Google is hiring GenAI Forward Deployed Engineers, and it shows where AI adoption is really heading.
These are not traditional consultants. They are embedded engineers who work inside customer environments, build and debug real systems, solve integration problems, and help move frontier AI products from promising demos into production.
The first phase of AI was about access. Companies tested ChatGPT, bought licences, created internal experiments, and encouraged teams to use AI inside their daily work.
That was useful, but it was not the hard part.
The hard part is making AI work inside the actual business. That means connecting it to the right data, fitting it into real workflows, controlling permissions, checking output quality, managing cost, monitoring reliability, and making sure the system keeps working when people, models, pricing or providers change.
Google’s move shows that deployment is becoming the bottleneck.
Datadog’s State of AI Engineering report points in the same direction. More than 70% of organisations now use three or more AI models, and the share using more than six models has nearly doubled.
So the AI market is not becoming winner-takes-all. It is becoming multi-model, task-specific, and harder to manage.
The question for CEOs is no longer only whether to use OpenAI, Gemini or Claude. The better question is which model fits which task, how easily the business can switch models, and whether the system has enough governance around cost, quality, reliability and risk.
Cost is becoming a serious issue, too. Datadog found that 69% of input tokens in customer traces were system instructions rather than user requests. Only 28% of relevant LLM calls used cached-read input tokens, even though caching can reduce cost and improve speed.
Put simply, many businesses may already be paying for avoidable AI waste before their usage has even properly scaled.
Reliability is another warning sign. Datadog found that rate limit errors were the most common LLM call failure type in March 2026. Once AI starts supporting sales, operations, customer service, reporting, finance or internal decision-making, failed calls and slow responses stop being technical details. They become business problems.
This is why the new AI role matters.
The companies that win the next phase will not simply be the ones with the most tools. They will be the ones that can operate AI properly as part of the business.
That means AI engineering, governance, observability, model flexibility, cost control and practical implementation capability.
For SMEs, there is also a continuity risk. If one expensive specialist builds an AI system and leaves twelve months later, the business may be left with something nobody else understands. That does not create resilience. It creates dependency. The next phase of AI adoption will need more than software access. It will need people who can build, manage, improve and maintain AI inside real commercial environments.
The first-mover question is no longer, “Which AI tools should we use?”
It is, “Who can actually make AI work properly inside our business?”
For readers who want the deeper technical layer, the Datadog State of AI Engineering report is the best supporting read.
The Briefing: What Else You Need to Know
Microsoft Is Building The AI Integration Layer For Enterprises
Microsoft has launched Frontier Company with $2.5 billion in backing to help major businesses choose, integrate and customise AI tools across different model providers.
The signal is clear. Enterprise AI is moving beyond chatbot access and into model orchestration, internal data integration, workflow design and measurable return on investment. Microsoft will work with clients, including Unilever and Novo Nordisk. Judson Althoff also said Microsoft made a mistake by originally tying Copilot only to OpenAI models.
The next advantage may come from flexible AI systems, not single-vendor dependency.
Meta May Turn Excess AI Compute Into Cloud Revenue
Meta is reportedly building a cloud business to sell excess AI computing capacity and access to models hosted on its infrastructure. This shows how Big Tech may turn huge AI infrastructure spending into direct revenue, not just internal product capability.
Reuters reported that Meta shares rose more than 10% after the news, while CoreWeave and Nebius both fell sharply. AI infrastructure is becoming its own business model. The companies that overbuild compute may become the next cloud landlords.
National Grid Invests In The Power Layer Behind AI
Britain’s National Grid is investing $1.75 billion in Joulent to support power infrastructure for AI-driven data centre demand. AI is no longer only a software story. It is now reshaping electricity demand, infrastructure investment and data centre economics.
National Grid is taking a 35% stake in Joulent. The first project is a 2.67-gigawatt facility supporting a Microsoft-operated data centre campus. Compute access, electricity contracts, and data centre geography are becoming part of AI business strategy.
Bank Of England Pushes For Agentic AI Controls
The Bank of England has said agentic AI may need bespoke regulation, stronger recovery plans and possible market-wide kill switches.
Autonomous agents are now being treated as a different risk category from ordinary software automation. Deputy Governor Sarah Breeden said existing frameworks were not built for autonomous agents. Reuters also reported that 52% of finance firms are already using agentic AI.
The agentic AI conversation is moving from productivity to control, oversight and fail-safes.
AI Skills Are Becoming The Global Hiring Divide
AI-related hiring is rising even as broader tech hiring remains under pressure. In India, Reuters reported that AI hiring in the IT sector rose 16% year on year in June, while overall IT recruitment fell 3%. Across 14 Indian sectors, AI and machine-learning roles rose 25%, based on Naukri data from more than 150,000 firms.
