From Whitehall to WhatsApp: The rise of conversational government
Chatbots and secure messaging are reshaping citizen access to public services, offering speed and simplicity, but raising questions of trust and inclusion
Interacting with government often meant queuing in town halls, navigating dense websites or phoning busy call centres. Today, a different future is emerging – one where citizens message their council about bin collections on WhatsApp or ask an artificial intelligence (AI) chatbot on Gov.uk how to register a small business.
This shift from Whitehall to WhatsApp is more than a technological upgrade. It reflects a profound change in how the state communicates with the public.
A UK government survey on public attitudes to AI found that 60% of people had used a chatbot in the past three months, with 44% using them at least monthly.
More than half already believe AI plays a role in public services, and expectations are moving fast. Citizens increasingly want their interactions with government to mirror the immediacy and simplicity of the private sector.
Sultan Mahmud, BT’s director for health and communities, explained that “navigating government websites can be overwhelming, especially when faced with jargon, endless menus and outdated links”.
“Conversational interfaces change that,” he said. “Instead of hunting through over 700,000 pages, users can simply ask, ‘How do I renew my parking permit?’ and be guided step-by-step to the right form or service.”
The appeal is obvious: chatbots cut through bureaucracy, offer 24/7 availability and make interactions feel less intimidating. Yet the challenge is ensuring that the simplicity of a chat window doesn’t mask deeper risks around trust, accuracy and inclusion.
Designing for trust in a digital democracy
The promise of conversational government rests on trust – both in the technology itself and in the institutions deploying it. Secure messaging platforms like WhatsApp are familiar and encrypted, but as Will Richmond-Coggan, partner at Freeths, emphasised, “a string of recent high-profile, and potentially harmful, data breaches across various government departments suggest that there is still a long way to go before platforms like this can be confidently entrusted with large volumes of citizens’ data”.
Purpose-built apps such as the Gov.uk App, currently in beta, may offer stronger safeguards. It combines generative AI (GenAI) chat with features like a digital wallet for documents, including driving licences.
Daryl Flack, partner at Avella Security, tells Informa TechTarget that “trusted government apps of this nature will likely be required to meet strict security requirements such as local sovereign data residency, government-owned encryption keys, robust identity verification and compliance with privacy laws such as GDPR [the General Data Protection Regulation]”.
Trust is not just a technical question, but a cultural one. Nina Midgley De-Jong, senior director for public sector at BIP UK, explained that “security can be designed; trust must be earned”.
“Change management cannot be an afterthought,” she said. “Pockets of resistance can be reduced when people feel heard, their ethical concerns acknowledged, and their voices included in shaping the journey.”
This distinction – between designing secure systems and earning public confidence – will determine whether citizens embrace conversational government as an empowering tool or reject it as another layer of bureaucracy.
Balancing speed, accuracy, and empathy
The early pilots of Gov.UK Chat highlight both promise and peril. Nearly 70% of users found it helpful, with around 65% reporting satisfaction. But experts remain wary of hallucinations – AI-generated errors that sound plausible but are dangerously wrong.
“An inaccurate output from a public sector chatbot has the potential to substantially alter a person’s quality of life,” said Anshuman Singh, CEO of HGS UK. “This poses an important question: how does the sector confidently mix the brain (accuracy of previous systems) with the brawn (speed of generative AI)?”
His solution: integrating GenAI only with vetted internal sources to ensure trustworthy outputs. The question of empathy is equally complex. Most citizens rank speed and accuracy above emotional connection when using chatbots – a GWI survey found only 5% considered empathy an important feature.
Read more about government and technology
- The UK AI Opportunities Action Plan reads like a technology leader’s wish list. But public sector AI needs to be designed with human oversight.
- Amid an economic crisis, public scepticism and confusing messages, could the government’s digital identity programme fail before it even gets off the ground?
- The government has launched the first service on its digital wallet, to apply and download a digital version of the HM Armed Forces Veteran Card.
Yet sensitive services, from benefits appeals to healthcare advice, demand more than efficiency. “Whether in retail, healthcare or the public sector, empathy is critical for fostering strong relationships,” said Singh.
“Government agents simply don’t have the time or resources to solve every query. This is where automation has a role to play – but it must ensure the right firepower is invested into the right query.”
Cobus Greyling, chief evangelist at Kore.AI, agrees. “Striking the balance between automation and human support is critical for user trust,” he said. “Automation can speed up routine services, but citizens should default to human help when tasks are sensitive or high-risk.”
In other words, efficiency is not a substitute for empathy. The real opportunity lies in designing seamless “hand-offs”, where AI resolves routine queries and frees human advisers to focus on complex or emotional cases.
Inclusion at the heart of digital transformation
For all the potential of conversational tools, a critical danger remains: exclusion. Nearly half of central government services are still not fully digitised, with manual or paper-based steps that disadvantage those unable to go online. Amnesty International recently warned that the unchecked use of AI risks excluding people with disabilities and other marginalised groups.
José López Murphy of Globant emphasised that “conversational tools can lower barriers for many, but some people still lack connectivity, devices or digital skills”.
“Accessibility should be about optionality,” he said. “Keep multiple paths – digital, phone and in-person, running side by side.”
Local organisations will be pivotal in bridging this gap. Libraries, Citizens Advice centres and charities already help citizens struggling with digital tools. Martin Neale, CEO of ICS.AI, points to Derby City Council as an example where libraries provide direct support to residents using AI tools. “These initiatives ensure that conversational government is not a one-size-fits-all solution, but one that is complemented by practical, community-level support.”
Midgley De-Jong reinforced this point. “Local authorities and community organisations can provide the human connection and trust that ensures no citizen is left behind,” she said. “Together, they make inclusion not just a policy commitment, but a lived reality.”
Inclusion is not simply a design feature. It is a measure of whether conversational government strengthens democracy by extending access or weakens it by leaving the most vulnerable further behind.
Looking ahead: What success will mean
Five years from now, success for conversational government will not be measured by the number of chatbots deployed, but by the lived experiences of citizens. As Juan José López Murphy put it: “A smooth chatbot doesn’t matter if it still takes six months to get a driving test or four months to see a specialist. Success would mean citizens experience faster outcomes, wider inclusion and greater trust in government.”
For Richmond-Coggan, success will only count if it “enhances citizens’ trust in government and improves the efficiency and accuracy of their interactions”.
“Crucially, success cannot be reconciled with even one large-scale failure which results in harm to citizens,” he said.
Midgley De-Jong imagines a system where services are joined up seamlessly. “Instead of multiple separate processes, it notifies me when my passport is due for renewal and asks my permission to update my driving licence at the same time,” she said. “My time is saved, my stress reduced, and taxpayer money reallocated to frontline services.”
Ultimately, the rise of conversational government is not about technology alone. It’s about reimagining the social contract between citizens and the state.
“A ‘conversational’ government has enormous potential, but it must be implemented with care,” said Martin Taylor, co-founder of Content Guru. “If designed well, these services can reduce strain on government systems, free up human workers to look after the most vulnerable citizens and make interactions more intuitive for everyone.”
The future of public service is being rewritten in plain language. The challenge now is ensuring that this new conversational era is trustworthy, inclusive and genuinely human at its core.
