“Roughly five hours is what we probably would spend each week per person. And then in terms of a realistic estimate, let's say we look at 10% efficiency gains. I think in terms of the impact on the annual budget, we probably are looking at something like in the north of £5,000,000 per year.”
Presented in partnership with
UK AI Readiness · G2 AI Custom Research · September 2026
The Connected Mandate: UK leaders want one connected system.
UK business leaders describe a daily working reality built on disconnected software: workflows that hop between applications, data moved by hand, and hours lost in the seams between systems. The cost of that fragmentation lands three ways at once: in money, in people and in control. AI raises the stakes, because organisations are now layering it on top of these same fragmented estates while treating integration as a condition of competing at all. This report maps the gap between the connected organisation leaders say they need and the estate they actually run, and what closing it requires.
Analyst: Molly Burke
Analyst: Aurora Martina Granata
A connected organisation means integrated systems and a single, trusted source of truth that product, customer and enterprise information can move through reliably.
77%of UK business leaders operate in moderately to highly fragmented environments where the software they use every day is not integrated.
90%flag security or compliance exposure in fragmented estates, led by data exposure and identity-management risks. Fragmentation is a control problem, not just a productivity one.
61%of leaders who discussed outcomes report real time, productivity or work-quality gains from AI, but only 7% can point to a quantified financial return. Most value is still unmeasured.
73%call becoming a connected organisation extremely important or business-critical, and 58% say integration is essential or critical to competing over the next two to three years.
Executive summary
How AI-ready are UK organisations today?
This report examines how UK business decision-makers experience technology fragmentation, the consequences it creates for operations and control, and the extent to which organisations are prepared to realise AI value. It also explores how leaders conceptualise connectedness and AI sovereignty, with implications for how AI adoption, platform unification, and continued fragmentation influence both risk exposure and competitiveness.
UK business leaders describe a daily operating reality in which routine workflows cross five to nine applications, and 77% work in environments where 60% or less of their daily software is connected. The cost of that fragmentation lands three ways at once: in money, with payroll diluted by busywork, duplicative licences, and entire roles devoted to manually bridging systems; in people, with roughly one in five leaders volunteering unprompted emotional language about drudgery, frustration and morale; and in control, with 90% flagging security or compliance exposure and leaders describing incidents that ran undetected because nobody was sure who owned the seams between systems.
AI raises the stakes on all of it. 68% of organisations now use AI regularly or have embedded it broadly, and 61% of leaders who discussed outcomes report real time, productivity or work-quality gains. But only 7% can point to a quantified financial return, and 45% of active AI users operate without a fully established governance framework, with several leaders deferring governance openly because early gains arrive with or without it.
Leaders agree on the destination: 73% call becoming a connected organisation extremely important or business-critical, and 58% say integration is essential or critical to competing over the next two to three years. What they lack is a full map of the terrain: 52% of those who defined sovereign AI showed limited awareness or misunderstanding of the term, even though three-quarters say sovereignty already factors into their AI investment decisions.
Durable AI value and future competitiveness both depend on connected systems, governed adoption, and a clearer understanding of what those actually require.
Themes shaping the market
Five findings run through this research, from the fragmented estate to the strategy leaders say they need.
- Operations
Fragmented technology impacts productivity and spending
Workflows routinely span five to nine applications, and fragmentation's largest cost is staff time, lost productivity and rework, named by 64% of those who discussed cost.
- Control
Disconnected systems blur security ownership and widen exposure
Every seam between systems is a boundary that somebody has to own, and often nobody clearly does. 90% of leaders flag security or compliance exposure.
- AI value
AI is proving useful, but governance is trailing adoption
61% of leaders who discussed outcomes report productivity gains, only 7% can quantify a financial return, and 45% of active AI users lack a fully established governance framework.
- Strategy
Integration is now a growth and competitiveness question
Leaders describe integrated systems as growth infrastructure, with 58% calling them essential or critical to competing over the next two to three years.
- Readiness
Sovereign AI is a blind spot inside an otherwise clear strategy
Sovereignty already steers investment decisions, but 52% of the leaders who defined sovereign AI showed limited awareness or misunderstanding of the term.
