Business & Technology
NatWest launches AI ethics accreditation for all staff
KAREN JOY BACUDO
Finance Editor
NatWest has launched a bank-wide accreditation in AI and data ethics for staff, available to all 60,000 employees across the group.
The programme is designed to help employees use artificial intelligence responsibly in day-to-day work as AI tools become more common across the organisation. It follows an earlier course developed with the University of Edinburgh and delivered to about 90 employees across different business areas and roles.
Under the new scheme, staff will complete eight e-learning modules linked to NatWest’s AI ethics principles, followed by a half-day collaborative session focused on practical application. The course also includes guidance on identifying, assessing and managing ethical risks when using AI, including how to judge whether AI-driven decisions are fair and how training data can affect outcomes.
Employees who complete the course will receive an internal accreditation from NatWest and are expected to continue learning as AI standards and tools evolve. The programme is expected to take two to three months to complete, with the first modules launching in June and the wider rollout continuing through October.
Wider AI use
The accreditation forms part of a broader push by NatWest to expand the use of AI, data and digital tools across the bank while maintaining oversight of how those systems are used. All employees have completed core AI modules, and more than half have chosen to take additional training.
NatWest has also been introducing generative AI tools for internal use in research, analysis, customer interactions, fraud prevention and complaints handling. For customers, the bank highlighted AI-enabled virtual assistants, including Cora, which can help people manage finances and access more personalised support.
The lender has also been involved in external work on AI governance. It has supported the Responsible AI Network, an initiative that brings together practitioners, academics and businesses to discuss ethical AI, and has collaborated with the Financial Conduct Authority through the regulator’s AI Live Testing programme on security and ethical practices in agentic AI technology.
Dr Paul Dongha outlined the bank’s reasoning for the new training.
“As AI becomes increasingly embedded in how we serve customers and run our bank, it’s important that we equip colleagues with the skills and confidence to use it responsibly. Building on our existing AI and data ethics training, this accreditation gives our colleagues even more practical tools to recognise risks, ask the right questions and make better decisions in their day-to-day roles. By bringing this to all our colleagues, we’re building a strong culture of responsible AI, helping ensure we continue to be a trusted partner to our customers by meeting their needs, improving their banking experience and delivering consistent outcomes,” said Dr Paul Dongha, Head of Responsible AI and AI Strategy, NatWest Group.
Training focus
The original programme with the University of Edinburgh was developed as a bespoke course to improve understanding of AI and its ethical implications in a banking setting. NatWest said it was delivered across three cohorts and combined academic input with case studies intended to help staff apply the lessons in practical situations.
By extending that work across the whole bank, NatWest is seeking to make AI and data ethics training part of routine professional development rather than a specialist activity. The focus on structured learning and internal accreditation suggests it wants a common framework for staff using AI tools in customer-facing and operational roles.
One employee who took part in a pilot of the programme described how AI is already being used in daily work.
“AI is already becoming a really useful part of my day-to-day work, helping me find information more quickly and respond to customers more efficiently while making sure they get the right support. I can see it playing an even bigger role going forward, so having this kind of training will be really important in helping me use these tools confidently, understand where the risks are, and make sure I’m using them in a responsible way,” the staff said.
Business & Technology
Microsoft E7 price shift pushes UK firms to review AI
FLR Spectron has urged UK businesses to reassess their Microsoft licensing ahead of Microsoft 365 E7 price changes, arguing that the shift should prompt companies to review how they manage the spread of AI tools in their systems.
According to figures cited by FLR Spectron, Microsoft’s changes affect the pricing of its existing Microsoft 365 E5 and E3 tiers, while the newer E7 bundle remains at USD $99 per user. E5 rises from USD $57 to USD $60 per user, and E3 from USD $36 to USD $39, changing the economics for businesses weighing a gradual rollout of AI services against a broader bundle.
FLR Spectron’s case rests on more than software cost. Many organisations, it argues, are adding AI agents to enterprise applications without the controls needed to identify, monitor or limit what those systems can access.
That concern reflects a wider shift in corporate software. FLR Spectron pointed to a Gartner forecast that 40% of enterprise applications will feature task-specific AI agents by the end of 2026, up from less than 5% in 2025.
