Perspectives

Straight talk on CX, AI, and delivery.

Opinions formed in practice, not in PowerPoint. No hedging, no vendor worship - just what we've seen work and what we've watched fail.

14 articles

Our thinking on contact centres, AI in CX, the metrics that mislead, and what transformation that actually sticks looks like in practice.

No-code CX design has arrived. The sign-off process hasn't.

AWS just showed contact centres how to let business users build agentic CX flows without engineering. The tooling is ready. Most organisations' governance for it is not.

Everyone's contact centre has AI. Almost nobody has redesigned the workflow.

Generative AI adoption in customer service is now baseline. Redesigning the workflow underneath it isn't - and that gap is where deployments actually fail.

Agentic AI in the contact centre needs its own governance layer

85% of contact centres now blend human and AI agents in live customer conversations. Almost none have built the governance layer that blend actually requires.

Voice AI has arrived in the contact centre. Here's what production looks like.

Production deployments grew 340% in a year. The technology works. The gap between demos and production reality is telling you exactly where the work still needs to go.

The contact centre transformation that actually stuck

Nine out of ten contact centre transformations fail. Not technically - organisationally. Here's what the ones that lasted had in common.

The agent attrition problem nobody talks about

Forty percent of contact centre agents leave within twelve months. The industry treats this as inevitable. It isn't.

Why your CCaaS migration will fail

Modern platforms are genuinely capable. They fail because organisations treat this as an IT project when it is fundamentally a business transformation.

CX platform shootout: Salesforce vs ServiceNow vs Zendesk vs Freshdesk

Choose the wrong platform and you'll spend two years on something that makes your agents slower. Here's how to decide without the noise.

The real cost of 24/7 customer service

Every brand wants to be always on. Before you promise customers that someone will always answer, ensure someone competent actually will.

How to measure customer satisfaction without annoying your customers

Response rates to post-call surveys have collapsed. Your customers do not exist to provide feedback. There is a better way.

Why your chatbot handles 40% of enquiries but solves 0%

Your chatbot dashboard looks impressive. But containment is a meaningless metric because it is trivial to manipulate.

First call resolution is killing your customer experience

Most contact centres celebrate FCR like the holy grail. Obsessively chasing it creates perverse incentives that damage customer relationships.

Omnichannel means nothing if your agents cannot access history

Most 'omnichannel' operations are multi-channel operations pretending sophistication. Here's what the gap actually costs you.

AI agents are not the future of customer service

The hype machine has it wrong. Companies that chase full automation disappoint customers and burn budget. Here's what actually works.

18 articles

On pilots, rollouts, operating models, and the gap between AI hype and AI value - including why 95% of GenAI programmes produce no measurable P&L impact.

The model labs are becoming systems integrators. That changes who you call.

OpenAI, Microsoft and Amazon are building multi-billion-pound enterprise delivery arms. Before you hand them your AI programme, understand what they're actually incentivised to sell you.

Over 40% of agentic AI projects will be cancelled. Here's why.

Gartner predicts over 40% of agentic AI projects face cancellation by 2027. The problem isn't the agents. It's everything around them.

$30 billion in GCC data centres. Zero guarantees.

The Gulf is building AI infrastructure at an extraordinary pace. Compute capacity alone does not translate to AI value.

You've deployed AI agents. Now who's governing them?

54% of enterprises have AI agents in production. Only 21% have a governance model for them. This is the unaddressed risk in most AI programmes right now.

The 79% problem: why enterprise AI programmes stall

79% of enterprises report serious challenges adopting AI - a double-digit increase from last year. The gap isn't strategy or technology. It's delivery discipline.

Hybrid CX teams: designing the human-AI operating model

The question is no longer whether AI belongs in your CX operation. It's how you design a team where humans and AI actually work together.

When AI meets customer experience: three rules that keep you out of court

AI in customer experience is not a technology problem. It is a risk management problem.

The skill gap nobody is training for

The skills required to work effectively alongside AI are completely different from what we've been training for decades. Almost nobody is teaching them.

Bahrain's FinTech Regulatory Sandbox - The GCC's Most Open AI Testing Ground

The CBB sandbox lets you test AI in credit, fraud, and customer applications with regulatory tolerance unavailable elsewhere in the Gulf.