Inside the interviews
Six voices on what fragmentation costs and what AI needs.
Respondents in their own words: what a single workflow crosses, what the hours cost, where the seams open, and what AI still needs.
Voices 1–2 of 6
“It's a very crowded market which we operate in, so I think AI and the insights and operational efficiencies that AI provides is going to be a competitive advantage. That's going to allow us to maintain and grow our market share, and certainly allow us to develop and launch more products.”
“It's too early to tell ROI. We haven't got the right metrics in place for that, but at least from a development standpoint, for software engineers, the ROI is pretty clear. We've seen acceleration of coding by a factor of four.”
“We had an issue whereby we give access to one of these user emails, and the access granted was more encompassing than we had initially thought, which allowed them access to all of our data on the system.”
“To be honest, I'm not sure what sovereign AI would mean. Using AI to me means saving time. It's moving forward, and it's staying up with current technology levels.”
“I would estimate that we use up to 10 products to achieve what should really be undertaken by one.”
Audio excerpts are presented with lightly edited transcripts for clarity. Participant organisations are anonymised.
Chapter 1: The Fragmentation Tax
Fragmented technology impacts productivity and spending
Today's organisation operates within a largely fragmented tech environment. Leaders describe a frustrating status quo: multiple disconnected tools, the crushing cognitive load of frequent switching between apps, and dozens of hours lost each week across their teams hunting for information or doing manual data entry. On top of lost productivity, a lack of integration creates budgetary bloat: companies buy duplicative tech subscriptions, dilute payroll through busy work, and employ people rather than algorithms to serve as the integration layer between systems that don't automatically talk to each other.
How many software applications do UK teams use to complete a single workflow?
One finance leader's month-end reporting runs through Excel, email, an ERP, two separate budgeting and forecasting systems, back into Excel, and finally PowerPoint. An accounting team manually extracts reports from its ERP, searches invoices in its procurement platform and copies the results into Excel by hand; the team is experimenting with connectors, but “there's nothing that works yet.” Even buying software is fragmented: one purchase request moves through a shared document, is re-keyed into a risk system, reviewed offline, then raised in a fourth tool, with “no integration between them. It's all manual data transfer.” A consulting delivery lead summed up the daily reality: “At every single stage of the journey, there is a breakdown where we have to copy and paste.”
Operational systems create the heaviest integration burden: 48% describe workflows spanning ERP, CRM, line-of-business and reporting tools, compared with 31% using collaboration, productivity and work-management stacks.
Most workflows cross five to nine applications
One square per respondent, by the number of applications a single workflow spans.
- 22%Low: 1 to 4 applications
- 59%Moderate: 5 to 9 applications
- 17%High: 10 or more applications
- 2%Other responses
Source: G2 AI Custom Research, prepared for OneAdvanced. Based on 100 respondents. One square equals one respondent (1%).
| Applications per workflow | Respondents |
|---|---|
| Low application count (1 to 4 per workflow) | 22 |
| Moderate application count (5 to 9 per workflow) | 59 |
| High application count (10+ per workflow) | 17 |
| Other responses | 2 |
What do data silos cost a business in time and money?
Fragmented workflows impose their largest operational burden through staff time, lost productivity and rework: 64% of respondents who discussed the cost of fragmentation cited these issues. Teams spend hours each week switching tools, duplicating updates and validating data, turning disconnected systems into delayed work, avoidable rework and capacity that could otherwise support delivery.
The remaining costs extend beyond time, with 31% citing payroll spent on busy work, licensing bloat, or compensatory headcount wherein entire roles are devoted to manual data entry. A small number of leaders (3%) also highlighted delivery, customer and growth opportunities. These reported team-level impacts may understate the impact to each organisation as a whole, as siloed data can constrain revenue targeting and missed deadlines can convert routine coordination failures into project costs.
Where do employees lose the most time each week?
Hours each week, lost across the whole workflow: 34% of leaders say approval, review and stakeholder dependencies cause the greatest day-to-day delays, while 33% lose most time searching for information and switching between applications, and 30% cite manual data entry, reconciliation and rework.
Why are data silos problematic for businesses?