Cost pressure
For buyers already using much of Microsoft’s security, compliance and identity software, the unchanged E7 price could make the bundled option more attractive than assembling comparable products separately. FLR Spectron estimates the package now offers savings of about 15% compared with buying the components individually.
That calculation matters as finance and IT teams come under pressure to justify growing AI budgets. Many businesses have experimented with generative AI in isolated functions, but fewer have developed a licensing and governance model for broad use across the workforce.
FLR Spectron also argued that the issue goes beyond staff using public AI tools. AI agents embedded in business applications can operate with significant permissions while remaining difficult for security teams to detect and supervise.
Kamran Bahdur, Chief Information Officer at FLR Spectron, set out that risk in direct terms.
“Every AI agent a business deploys is a new identity with access to company systems, and most are being brought inside far faster than anyone can secure them. Each one is effectively another door into the network, often holding privileges on a par with a senior administrator, and the majority are invisible to the security team. As organisations rush agents into production, they widen their attack surface with every deployment. E7 addresses that risk head-on by providing businesses with Agent 365, which brings those agents under the same identity, monitoring and audit controls you’d apply to an employee so you can finally see them, govern them and switch them off,” Bahdur said.
Governance gap
FLR Spectron argues that the main challenge for many employers is the gap between AI adoption and internal governance. While software vendors are introducing AI features quickly, organisations often still rely on fragmented identity, security and data management tools.
Microsoft’s E7 offer, it said, seeks to bring together Defender, Entra Suite, Purview and Agent 365 in a single framework. In FLR Spectron’s view, that gives IT departments a clearer way to observe agents, apply policy controls and maintain audit oversight.
The firm also highlighted employee use of unapproved AI tools. Many workers, it said, are already turning to external services without formal authorisation from IT teams, creating risks around data leakage and compliance exposure, including under GDPR.
Bringing AI use into approved systems is one route companies are considering to reduce that risk. Sanctioned tools can be linked to internal identity controls and data policies in ways consumer-facing services often cannot.
Still, FLR Spectron stopped short of saying E7 is the right answer for every organisation. Its suitability, the firm said, depends on whether a business is ready to move from AI experiments to wider operational use.
Fabio Carvalho said many companies remain at an early stage because the basics are not yet in place.
“There’s a big difference between trialling AI and running it. Most organisations are stuck at the first because the foundation underneath isn’t there: identity, data and management are scattered across half a dozen tools, and you can’t safely scale agents on top of that. E7 provides the foundation and infrastructure for businesses to embed AI into their operations, and while the price increase is certainly a prompt, it’s this future-proofing that is encouraging earlier adopters to make the move now,” Carvalho said.
The broader issue for UK businesses is that AI spending is becoming more closely tied to questions of security architecture and software consolidation. As AI agents move from small pilots into day-to-day workflows, licensing decisions are increasingly bound up with who controls those systems, what data they can access and how quickly they can be shut down if something goes wrong.
Business & Technology
HiBob launches Slack integration for HR data access
SOFIAH NICHOLE SALIVIO
News Editor
HiBob has launched a Slack integration for its Bob HR platform, using the Model Context Protocol to bring workforce data into Slackbot.
The integration lets employees, managers and HR teams query Bob from within Slack using natural language and carry out some HR-related actions without leaving the messaging platform. Users can ask about compensation, performance reviews, team joins and upcoming compensation reviews through Slackbot.
The move reflects a wider push by software suppliers to place business data inside conversational interfaces as companies experiment with AI agents in day-to-day work. HiBob is positioning workforce information as a missing layer in those systems, arguing that organisational structure, reporting lines, tenure and team changes shape business decisions in ways general business data alone does not capture.
Slack is central to that argument because many companies already use it as a front end for internal communications. By connecting Bob to Slack through the Model Context Protocol, HiBob aims to make employee records and HR processes available where managers and staff already work and ask questions.
Examples include checking an employee’s compa-ratio, retrieving a performance review and counting how many people joined a team in a quarter. The integration can also identify who is due for a compensation review in the following month.
Workforce context
The launch comes as software groups and corporate buyers focus on whether AI systems can do more than retrieve information from separate applications. A growing part of that debate concerns context: whether an AI tool can understand enough about an organisation and its people to support decisions, recommendations and workflows in a useful way.
