The UK data regulator turned its attention to AI agents this week. Three vendor announcements show where agent tooling is heading.
ICO secures data protection changes from ten AI developers and turns to agents
On 8 October the Information Commissioner's Office published the results of its supervision of the largest foundation model developers working in the UK. Amazon, Anthropic, Apple, Cohere, DeepSeek, Google, Meta, Microsoft, OpenAI and Stability AI have made, or committed to make, changes. These cover clearer information on how personal data is used to train models, better routes for people to exercise their data rights, and tougher assessments of safeguards.
The ICO also opened a six-week call for evidence on the data protection risks of agentic AI, closing on 20 November 2026. It has made enquiries with OpenAI, Anthropic, Meta and the UK's AI Security Institute about recent agent testing.
What it means: the regulator now treats AI agents as a risk in their own right. If an agent in your business reads or acts on personal data, you should be able to say what it can access, what it did and who checked it. Write that down before you scale it up.
Source: Think Digital Partners
Distology signs Harmonic Security for the UK, Benelux and DACH
Distology, the Stockport-based security distributor, has signed an exclusive deal to distribute Harmonic Security in the UK, Benelux and DACH. Harmonic shows which AI tools staff are using, what data they share with them, and which tools were adopted without approval. Its focus is visibility rather than blocking.
What it means: most firms do not know which AI tools their staff use or what goes into them. You cannot write a sensible AI policy without that picture. Get it first.
Source: IT Brief UK
Kore.ai launches Autoloop to tune live AI agents
Kore.ai announced Autoloop on 9 October for customers on the Artemis edition of its Agent Platform. It measures agents against business goals such as task completion, accuracy, safety, customer experience and cost. It then finds where they fall short, makes changes and rechecks the result. Live interactions trigger further rounds of changes.
What it means: an agent that changes itself is a moving target. If you use a tool like this, keep a record of every change, limit what it is allowed to change, and keep a human sign-off for anything that touches customers or money.
Source: IT Brief Canada
Avalara makes its Aviator tax agents generally available
Avalara has made Aviator, its agent hub for tax and compliance, generally available. It was first announced at Avalara's CRUSH conference on 23 September. The agents map products to tax codes, validate exemption certificates and draft tax rules, then simulate the impact before a rule goes live. Each action is recorded, and decisions that need judgement are passed to staff. Avalara points to e-invoicing mandates in France, Germany and Spain, and to EU cross-border B2B e-invoicing from 2030.
What it means: this is the right shape for agents in regulated work. The agent does the repetitive steps, keeps an audit trail and hands judgement to a person. If you sell into the EU, the e-invoicing timetable is worth tracking now.
Source: CFOtech UK
Kontinuity builds AI automation for established businesses. If you want to know which of these actually fits your business, book a call.