43% identify blocked AI use, data silos and manual work as the biggest non-financial consequence of software fragmentation. Data silos are problematic for AI in particular, which makes data integration a practical precondition for effective AI adoption.
How do data silos affect business decision-making?
35% cite poor visibility and slower decision-making as the leading challenge created by fragmented software and data silos, while a further 22% point to security, compliance and governance risk. Reported team-level impacts may understate the impact to each organisation as a whole: siloed data can constrain revenue targeting, and missed deadlines convert routine coordination failures into project costs.
Where the cost of fragmentation lands
All 100 respondents, by the cost of fragmentation they named first.
- 64%Staff time, productivity and rework
- 31%Payroll, licensing and extra headcount
- 3%Delivery, customer and growth opportunities
- 2%Other responses
Source: G2 AI Custom Research, prepared for OneAdvanced. Based on 100 respondents. Each percentage point is one respondent.
| Cost named | Respondents |
|---|---|
| Staff-time, productivity and rework costs | 64 |
| Payroll, licensing and additional-headcount costs | 31 |
| Delivery, customer and growth opportunity costs | 3 |
| Other responses | 2 |
Voice of the respondent
“We've got our procurement software, then we've got our ERP, and then we've got our Excel offline workbooks that are used for general preparation. In all three you have to manually extract the reports. You have to manually search for invoices that you need detail for, and then manually copy into Excel.”
Voice of the respondent
“I would say headcount, we probably could reduce by up to 20% if the systems were more integrated. Budget as well from a tech standpoint, we could also probably save another 20% because we wouldn't have to update and maintain the licensing or pay for development, especially on the in house applications that we have.”
The emotional cost of fragmentation
22 of 100 leaders, unprompted, used emotional language to describe fragmented work
Nobody in this study was asked how fragmentation feels, yet 22 of 100 leaders volunteered emotional language anyway. Every phrase below is verbatim, with light transcription cleanup only.
Cognitive load. Frustration.
Platform Engineering, FinTechthe feeling of drudgery rather than a good job done well
Group CFO, Constructionit causes embarrassment in front of the client
Head of Business Solutions, Business ServicesEverything is done manually. It's very frustrating.
Compliance Manager, Consumer CreditTime for manual entry. Annoyance.
Head of IT, Manufacturingbasically drives me crazy
Technical Project Manager, Healthcarenot great for employee engagement or workforce morale
Senior Leadership Team, Logisticsfrustration in that you can't get a job done
UK Operations, Beauty and Personal Carethankless tasks that could be automated
Customer Care Manager, Financial Sectorcreates frustration and dissatisfaction
Head of Marketing and Technology, Financial Servicesdifferent versions of the same document… very annoying
Risk Consulting, Financepeople have to work overtime to actually meet the deadlines
AI Solutioning, Consultingthe frustration factor with employees of thinking that things should be easier
Senior FP&A Manager, DefenseI always worry that we're gonna do something kind of embarrassing
Senior Lecturer, Higher Education
And when the admin burden lifts: they could basically get their life back, reduce burnout of staff
Technical Project Manager, Healthcare, on clinicians after AI removed administrative typing
22 of 100 interviews included emotional terms describing fragmented systems, manual work, switching or time lost; a further group expressed emotion about AI adoption itself, counted separately. Selected phrases shown.
What this means
Map fragmentation at the workflow level, not the application level. An application inventory understates the problem: the real cost shows up inside a single end-to-end workflow, where five to nine tools and repeated hand-offs create the time loss and rework 64% of leaders describe. Start by tracing one high-volume workflow start to finish before scoping any wider integration effort.
OneAdvanced perspective
Marko PerisicChief Product Officer, OneAdvanced
“When teams spend their day moving between disconnected systems, productivity suffers and valuable insight gets trapped in silos. The organisations creating competitive advantage today are those simplifying complexity, connecting workflows and freeing people to focus on high-value work rather than administrative overhead.”
Chapter 2: The Security Seams
Disconnected systems blur security ownership and widen exposure
A third but no less important cost of tech fragmentation is control. Almost all (90%) leaders we interviewed said the fragmented nature of their software environment increases their security or compliance exposure. In a connected environment, security is one perimeter, one identity model, and one set of controls. In a fragmented one, every seam between systems is a boundary that somebody has to own, and often nobody clearly does.