HiBob argues that workforce intelligence should sit alongside financial, sales and operational information when those systems are asked to assist with work. In practice, that means linking HR data to conversational tools used by employees and managers rather than restricting it to a standalone HR system.
“The next generation of enterprise software won’t be built around applications. It will be built around AI agents,” said Ronni Zehavi, Chief Executive Officer and Co-Founder of HiBob.
He set out the company’s view of the role employee data should play in those systems.
“Those agents can only make effective decisions if they understand the people behind the work. Workforce intelligence is a critical layer of business context, and until now it has largely been missing from AI-powered workflows. By bringing Bob into Slack, we’re helping organisations ensure that people insights are part of every decision. Modern companies don’t need more disconnected systems. They need better context delivered where decisions are made,” Zehavi said.
Broader shift
The integration also highlights the growing use of the Model Context Protocol to connect AI assistants with business software. The protocol has gained traction as companies look for more standard ways to let AI tools access data and trigger actions across applications without building separate integrations for each workflow.
For HR software providers, that creates an opportunity to move beyond back-office administration and into more regular employee interactions. Managers checking team changes, compensation timing or review history from within Slack is a different pattern of use from logging into a dedicated HR platform for occasional administrative tasks.
That could make workforce data more visible across organisations, though it also raises familiar questions about permissions, accuracy and the limits of what should be surfaced through chat interfaces. HiBob said the information available through Slackbot comes from Bob, which it described as a source of trusted people data.
Bob is HiBob’s platform for human capital management, payroll and finance. More than 5,400 organisations use the software worldwide, including eToro, Fred Perry, Huel, team.blue, SmartRecruiters and Save the Children.
The release places HiBob within a broader group of software suppliers trying to connect structured business systems to AI tools that employees can use in plain language. Rather than asking workers to switch between systems, these suppliers are increasingly trying to bring information and actions into a single conversational layer.
HiBob’s argument is that AI systems will not be fully useful to businesses unless they can account for the people behind the work as well as the work itself. The Slack integration is its attempt to put that principle into a tool employees already use throughout the working day.
Business & Technology
Thames Water update amid Oxfordshire town’s burst water pipe
After it was reported that households in the OX29 area were running without water this morning (Tuesday, June 23), Thames Water has said the issue has been fixed.
At around 7.30am, the company said it was aware of an issue impacting customers in the Witney, OX29 area and was working on a solution.
READ MORE: Widespread 60mph speed limit reductions examined in Oxfordshire review
At the time West Oxfordshire District Council said it had been informed of the problem and encouraged people to check in on vulnerable neighbours, friends and relatives possibly impacted.
However, now, it appears the issue has been fixed.
At 2pm, a spokesperson for Thames Water said: “We are pleased to confirm that repairs to the burst water pipe at Poffley End Lane have now been successfully completed, and water supplies are being restored to affected customers.
“As water returns to the network, you may initially experience lower water pressure than normal.
“This is expected while the system recharges and pressure gradually builds back up across the area. Your water supply should continue to improve throughout the afternoon.
READ MORE: Town’s armed forces event CANCELLED amid ‘exceptionally hot’ 37°C weather
“If you live in a block of flats and are still experiencing issues once supplies have returned, you may need to contact your managing agent to arrange for any internal booster pumps to be checked and reset.
“Often when your water returns it may be discoloured at first (rusty, white or milky in colour), this is normal.”
They added: “We are sorry for the disruption caused and appreciate your continued support.”
-
Oxford News4 weeks agoOxfordshire families invited to free day of fun in Bicester
-
Business & Technology4 weeks agoNew ‘high-quality’ mushroom business launched in Oxford
-
Oxford Events4 weeks agoWhat’s on in Oxford and Oxfordshire this June
-
Student Life3 weeks agoTransgender rights protest in central Oxford following updated EHRC guidance
-
Oxford News3 weeks agoJeremy Clarkson hits back with sweary response over BGT backlash
-
Crime & Safety3 weeks agoPhotos as 1979 Pontiac Firebird ‘bursts in flames’ at Tesco
-
UK News4 weeks agoUS strikes Iran missile sites and mine laying vessels as Trump’s promised peace deal remains elusive | US-Israel war on Iran
-
Business & Technology4 weeks agoNHS IT outages disrupt 274,620 patient interactions
