI readiness into something you can measure. Score your organisation across six dimensions — role and task mapping, workforce AI literacy, work redesign and governance, reskilling and redeployment, flexible talent access, and measurement — to see where you stand and where to act first. Built on BCG's 10-20-70 rule (70% of AI value is people and process) and the WEF's finding that 39% of job skills will change by 2030, and calibrated for UAE and KSA enterprises moving fast.
Spotlight: AI-Assisted Hiring with Hyrra.ai
Hyrra.ai. Hyrra.ai puts AI agents to work on the CV pile — automating first-pass screening and cutting time-to-shortlist, with a recruiter in control of every shortlist and rejection.
