Trend 1: The Governance Gap is Reaching Crisis Point

AI adoption has raced ahead of the policies and oversight needed to manage it, with research flagging this as a genuine business risk. Without clear governance, organisations are exposed to compliance failures, reputational damage, and costly mistakes, like GDPR breaches from unvetted outputs, a chatbot giving harmful advice, or a flawed AI report shaping a major decision.

• More than half of professionals use GenAI without formal approval; more than a quarter work with no policy at all.
• When senior leaders are asked who owns agentic AI, often the answer is that nobody agrees.
• 38% of large organisations have appointed a Chief AI Officer, but often there is no consensus on where that role reports.

Call to action: Those who treat it as a strategic priority will move considerably faster than those still waiting for clarity to arrive on its own.

1. Treat governance as a leadership priority
2. Assign clear ownership
3. Establish a policy baseline

Relevant courses:

AI Steering Committee | Hiring an AI-First CTO | Helping Your Team Use AI Responsibly

Read more:

LexisNexis: Future of Work Report 2026 | HBR: Who in the C-Suite Should Own AI? | Gartner: The 2026 CIO Agenda | MIT Sloan: Action Items for AI Decision Makers in 2026

Trend 2: AI is Becoming the Front Door to Your Brand and Your Services

Something important is shifting in how customers find, evaluate and choose brands.

• AI is increasingly the first point of contact: answering questions, surfacing recommendations and shaping perceptions before a human ever gets involved.
• The organisations that will be found, recommended and chosen are those that have invested in differentiation and authority, not just discoverability.

Call to action: Audit and optimise your brand's AI presence

1. Test your brand across at least three AI platforms using structured prompts covering brand description, competitive positioning, recommendations, pricing, thought leadership and customer sentiment.
2. Score and analyse the responses across five dimensions - accuracy, positioning, tone, completeness, and sentiment - to identify where gaps exist.
3. Prioritise and act on your gaps, tackling high-impact issues first, then set a regular audit cadence to track improvement over time.

Relevant courses:

How to Audit and Optimise Your Brand's AI Presence

Read more:

HBR: Preparing Your Brand for Agentic AI | FT: The AI Pension Advisers Are Already Here | Gartner: 2026 Priorities for Tech CMOs

Trend 3: The Frontline Is Where AI Pays Off But Only If the Interface Lets It

The most measurable AI returns are showing up in operations, on the shop floor, and in the daily work of frontline teams.

• The organisations pulling ahead aren't deploying ambitious transformation programmes from the top; rather, they're embedding AI into the systems that govern how work actually gets done, and backing it with serious investment in people and skills.
• There's a catch, though: the interface through which people access AI matters enormously.
• Research is showing that the standard chatbot format creates genuine cognitive load that can wipe out productivity gains entirely.

Call to action: Gains disappear when the UI adds cognitive load or fights the job. Design the work first, then decide how AI shows up in it.

1. Redesign workflows first, then choose interfaces that match tasks, not just a default chat window.
2. Invest seriously in people and skills alongside the technology, so changes stick beyond the pilot.

Relevant courses:

Leading Teams Through New AI Workflows

Read more:

McKinsey: How Danone is Reinventing FMCG Operations | BCG: AI-First Hotels - Leaner, Faster, Smarter | BCG: Four Ways GenAI Improves the Lives of Frontline Workers | One Useful Thing: Claude Dispatch and the Power of Interfaces

Tool 1: Gemini Notes (AI Call Transcription in Google Meet)

Available on: Google Workspace Business Standard and above

What it is: Gemini Notes is Google's built-in AI meeting capture for Google Meet on supported Workspace plans: it transcribes the call, drafts a structured summary with action items, and saves the record in Drive next to the calendar event.

We've been recording and transcribing client calls for years using tools like Otter.ai and Read.ai. It's a core part of how we work: a recording lets you focus on the conversation and prevents misunderstandings later. Gemini's built-in transcription changes the equation for Google Workspace users.

Why we love it: There's no extra app, no separate subscription, and the notes land directly in Drive alongside the calendar event. It produces a structured summary with action points automatically, and the quality of the task extraction is genuinely useful.

Limitations: The errors we've seen are human-grade: occasionally missing the importance of something discussed briefly, or misattributing an action when the conversation wasn't clear. We always cross-check against the full transcript. One honest limitation: it only works inside Google Meet. No Zoom, no Teams.

Value: Every call produces a searchable, structured record without anyone doing extra work. That compounds quickly. If your team already lives in Google Workspace, this is worth switching on.

Tool 2: Linear via MCP (AI-to-Task-Tracker Pipeline)

In use for: approximately one year

What it is: Linear is a project tracking tool where you create tasks and assign them to people. Productive meetings generate action items, but manually logging them in a tracker is tedious. By connecting Claude to Linear, we can take meeting notes, extract the action items, and create each one as a task directly from the AI chat.

Why we love it: Claude connects directly to Linear via MCP, so in the same conversation where we're reviewing a transcribed call, we can extract action items and push them straight into Linear as properly formatted tasks with owners and deadlines. For software development clients, turning a requirements conversation into a developer-ready task list used to be a manual step that ate time and introduced errors. Now the pipeline handles the structure and we focus on checking the substance. That review step isn't optional: the AI gets the list roughly right, but it doesn't know your project context well enough to be trusted without a human in the loop.

Limitations: Linear isn't perfect either. If you're used to the freeform flexibility of a spreadsheet for ranking and reprioritising, it can feel constraining. We still drop back to Google Sheets for heavy prioritisation work.

Value: This same pattern works well beyond task tracking. AI extracts structured data from unstructured conversation, a human validates it, and a tool receives it ready to act on. Apply it to CRM updates, status reports, and dozens of other "someone needs to type this up" jobs.

Tool 3: NotebookLM Slide Decks (Source-Grounded Presentations)

Feature launched: November 2025 (per-slide editing and PPTX export added February 2026)

What it is: Google's AI-powered research tool that lets you upload sources, chat with them to extract insights, and generate slide decks from your material.

Why we love it: We've been using this for client-facing presentations about AI adoption trends and strategic recommendations. AI moves fast and we don't want to spend half a day on artifacts that need updating in a month. Our workflow: build a proper deck outline in Claude first with the key points per slide clearly specified, then upload that into NotebookLM as a source and let it generate the visual presentation. The outputs are polished, visually coherent and grounded in the source material you provide.

Limitations: Each slide is rendered as an image, so you can't click in and fix a typo. Google added per-slide AI revision in February 2026, which lets you prompt changes to individual slides without regenerating the whole deck. It's not a good fit for businesses wanting branded templates or direct control over post-production editing.

Value: Building a structured outline in one AI tool and handing it to another for visual execution separates the thinking from the production. You invest your time on what each slide needs to say, and delegate the formatting to a tool that does it faster than you ever will.