Where do the biggest security risks in a fragmented estate sit?
Having a large number of disconnected tools often results in misunderstandings about who is responsible for keeping them secure. The leading fragmentation-induced cybersecurity vulnerabilities leaders cite are clustered around data exposure, access-control and identity-management risks (47%).
One financial services delivery head recalled: “We have had situations where an SMS pumping attack was carried out on the organisation, which went undetected because no one was managing or reviewing the cost or logs of that application, because it wasn't clear who was responsible.” A procurement leader running multiple regional ERPs described what a single compromised account could do across disconnected systems with poorly segregated roles: “change suppliers' bank details, approve payments, extract sensitive information.”
Identity is the leading exposure in fragmented estates
- 47% Data exposure, access-control and identity-management risks
- 39% Expanded attack surface, inconsistent controls and weak monitoring
- 14% Shadow IT, shadow AI and uncontrolled data sharing
Source: G2 AI Custom Research, prepared for OneAdvanced. Based on 90 respondents who discussed security risk.
| Risk named first | Share |
|---|---|
| Data exposure, access-control and identity-management risks | 47% |
| Expanded attack surface, inconsistent controls and weak monitoring | 39% |
| Shadow IT, shadow AI and uncontrolled data sharing | 14% |
Identity is the leading exposure
47% of leaders who discussed security risk name data exposure, access-control and identity-management risks first, describing overly broad access grants and compromised credentials reaching systems outside central control.
Fragmentation defeats a single security posture
39% point to expanded attack surfaces, inconsistent controls and weak monitoring, with unclear ownership leaving parts of the estate unwatched and unpatched. A manufacturing IT head put the life cycle problem directly: with fragmentation, “is the system siloed, forgotten about, not updated, not patched?” Systems excluded from single sign-on run their own identity management, and credentials get reused across boundaries.
When something goes wrong, fragmentation slows discovery
One operations leader described fraud that “was clear on one system” but spread because catching it required “manual identification by the responsible parties” rather than automatic flagging across tools. A healthcare project lead described patient-data opt-outs that silently failed, caught only because someone manually checked a dataset before delivery. A finance team's response to a stolen, unlocked phone slowed because the incident had to pass by hand from IT to finance across disconnected processes.
What are the cybersecurity risks of sharing company data with AI tools?
14% name tools adopted outside formal channels that spread company data with no consistent oversight. Organisations are connecting AI tools to these same fragmented estates, and the seams are already leaking: one data science leader reports staff putting “sensitive personal data into the AI” tools, something “noted more than once,” while another describes the core difficulty as “having a single consistent security plane across all of the applications.”
Voice of the respondent
“Our systems that aren't accessed through single sign-on rely on their own onboard identity management, and usually they are accessed via the user's email. So we have had a number of incidents where third parties have gained access to a colleague's email account and used it to then gain access to those unconnected systems.”
Voice of the respondent
“It has happened previously that certain teams don't have access to information in certain software. So someone from another department will lend them credentials, which is obviously a security breach.”
Voice of the respondent
“Probably difficulty in having a single consistent security plane across all of the applications will be one. Although all work pretty good in terms of security, I think having different systems with different governance approaches makes it difficult to have a one size fits all or a highly standardised cyber approach.”
What this means
Treat fragmentation as a security issue, not only an efficiency one. 90% of leaders tie disconnected systems to security or compliance exposure, led by data-exposure and identity-management risk. Any business case for integration should name the access-control and monitoring gaps a fragmented estate creates, not just the productivity cost.
OneAdvanced perspective
Andrew HendersonChief Technology Officer, OneAdvanced
“Every disconnected system creates another point of risk. Security today isn't just about protecting individual applications, it's about maintaining visibility, governance and accountability across the entire technology estate. The more connected your systems, the stronger your ability to manage risk with confidence.”
Chapter 3: AI Value and Governance
AI is proving useful, but governance is trailing adoption
AI adoption in UK organisations is widespread, with over two-thirds (68%) of the leaders in this study reporting that their teams use AI regularly in selected functions or that AI is now embedded broadly across their organisations. But reported ROI is lopsided in favour of productivity: leaders can point to faster, better-quality work and hours returned to their teams, yet very few can put a financial number on what AI is delivering to their companies. AI governance is similarly running behind the curve: among organisations actively using AI, 45% operate without a fully established governance framework.