In the UK, government projections suggest jobs directly involving AI activities could rise from 158,000 in 2024 to 3.9 million by 2035. At the same time, 56% of UK employers using or planning to use AI still rate their organisation’s AI knowledge as beginner or novice.
In the US, Indeed’s Hiring Lab says employers are concentrating limited hiring around AI-related roles and skills, even while the wider market stays weak. BCG also estimates that 50% to 55% of US jobs could be reshaped by AI over the next two to three years.
The pattern is global: companies may hire more selectively, but AI capability is becoming one of the clearest workforce priorities.
CMOs Are Optimising Brands For AI Platforms
Top CMOs are adapting brand and content strategies so companies appear more effectively inside AI platforms such as ChatGPT and Gemini.
Business Insider reported that brands including Coach, American Eagle, Chime, Bobbie and Fruitist are now thinking seriously about AI-platform visibility. The article also cited data suggesting users are 2.5 times more likely to visit a recommended brand’s site after seeing it in AI search.
China’s Z.ai Adds To The Global Frontier Model Race
Reuters reported that Chinese AI startup Z.ai’s GLM-5.2 model is narrowing the performance gap with leading U.S. frontier models. The story is not simply about one model catching another.
It shows that frontier AI capability is becoming more distributed across regions, providers and infrastructure ecosystems.
Reuters said GLM-5.2 rivals leading closed-source models on coding and agent tasks. The model also supports domestic chip infrastructure, including Huawei systems. For business leaders, the point is that AI strategy is becoming more multi-model, multi-region and operationally more complex.
Model choice may increasingly depend on performance, cost, access, infrastructure, sovereignty requirements and the specific task being handled.
That reinforces the wider theme of this issue. Businesses that build flexible AI systems will be better placed than those relying too heavily on one provider, one model or one default setup.
The Shift - AI Adoption Is Becoming a Workforce Design Problem
Most businesses still talk about AI adoption as if the question is whether teams have access to enough tools. That is no longer the right question.
Google is already showing where work is heading. Sundar Pichai said 75% of all new code at Google is now AI-generated and approved by engineers, up from 50% last autumn. He also said Google is moving towards agentic workflows, where engineers manage autonomous digital task forces instead of only writing code themselves.
Google Cloud is also hiring GenAI Forward Deployed Engineers. These are not traditional consultants. Google describes them as embedded builders who work inside customer environments, code and debug real systems, and help move frontier AI products into production-grade workflows.
That is the bigger shift.
The future technical role is not only “a person who writes code”. It is becoming “a person who can build, check, integrate, govern and ship AI-enabled systems inside real businesses”.
The UK government is moving in a similar direction at the national level.
Its AI Opportunities Action Plan focuses on skills, compute, business adoption and the future of work. The government says it has completed 38 of the 50 actions in the plan, designated five AI Growth Zones, committed £2 billion to expand UK compute capacity twentyfold by 2030, and created a Sovereign AI Unit backed by up to £500 million.
So AI is no longer being treated as a software trend. It is being treated as infrastructure for the economy.
What Businesses Should Diagnose First
The first thing to check is whether AI is actually improving work, or simply creating more tool usage.
The UK government’s Digital and Technologies AI Adoption Plan is clear on this. The challenge is not access to AI. Many firms are already experimenting with tools. The real problem is operationalising AI at scale and integrating it properly across workflows, systems and products.
That means businesses need to ask better questions.
Where can AI save time? Which use cases matter most? Who owns the workflow? What data can safely be used? What guardrails are in place? How will the business measure whether AI is improving productivity, quality or customer experience?
The government’s commitments are also becoming more practical. It wants sector-led programmes to identify high-value AI use cases, an AI Adoption Framework to help firms move from experiments into deployment, and expanded support through programmes such as BridgeAI to help thousands of businesses adopt AI before the end of this Parliament.
The skills gap is also more specific than “we need more AI people”.
Businesses need leaders who can choose the right use cases, managers who can redesign workflows, technical teams who can integrate AI safely, and employees who know how to use AI without creating risk.
There is also a junior talent issue. If AI automates the routine tasks that once helped entry-level people learn, businesses need a new way to build experience. The government points to Autotrader as one example, where trainees from graduate programmes, apprenticeships and T-levels made up around 10% of its engineering team in 2026.
The Big Question for Businesses
The big question is not, “Are our people using AI?”
It is, “Are we redesigning the business so AI actually improves how work gets done?”
If the answer is no, AI can quickly become another layer of noise.