Why your GenAI pilot will join the other 95%

MIT reviewed 300+ AI initiatives and found 95% produce no measurable P&L impact. The failure mode is consistent, almost entirely self-inflicted, and it has nothing to do with the model.

The technology worked. The rollout didn't.

AI deployments do not fail because of the technology. They fail because of behaviour, expectations, and measurement. Here's the pattern - and how to break it.

Your people are not the bottleneck. Your operating model is.

Companies are buying AI and bolting it onto operating models designed for a different era. The result is neither efficient nor human. Here's what actually needs to change.

Your AI agents are only as good as their plumbing

APIs gave AI reach. MCP is becoming the standard for reliable agent-tool connections. A2A is what makes multiple agents work together without chaos. The infrastructure layer is now the competitive advantage.

What constitutes a successful AI delivery

Most organisations measure AI success by what the system does. The better question is what it achieves. Four markers that separate genuine delivery from a project that happened to go live.

Your customers don't want a conversation. They want a resolution.

Conversational AI set the wrong expectation. The next era of CX isn't about AI that responds - it's about AI that executes. Reliably, auditably, and without losing control.

Why Western AI Models Fail in Arabic - and What to Do About It

Arabic NLP is not a translation problem. It is an architecture problem - and most Western AI deployments in the Gulf haven't solved it.

Why most AI pilots fail - and how to fix them

Everyone has an AI pilot. Few have AI in production. The pattern is consistent - and so is the fix.

Off-the-shelf AI vs custom build: how to choose without bankrupting yourself

The vendor demo was slick. Six months later you're on their third escalation engineer. Here's how to navigate the gap.

6 articles

What makes large programmes deliver - and why most £5 million initiatives quietly fail. Including how to choose between in-house, boutique, and the Big 4.

The Big Four cut the graduate class. The invoice arrives in 2035.

PwC scrapped its hiring pledge and cut headcount. Graduate intake across the Big Four is down as much as 29% in two years. GCC clients who lease their delivery capacity should be asking who leads their programmes a decade from now.

Outcome-based contracts are quietly rewriting the AI programme business case

A quarter of one major firm's global AI fees are now outcome-based, not time-billed. Here's what that shift means for how you scope, price, and govern an AI programme.

The consulting mid-market is disappearing. Good.

AI is splitting consulting into global-scale platforms and specialist boutiques. The firms stuck in between are running out of time.

Why your £5 million programme will deliver nothing

The business case was elegant. The steering committee had PowerPointed for months. Then implementation started, and everything slowed down.

In-house team vs consultant vs Big 4: a decision framework

You need expertise you don't have. Should you hire, contract, or call the Big 4? The answer depends on what you're willing to pay for bureaucracy.

The implementation roadmap that actually works

65–70% of CX technology implementations fail to deliver their promised benefits. This roadmap is built on 12 years of programmes that stick.

12 articles

Regulation, ethics frameworks, and data sovereignty - covering what the GCC's emerging AI governance landscape actually requires from firms delivering programmes in the region.

Brussels delayed the AI Act's hardest rules to 2027. The disclosure duty didn't move.

The EU AI Act's high-risk system obligations were pushed to December 2027 and August 2028. Its chatbot and AI-content disclosure duties activated on 2 August 2026 anyway - and that's the one that lands on CX teams.

Saudi Arabia solved the AI training-data question. The answer has three conditions.

Saudi Arabia's new Copyright Law enters into force on 12 August with the Gulf's first statutory AI training-data exception. The exception comes with three conditions - and the conditions are where the risk sits.

Saudi Arabia's AI law isn't finished. Its enforcement already is.

SDAIA's draft risk-tiering policy is still in consultation. Its enforcement committees have already issued 48 PDPL rulings and fines up to SAR 10 million. Here's the gap procurement teams are missing.

The first binding law for AI agents just took effect. It wasn't written in the Gulf.

China's Implementation Opinions on AI Agents became enforceable on 15 July 2026 - the world's first dedicated regulatory regime for agentic AI. No GCC jurisdiction has anything like it yet. That won't last.

The UAE built one front door for AI and data. Most firms are still knocking on three.

The UAE's new Federal Authority for Artificial Intelligence and Data consolidates three bodies into one. Here's what actually changed, and what didn't.

Shadow AI is already inside your GCC compliance perimeter

GCC regulators have moved fast on AI governance. Most institutions still have no idea what AI their own staff are actually using. That gap is the real exposure.

