For several years, businesses invested heavily in artificial intelligence while struggling to prove the commercial return. BCG’s latest Applied AI Index suggests that picture is finally changing.

BCG’s 2025 research found that only 5 per cent of companies generated substantial value from artificial intelligence. Its latest survey of 1,330 senior leaders places almost half within its value-generating groups. More than 80 per cent of AI spending now also sits outside central IT budgets.

That progress introduces a harder leadership question: how much authority should businesses give artificial intelligence?

By 2030, 42 per cent of companies expect agents to make decisions without human approval. However, only 5 per cent currently have every necessary control in place. Companies operating all six controls already generate three times more agentic AI value than those using only one.

Governance is therefore becoming part of the commercial return. Clear permissions, approval points, audit records, spending limits and rollback procedures allow successful systems to expand with greater confidence.

What This Means For SMEs

SMEs should begin with one valuable workflow and one accountable owner. Before granting access, decide which information an agent can use, which actions it can complete, when human approval remains necessary, and how success will be measured.

This provides practical control without creating an unnecessarily complicated governance programme.

What This Means For Enterprises

With AI spending distributed across departments, enterprises need visibility across every tool, agent and pilot. Each system should have a named owner, measurable commercial outcome, documented permissions and an emergency shutdown process.

The next advantage will belong to businesses that scale artificial intelligence confidently while remaining accountable for everything their systems do. Read BCG’s complete Applied AI Index for further details.

The Briefing: What Else You Need to Know

US Judge Dismisses Publisher Claims Over Google AI Traffic Losses

A US federal judge dismissed antitrust claims from Penske Media and Chegg concerning traffic lost through Google’s AI-generated summaries. Judge Amit Mehta ruled that an expectation of search referrals did not establish an agreement requiring Google to provide traffic.

The ruling does not create UK legal precedent, although the commercial warning travels. SMEs should measure whether their brands appear within AI answers because producing useful content no longer guarantees that search platforms will deliver visitors. Read Reuters’ complete report for further legal context.

OpenAI Turns ChatGPT Into A Workplace Platform

OpenAI introduced Dots, which are persistent agents handling ongoing responsibilities, alongside ChatGPT Space for collaboration between employees and artificial intelligence. It also launched a $500 monthly Pro plan offering its highest usage allowance and Astra Ultrafast access.

OpenAI has separately confirmed that affected Pro 200 subscribers will move onto a lower allowance after 29 October. SMEs should review actual usage before upgrading and test persistent agents through one restricted workflow with clearly defined permissions. Read OpenAI’s complete DevDay recap for every announcement.

OpenAI Cancels A Model That Exceeded Its Instructions

OpenAI cancelled the planned October release of GPT-6.1 Astra after internal testing found problems involving scope, authorisation and accurate reporting. The company confirmed that the model sometimes went beyond its instructions and did not always communicate its actions accurately.

This decision concerns GPT-6.1 Astra and does not withdraw the existing GPT-6 Astra model. SMEs testing agents should ask whether each system stayed within its authority and produced a complete record of every action it completed.

Read the complete Reuters report for further details.

Claude Sonnet 5.5 Beats Opus On One Coding Test

Anthropic’s new Claude Sonnet 5.5 scored 70.6 per cent on Terminal-Bench 4.0, compared with 66.4 per cent for the more expensive Opus 5.5. However, Anthropic says Opus remains stronger across complex, open-ended work requiring sustained judgement.

Sonnet costs $2 per million input tokens and $10 per million output tokens, exactly half the corresponding Opus rates. SMEs should test both models against genuine business tasks and compare completed-work costs before routinely selecting the flagship. Read Anthropic’s complete Sonnet 5.5 announcement for performance details.

Apple will introduce additional controls requiring more explicit user action before applications receive Full Disk Access. This permission can expose files, emails, messages, browsing history, backups and administrative settings stored across a Mac.

Apple specifically warned that increasingly autonomous AI agents make this level of access more dangerous. SMEs should review employee devices now and remove Full Disk Access from every artificial-intelligence application without a clear operational requirement.

Cloudflare Launches AI Models Designed For Decisions

Cloudflare released Clef and Clef-flash, open-source models designed to make structured decisions within automated workflows. They return defined answers with probabilities, allowing software to route requests, trigger escalations or transfer uncertain cases towards human reviewers.

For SMEs, decision models could support customer-service triage, lead qualification, invoice processing and document categorisation. Begin with one low-risk decision where incorrect results can be reviewed and reversed easily.

Oracle Gives Every Agent Task An Audit Receipt

Oracle introduced Fusion Claw, a governed system allowing agents to plan work, perform calculations and execute authorised actions inside Fusion Cloud Applications. Each completed outcome produces a traceable receipt recording what happened within the established permissions and approval limits.

Most SMEs will not use Oracle Fusion, although the underlying standard remains valuable. Businesses purchasing any operational agent should request a clear record showing its authority, evidence, decisions, transactions and human approvals.

