AI Services Framework

Opportunity overview

Yorkshire Water has made significant investments in its cloud data platform and has established an internal data science team. The organisation is now moving into a new phase of strategic maturity, aiming to apply AI in ways that deliver measurable value across the business. AI is recognised at the executive level as a priority capability to support efficiency, resilience, and improved customer service. YW operates in a heavily regulated environment, and any use of AI must align with regulatory requirements around transparency, ethics, security, and operational reliability. There are currently two priority domains for AI (although the scope of the framework will be for any AI/ML services required at Yorkshire Water): 1. Central Control Operations • Yorkshire Water operates a real-time control room that monitors extensive telemetry across clean and wastewater networks. • Operators are overwhelmed by alarms, many of which are false positives due to manually-set thresholds (e.g. "trigger at 80% tank level" without data context). • Decision-making often relies on decades of tacit knowledge, e.g. knowing which alarms to ignore or act on. • There is a strong opportunity to apply AI in: o Alarm triaging and rationalisation o Digital twins for simulation and scenario testing o Predictive maintenance and network optimisation o Performance is heavily regulated; service outages or pollution events result in significant financial penalties - hence the need for trusted, explainable automation. 2. Customer Service Operations • Yorkshire Water serves around 5 million customers, handling thousands of customer queries and calls, often with high manual effort. • Call centre volumes spike due to changes like pricing or hosepipe bans, and many interactions (e.g. moving house) are highly repetitive and ripe for automation. • Digital self-service is limited; the website is dated and lacks tools like live chat. • There is strong interest in deploying voice bots, LLM-powered agents, and NLP models to: o Automate simple tasks o Improve customer response time o Prioritise vulnerable customers (while navigating GDPR/sentiment rules) To accelerate delivery and unlock value across the organisation, YW is establishing a framework of specialist AI service providers. The focus is on working with partners who can bring genuine AI expertise and work collaboratively with internal teams to deliver real, operationally grounded solutions. YW is seeking to develop trusted partnerships with AI specialists who can support the delivery of practical, business-focused solutions. There is a strong emphasis on collaboration, with suppliers expected to work closely with internal teams and co-develop solutions in an agile, transparent way. The organisation aims to upskill internal staff as part of any delivery, ensuring that solutions can be owned, maintained, and adapted in-house over time. Suppliers must be able to work entirely within Yorkshire Water's environment, including existing cloud infrastructure (e.g. Azure), CRM (Dynamics), and other systems. No data is permitted to leave YW's infrastructure. The framework will seek suppliers with a range of AI-related expertise, including: • AI consultancy and use case advisory • Design, integration and implementation of AI tools and systems • Data science and platform engineering • Responsible AI and compliance support • Upskilling and capability-building for internal teams Suppliers must be able to work within Yorkshire Water's existing infrastructure and data environment, and deliver in a collaborative, agile manner.

Awarded to

Amounts belong to published awards or contracts. Where several suppliers share one value, it is labelled as shared and is not a supplier-specific allocation.

Classification

• CPV72200000 — Software programming and consultancy services
• CPV72300000 — Data services

Lots

• Lot 1£32m

Documents and notices

• Contract award notice on Find a TenderOpen source ↗
• Tender notice on Find a TenderOpen source ↗
• Preliminary market engagement notice on Find a TenderOpen source ↗