AI Leadership Weekly · Issue #93 · Tuesday 21 July 2026 · 08:00 BST
Good morning.
Google is expanding an existing research workspace, IBM is adding AI-assisted operations to established systems, and Vids is gaining more ways to create and edit video. These changes may reach teams through software they already recognise. That can simplify adoption, but it does not remove the need to check launch dates, account eligibility and the work that remains with people.
IN 60 SECONDS
NotebookLM becomes Gemini Notebook. Google announced Gemini Notebook, building on NotebookLM with expanded analysis capabilities and staged availability across account types.
IBM updates Power systems. IBM announced new Power systems and software, including AI-assisted operations.
Vids adds personal avatars. Google introduced new video-editing and personal-avatar capabilities in Vids, with access and usage limits depending on the plan.
CEO / COO / CXO CHECKLIST
CEO: Identify one useful improvement within software the business already uses.
COO: Distinguish launch announcements from features actually available to the team.
CXO: Check account eligibility, source permissions and the cost of adoption.
TOP STORIES
1. NotebookLM becomes Gemini Notebook
Google · 16 July 2026
What happened. On 16 July, Google renamed NotebookLM to Gemini Notebook. A cloud-computer update lets it write and run code against source material. Initial access covers Google AI Ultra users and qualifying Workspace customers; broader Pro access was due over the following weeks. It remains a standalone product.
Why it matters. The interesting change is the move from discussing documents towards analysing their contents. For a business team, a useful trial would combine the written explanation with the calculations that support it.
A fluent summary and a convincing chart should not be treated as independent confirmation of each other.
What to do. Use an approved, non-sensitive analysis you have already completed. Ask the tool to reproduce it, show the calculation and identify its sources. Compare the result with the original before trying unfamiliar material.
2. IBM updates Power systems
IBM · 15 July 2026
What happened. IBM announced Power Autonomous Operations on 15 July, an agent intended to monitor systems and help resolve capacity issues. General availability was expected on 23 September. Its claim of up to 15-times faster resolution came from an internal, controlled test involving eleven systems and human approval—not a measured customer-wide result.
Why it matters. This is an option to consider for improving the operation of an existing platform, rather than assuming that AI value requires replacing it. Whether it is worthwhile will depend on your workload, current support effort and reliability requirements.
What to do. Ask the platform owner which routine jobs consume specialist time and what failure would cost. Use that evidence to evaluate a future pilot. Do not make the vendor’s internal speedup your savings forecast.
3. Vids adds personal avatars
Google · 16 July 2026
What happened. Google began rolling out Gemini Omni and personal avatars in Vids on 16 July. Users can generate or revise clips through text, and create an account-linked avatar from a selfie and voice recording. Availability includes eligible Google AI and Workspace customers; avatars have age and regional limits.
Why it matters. For suitable work, this could lower the effort of producing and updating short explainers. The more interesting opportunity may be keeping useful material current, rather than producing more material nobody requested.
Choose carefully where an avatar adds value. Routine instructions and a sensitive message from a manager are not the same communication problem.
What to do. Test a short onboarding or product explainer. Judge comprehension, correction effort and ease of updating—not only how realistic the presenter looks. Make the use of an avatar clear to viewers.
SIGNALS FROM THE LAST MONTH
2 July · Microsoft funds hands-on implementation. Microsoft announced a $2.5 billion initiative involving 6,000 industry and engineering experts to work alongside customers on AI implementation. Source
30 June · Sonnet 5 arrives. Anthropic launched Sonnet 5 with introductory API pricing of $2 per million input tokens and $10 per million output tokens. Source
24 June · Gemini 3.5 Flash adds computer use. Google announced computer-use capabilities for Gemini 3.5 Flash, extending how the model can interact with software interfaces. Source
23 June · Azure adds an Observability Agent. Microsoft announced general availability of Azure’s Observability Agent, applying agents to the investigation and operation of cloud services. Source
IN BRIEF
More dated updates from the preceding 30 days.
9 July · OpenAI launches GPT-5.6. OpenAI released GPT-5.6 in multiple variants, giving customers different capability and cost options across its product and developer routes. Source
9 July · AlphaEvolve becomes available on Cloud. Google announced general availability of AlphaEvolve on Google Cloud, bringing its algorithm-optimisation approach to customer workloads. Source
7 July · FireSat adds three satellites. Google reported the launch of three additional FireSat satellites, extending work on AI-assisted detection of wildfires. Source
6 July · ChatGPT comes to PowerPoint. OpenAI made ChatGPT for PowerPoint generally available, offering in-application assistance with editable presentations, subject to account limits and workspace settings. Source
THE 15-MINUTE PLAYBOOK
Pick one proposed AI use case and the software the team already uses for that work.
Minutes 0–4: Describe the missing capability without naming a vendor. “Compare these documents and explain the differences” is more useful than “we need an AI platform”.
Minutes 4–8: Check available features, account entitlements and administrator settings. Record anything still on a roadmap separately. Do not assume a feature in a demonstration is included in your subscription.
Minutes 8–12: Identify the adoption work: information access, user training, testing, support and any changes to the way the team operates. Existing software still needs to earn its place in the process.
Minutes 12–15: Choose between testing an existing feature, buying an additional capability or building something specific. Write down the reason and what evidence could change the decision.
The aim is not to avoid investment. It is to avoid paying twice for the wrong answer.
DATA WAVE MOMENT
A useful technology decision starts with a clear view of the work and the capabilities already available. Build where the business needs something distinctive. Reuse where a suitable capability already exists. Improve the process when another tool would simply make the existing friction more expensive.
That is not less ambitious than buying a new platform. It is a better starting point for deciding where ambition belongs.
QUESTION FOR READERS
Which proposed AI project should begin with a closer look at the software you already have?
Brought to you by Data Wave — your AI & Data Team as a Subscription.