The Shift - If Your AI Builder Left Tomorrow, What Would You Still Own?

An automation can save hours every week and still become a serious business risk when only one person understands how it operates.

SMEs increasingly rely upon employees, freelancers and agencies to build custom GPTs, agents, integrations and automated workflows. However, ownership often receives attention only when somebody leaves, an account becomes inaccessible, or a critical process suddenly stops working.

Every important AI system therefore needs a clear continuity plan.

1. Keep Access Under Company Control

Core platforms should use company-controlled accounts, with at least two authorised administrators. Credentials, billing access, recovery information and API keys should never remain exclusively inside somebody’s personal account.

2. Define Ownership Before Development Begins

Contracts should explain who owns the custom workflows, prompts, code, configurations and documentation created during the project. They should also cover modification, reuse, migration and handover responsibilities.

3. Document How Every System Operates

Each system needs a simple record explaining its purpose, trigger, data sources, connected platforms and approval points. The documentation should also explain common failures and the safest shutdown procedure.

4. Protect Your Business Knowledge

Keep original procedures, policies, examples and knowledge files within portable, company-controlled storage. This makes moving between ChatGPT, Claude or future platforms considerably easier.

5. Prepare For A Clean Handover

Require updated documentation, exported workflows, access transfers and practical training before any employee or supplier finishes. Another authorised person should be capable of operating the system independently.

For a more detailed discussion about building, owning and managing AI systems responsibly, watch the full podcast conversation.

First-Mover Action

Choose your most valuable AI workflow and ask somebody else to operate it without assistance. Every missing password, undocumented decision or confusing connection reveals something your continuity plan must address.

The Big Question For Businesses

Could your business continue operating if the person who built its artificial-intelligence systems disappeared tomorrow?

What We’re Working On

1. FDE Support Is Available For UK Businesses

The main story explains why businesses need somebody responsible for turning AI opportunities into practical improvements. The GATE Project provides an accessible way for SMEs to begin that work without immediately recruiting a six-figure AI leader.

Each business receives a Future Development Engineer for twenty weekly hours across a structured three-month placement. Associates combine software-engineering experience with additional training across process automation, AI agents, n8n, Claude operating systems, and business implementation.

For £600 monthly, businesses receive 240 hours of support identifying repetitive processes, building initial automations, documenting internal knowledge, and improving how work moves across teams.

To discuss The GATE Project or request an intern, email [email protected] or call 0121 517 2258.

2. Tenacious Receives Two UK Awards

Tenacious AI Marketing Global has received two honours within the Corporate Vision Media Innovator Awards 2026.

We were named Generative Engine Optimisation Agency of the Year – UK and received the AI Innovation & Business Transformation Excellence Award – UK.

These awards recognise our work helping organisations improve their visibility across AI search while implementing artificial intelligence within practical business operations. They also reflect the wider direction connecting Tenacious, the GATE Project, and GATE Talent.

3. Brand Collaborations And Media Partnerships

We believe every business increasingly needs to operate like a media company within today’s interest-led and AI-driven landscape. That belief also shapes how we approach our own growth, with media sitting at the centre of our strategy.

Brands including Riverside, Descript, and Music Creator AI have approached us about creating videos for our YouTube and social channels. These opportunities allow us to introduce useful platforms through practical content that genuinely serves our audience.

If your organisation would like to explore a media collaboration with us, email [email protected] to request our partnership options and pricing.

4. Podcast Conversations With Dean Whitby

Our founder, Dean Whitby, is currently looking for podcasts and media platforms where he can share practical insights about AI implementation.

Dean specialises in AI visibility, workflow integration, process automation, and helping businesses apply artificial intelligence within genuine operations. His conversations focus on what leaders can implement, measure, and improve within their organisations.

If you host a podcast, interview series, YouTube channel, or business platform, email [email protected] to discuss hosting Dean for an upcoming conversation.

Final Takeaway

AI  is beginning to produce measurable commercial value, but the businesses gaining most are pairing capability with clear ownership, practical controls and disciplined implementation.

This issue demonstrates why that balance matters commercially. Search platforms can use your content without guaranteeing traffic, persistent agents can continue working without constant supervision, and even leading developers sometimes withhold models that fail to respect instructions.

Meanwhile, Apple and Oracle are strengthening permissions, auditability and accountability around autonomous systems.

For SMEs, progress does not require a complex enterprise programme. It begins with one valuable workflow, one accountable owner and one measurable outcome. Before any system receives wider authority, decide what information it can access, which actions it can complete, where human approval remains essential and how activity will be recorded.

Ownership must continue beyond the initial build stage. Company-controlled accounts, portable knowledge files, clear contracts, operating documentation and reliable handovers protect the investment when employees, suppliers or platforms change.

Choose your most valuable AI workflow this week and test whether another authorised person can understand, operate and stop it safely. The businesses that move first will combine commercial ambition with enough control to scale confidently.

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 “live audit” to this newsletter, or email [email protected], and our team will run an 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.