The UAE's central bank just wrote the Gulf's AI governance template

The CBUAE's new guidance note on AI/ML in licensed financial institutions isn't a checkbox exercise. It's the model every GCC sector regulator will copy next.

$30 billion in GCC data centres. Zero guarantees.

The Gulf is building AI infrastructure at an extraordinary pace. But compute capacity without governance, talent, and operating models is a stranded asset.

GCC AI compliance is no longer optional - a practical guide for market entrants

The UAE, Saudi Arabia, and Bahrain have all moved on AI regulation in 2026. Here's what market entrants actually need in place before they deploy.

Oman's National AI Ethics Charter - What It Requires from Consultants

Principles-based governance requires more judgement than prescriptive rules - not less. Here's how to build ethics review into an Oman AI programme.

AI Regulatory Requirements by GCC Country - A Practical Comparison

UAE, Saudi Arabia, Qatar, Bahrain, Oman, Kuwait: each has a different regulatory posture. Here's what it means in practice.

Data Sovereignty in the UAE and KSA - What It Means for Your AI Programme

A data residency addendum is not sufficient. Here's what sovereignty actually requires - and why the audit trail matters more than where the data sits.

15 articles

Market-specific analysis across the UAE, Saudi Arabia, Qatar, Bahrain, Oman, and Kuwait - covering national visions, procurement realities, talent constraints, and what it actually takes to deliver in each country.

The Gulf is building its own internet to escape the Suez chokepoint

Four subsea cable projects worth over $2 billion aim to route Gulf AI data around the Suez Canal and Bab-el-Mandeb strait. None of them are finished, and one can't currently get its cable-laying ships into position.

Saudi Arabia's data embassy law changes the infrastructure calculus for AI programmes

The first G20 nation to let foreign entities keep their own legal jurisdiction over data hosted on Saudi soil. What this means before you build your GCC AI infrastructure.

Dubai's agentic AI mandate: what the private sector actually needs to deliver

Dubai has set a two-year deadline for full private-sector agentic AI adoption. Only 21% of companies have governance mature enough to support it. Here's what closing that gap actually requires.

HUMAIN and the Saudi AI Ecosystem - What Consulting Firms Need to Know

Saudi Arabia isn't just buying AI - it's building it. HUMAIN, SDAIA, and sovereign AI ambition are reshaping how firms need to engage.

Saudisation in Technology Programmes - Designing Delivery That Actually Works

Saudi names on org charts with no real roles is a trap. Here's how to design AI programmes where localisation creates genuine capability.

TASMU and Qatar's National Digital Programme - A Delivery Perspective

TASMU is the vehicle for much of Qatar's government AI agenda. Here's what it covers, how it procures, and what the post-World Cup shift means for consultants.

Delivering AI in Qatar - What Makes This Market Different

A small, tightly connected decision-making community with enormous per-capita ambition. Qatar rewards firms that understand its relationship dynamics.

Using Bahrain as a GCC Market Entry Point - What Works and What Doesn't

The regulatory openness, EDB support, and Saudi proximity are real advantages. The mistake is treating the entry point as the destination.

Delivering in Oman - Why the Long-Term Approach Wins

Oman's consensus-driven culture makes it demanding on delivery pace. It's also one of the most loyal client relationships in the GCC for firms that invest correctly.

Kuwait's AI Infrastructure Gap - and the Opportunity It Creates

Kuwait has the fiscal capacity to lead on AI but consistently delivers less than its peers. Here's why - and why the gap is an opportunity.

Government Procurement in Kuwait - What AI Consulting Firms Need to Know

The CTC, parliamentary scrutiny, and CAIT make Kuwait the most complex procurement environment in the GCC. Here's how to navigate it.

GCC National Visions and What They Mean for Your AI Programme

Six countries, six visions, one critical question: what do they actually require from the firms delivering against them?

What the UAE's ethical AI audit requirements mean for your next procurement

Public sector AI projects in the UAE now require formal ethics audits. We break down what's required, what's not, and how to get ahead of it.

Entering the UAE as a professional services firm - what the D33 agenda changes

The Dubai D33 economic agenda reshapes how international consultancies should position, structure, and price their engagements in the emirate.

The GCC AI Talent Gap - and How to Build Programmes Around It

The UAE has ~7,000 AI specialists. Saudi has ~5,000. Both are well below the demand created by their national vision programmes. Here's how to design around it.