A funding crisis local government cannot “transform” its way out of

The BBC’s reporting on a projected £200m overspend in adult social care across London councils is not an outlier - it is a symptom of a national system under intolerable financial strain. Rising demand, workforce shortages and higher complexity mean adult social care is consuming ever larger shares of local budgets, crowding out other services and forcing “unfeasible savings” year after year. 

For senior leaders and practitioners, the uncomfortable reality is that more of the same will not close this gap. Without new ways of working, councils face a choice between rationing support, further eroding staff wellbeing, or slipping deeper into deficit. AI is often presented as a silver bullet; in truth, it is closer to a new junior colleague: it can take on repetitive work, but it still needs direction, supervision and good governance. Used that way, it can help shift the cost curve without compromising safety or quality.

If a resident uses a generic chatbot or even worse if a resident used a tool on the open internet, to ask whether they should reduce paid care because of cost, or whether a relative is “safe enough” to stay at home, the financial and human risks of bad advice ultimately fall back on the council, NHS and unpaid carers. At system scale, that drives higher downstream costs: crisis presentations, hospital admissions, safeguarding incidents and legal disputes.

By contrast, domain‑specific AI, built and governed for local government, can reduce demand, avoid unnecessary escalation and release staff time in ways that do show up in the budget. That is the space Datnexa operates in.

Hey Geraldine: releasing capacity and avoiding cost in adult social care

Hey Geraldine, developed with Peterborough City Council, is a personalised AI assistant for social care and health practitioners - designed from the outset to behave like a new junior colleague. It never replaces professional judgement; instead it handles the many “quick questions” that currently consume scarce occupational therapy and social work time.

Independent evaluation has already shown material financial impact:

• Around 15 minutes saved per conversation, across roughly 1,200 staff queries in an initial six‑week trial - over 300 hours of practitioner time returned to direct resident support.

• A long‑standing referral backlog cleared, unblocking hospital discharges and reducing the costly knock‑on effects of delayed decisions.

• More than £250,000 in cost avoidance in year one, through time savings, reduced reliance on agency staff and more efficient decision‑making - equating to roughly a 900% return on investment.

For senior leaders trying to close a multi‑million‑pound gap, this matters. Hey Geraldine does not ask you to find new money; it helps you use the workforce you already have more effectively by:

• Reducing duplication and variation in advice across teams and localities.

• Ensuring expensive professional time is spent on complex thinking and relationship‑based work, not on searching for information.

And because its knowledge base is curated with experienced practitioners and owned by the council, you retain control over the quality, tone and financial implications of the guidance it provides.

Social prescribing chatbot: shifting demand earlier, at lower cost

Evidence is growing that effective social prescribing reduces avoidable GP appointments, hospital attendances and other pressures on the wider system - provided people can actually find and use the offers available. Yet directories are often incomplete, out of date or hard to navigate, which means residents and practitioners default back to more expensive statutory options.

Datnexa’s social prescribing chatbot is designed to address this, while still acting like a junior colleague rather than an unregulated advisor:

• It talks to residents, carers and staff about what matters to them and what is getting in the way of living well, then signposts to relevant local activities, groups and digital tools - without attempting to diagnose or offer medical advice.

• A dedicated safeguarding layer monitors for signs of risk and interrupts to route people to urgent human help, instead of trying to “handle” crisis situations itself.

• The chatbot actively helps maintain the directory, highlighting duplicates, inconsistencies and gaps so teams can target their limited time where it makes the most difference.

For leaders, the financial impacts sit in:

• Reducing avoidable presentations to GPs and A&E by making lower‑cost community and voluntary offers easier to access.

• Maximising the value of investments you have already made in prevention and community capacity.

• Minimising waste in maintaining service information by letting AI do the heavy lifting on data cleaning and gap‑spotting.

Once again, the AI is doing what a good junior colleague would do: keeping lists up to date, having first‑line conversations, spotting when something looks risky - and then bringing humans in.

An AI “front door”: one conversation, fewer wrong turns

The BBC’s reporting highlights how much of councils’ financial pressure stems from late, complex needs and from people bouncing between services before reaching the right help. Every wrong door - every mis‑routed contact, repeated assessment or delayed decision - carries a cost.

An AI‑enabled front door can reduce those costs without replacing human contact. OSCAR, the model Datnexa is developing with partners involves:

• A single conversational entry point where residents and staff explain what is happening, in their own words.

• An orchestration layer that routes parts of the conversation to specialist agents (for example, social prescribing, adult social care, children’s services, housing) and then synthesises the responses into a clear answer.

• Consistent experiences across channels - websites, GP practices, contact centres - so people do not get materially different journeys depending on where they happen to start.

Financially, this can contribute to:

• Lower contact‑centre volumes and handling times by resolving simple queries and signposting accurately at first contact.

• Fewer inappropriate referrals into expensive, high‑threshold services.

• Better system‑wide data on demand, allowing targeted investment in the offers that actually deflect crises.

Crucially, the AI front door is not a “digital gatekeeper” making eligibility decisions in the dark. It is a new way to ask first questions, check resources, draft options and present them to residents and practitioners in a structured, auditable way.

Assessment and report‑writing tools: cutting paperwork, not corners

Assessment and review paperwork is one of the clearest drivers of cost and burnout in adult social care. Agency spend often follows when councils struggle to retain staff under that load.

With tools like AWS Quick, what once required specialist AI tools can be done with custom in-house workflows. Datnexa’s assessment and report‑writing tools are designed to reduce that burden while keeping practitioners firmly in charge:

• AI helps structure and draft reports in line with local templates and statutory requirements, based on information the practitioner has gathered.

• Practitioners review, edit and approve every output, retaining responsibility for analysis, recommendations and decision‑making.

• Over time, consistent structures support better auditing, benchmarking and learning - further reducing the risk (and cost) of poor‑quality documentation.

In financial terms, the gains come from releasing practitioner capacity, stabilising teams, and reducing reliance on agency workers to plug paperwork gaps - all of which are central to the overspends highlighted in the BBC article.

Three tests for AI that can genuinely help your budget

Taken together, Datnexa’s tools - Hey Geraldine, the social prescribing chatbot, the AI front door, and assessment and report-writing tools - offer councils a practical way to respond to the financial crisis in adult social care without sacrificing safety.

When you evaluate any AI proposal against that challenge, three tests can help:

1. Does it behave like a colleague - or an unregulated advisor?It should support practitioner thinking, not replace it; signpost, not diagnose; and escalate risk, not handle it alone.

2. Can you trace the financial impact?Look for clear pathways from the tool to reduced contact volumes, faster throughput, lower agency spend or avoided escalation - not just “innovation” for its own sake. And by increasing faster throughput the resident experience and satisfaction levels increase adding further value. 

3. Is governance built‑in, not bolted‑on?Demand clarity on knowledge sources, safeguarding mechanisms, audit trails and oversight roles before deployment, not afterwards.

Doing nothing on social care is, as MPs have pointed out, financially untenable. The question for senior leaders and practitioners is not whether to use AI, but whether to use the right kind, in the right way. Datnexa’s commitment is to help councils do exactly that: deploying AI that behaves like a junior colleague - diligent, accountable and aligned to local practice - so that your scarce resources can stretch further without increasing risk.