Teams may use different tools without shared standards. Managers may struggle to judge AI-assisted work. Junior people may lose the training ground they need. Leaders may see activity everywhere without clear productivity gains.
The first-mover advantage will belong to businesses that move beyond tool access. They will build clearer workflows, stronger training, safer guardrails and better ways to measure AI’s impact.
Because the businesses that win with AI will not be the ones with the longest list of tools. They will be the ones who know how to redesign work around them.
What We’re Working On
1. Live AI Visibility Audits Every Week
We are continuing to run live AI Visibility Audits every week for businesses that want to know how they appear inside AI search.
Each audit looks at whether AI understands the business clearly, whether the brand is being recommended, which competitors appear instead, and what visibility gaps need fixing. This is designed to make AI visibility practical for businesses.
If ChatGPT, Gemini, Claude or Perplexity are already influencing how buyers compare companies, businesses need to know what those systems are saying about them.
The aim is to make AI visibility more practical, more understandable and more useful for real businesses.
If you want your business to be one of the first audited live, email [email protected] with the words “AI Live Audit”.
We will pick two businesses for each live session and walk through their visibility, opportunities and next steps in public.
2. The GATE Project: FDE-Style Support For SMEs
Google is hiring Forward Deployed Engineers to help enterprise businesses move AI from pilots into production. Most SMEs cannot access that level of implementation support at enterprise cost, but the need is still there.
That is where The GATE Project comes in.
Through The GATE Project, businesses can access trained AI implementation interns for £600 per month, working 20 hours per week for three months.
Think of it as an SME-friendly version of the same implementation gap Google is addressing: practical support from people trained to help build agentic AI, improve workflows, and automate real business processes.
The aim is to give smaller businesses access to AI implementation capability without needing to hire a full-time specialist immediately.
To discuss The GATE Project or request an intern, email [email protected] or call 0121 517 2258.
3. The GATE Talent: A Managed AI-Ready Talent Pipeline
We are also building The GATE Talent, a platform designed to help businesses hire skilled candidates across technical, commercial and operational roles.
This includes cyber security, front-end engineering, back-end engineering, full-stack engineering, DevOps, UX, marketing, product management, sales, project management and business management.
The difference is that candidates can also be upskilled in AI engineering, agentic AI and automation. That means businesses are not only hiring for the role they need today. They are hiring people who can help them build, improve and automate how the business works.
The talent will be upskilled every fortnight, so their capabilities continue to develop as AI changes. They will also be managed, trained and supported, with tax and operational admin handled as part of the process.
Businesses will be able to hire talent for £24,000 per year, with the full delivery process managed by our team.
We are currently building the website so businesses can choose the talent they need, then move through a managed 12-week delivery process.
4. The Next 10-Industry AI Visibility Report
We are also working on the second round of our 10-industry AI visibility research. In the first round, we ran 300 buyer-style prompts across 10 industries and three AI platforms to see which companies were being recommended by AI.
This next round will show what has changed since last month. We will be looking at which companies gained visibility, which companies dropped, who the winners and losers were, and what the data tells us across each industry.
The full breakdown will be shared in next week’s newsletter.
Final Takeaway
The first AI phase was about access, experimentation and speed. The next phase is about operation, integration and resilience.
This week’s stories all point in the same direction. Google is hiring engineers to help businesses move AI into production. Microsoft is building an enterprise AI integration layer. Datadog’s research shows companies are already managing multiple models, rising costs and reliability issues. The Bank of England is looking at agentic AI controls, while National Grid is investing in the power layer behind AI demand.
The message for business leaders is clear. AI is no longer just another tool sitting inside the software stack. It is becoming part of how businesses operate, hire, compete, manage cost, protect risk and build future capability.
That also changes the skills question. Businesses do not only need people who can use AI tools. They need people who can redesign workflows, manage AI systems, understand model choice, control spend, build guardrails and keep improving how work gets done.
The first-mover advantage will not belong to businesses with the longest list of AI tools. It will belong to businesses that can make AI work properly inside the organisation.
The question is no longer, “Are we using AI?”
The better question is, “Are we building the people, systems and processes needed to operate AI well?”
Want to Know What AI Is Saying About Your Business?
If you want to understand how visible your business is in AI search and how AI compares you against your competitors, we can help.
Reply “audit” to this newsletter, or email [email protected], and our team will run a AI Visibility Audit so you know exactly what AI is saying about your business.
Is this newsletter right for you?
We send this briefing to help first movers stay close to the AI news, trends, and signals that matter.
If that is useful to you, we would love to keep sharing it with you. If it is not quite what you are looking for, no problem at all, you can unsubscribe using the link below.
Either way, thank you for giving it a read.