Organisations without frameworks report productivity gains at roughly the same rate as those with mature governance. Some leaders defer governance deliberately, fearing that rigid rules will slow experimentation. Others are building the enforcement habit now, reviewing every tool before it touches company data, and report no slowdown for doing so. In the long term, however, standing up AI governance now could lend a competitive edge: in this study, organisations that manage AI security through approved tools and formal review tend to keep adopting, while organisations with security concerns but no controls are the ones whose AI use stalls.
Read together: for every organisation that can put a financial figure on AI, nearly nine more are seeing gains they cannot yet price, and another four see little or nothing to report. That ratio is the truest measure of how AI-ready UK organisations are today.
Why can so few organisations quantify AI ROI?
AI is delivering operational value faster than it is proving financial impact. Among interviewed leaders who discussed AI ROI, 61% cited time savings, productivity or work-quality gains, while only 7% reported a quantified cost, revenue or profitability effect. A further 30% described their return as limited, absent or unmeasured. Reading those numbers together: for every organisation that can put a financial figure on AI, nearly nine more are seeing gains they cannot yet price, and another four see little or nothing to report.
This could be because most AI deployment is recent and partial, and the measurement infrastructure was never built: one technology leader reports developing code four times faster while conceding, “It's too early to tell ROI. We haven't got the right metrics in place for that.” The handful of organisations that could cite the monetary returns of using AI extrapolated ROI from priced units of high-volume work such as service desk ticket processing.
In leaders' own telling, fragmentation itself helps account for the difficulty in reporting ROI. Many (43%) name blocked AI use and data silos as among the biggest non-financial consequences of fragmented software, and 21% name fragmented systems and poor data quality as the main constraint on AI value. An AI system is only as good as the data reaching it, and leaders consistently describe disconnected systems as the reason their data cannot be trusted, assembled, or measured end to end.
AI returns hours far more often than pounds
Respondents who discussed AI outcomes, by the return they described. The top response leads; the rest are ranked beneath it.
61%Top response: time savings, productivity and work-quality gains, named by 61 of 99 respondents.
- Time savings, productivity and work-quality gains61%
- No, limited or unmeasured ROI to date30%
- Quantified cost, revenue or profitability impact7%
- Other response1%
Source: G2 AI Custom Research, prepared for OneAdvanced. Based on 99 respondents who discussed AI adoption, value and barriers. Percentages are shares of the 99 respondents; the one response naming security, governance, cost and unclear ROI constraints is shown as other.
| Return described | Respondents |
|---|---|
| Time savings, productivity and work-quality gains | 61 |
| No, limited or unmeasured ROI to date | 30 |
| Quantified cost, revenue or profitability impact | 7 |
| Other response | 1 |
What is holding back AI readiness at large UK organisations?
43% of leaders say skills, confidence, adoption and organisational change gaps are what is holding back AI readiness and preventing them from extracting greater value from AI. When asked about the single biggest barrier to becoming an “AI-enabled organisation,” 51% of leaders cited similar criteria.
How many organisations have an AI governance framework?
Among the organisations that have actively deployed AI, 55% do so within a governance framework that establishes rules for how AI is adopted, used and controlled; 31% operate with a fledgling framework and 14% with none at all. Across all 97 respondents who discussed governance, 44% have an established framework that is actively evolving or enforced, 33% have one in development or partially implemented, and 23% have none. Among organisations not yet using AI, only 6 of 31 have an established framework in place.
There is no near-term penalty for skipping governance
Organisations with no governance report productivity gains at rates comparable to those with mature governance, so the business case for formalising feels abstract. One transformation manager at an organisation using AI for a year and a half: “We don't currently have a governance framework for AI, but I think we will in the future. I actually don't know if one's being worked on.” Others worry that rules will choke momentum, describing the challenge of governing AI “without creating a massive backlog and a massive break on your AI innovation efforts.”
Controls enable adoption rather than slowing it
51% of leaders say their concern about the cybersecurity risks of sharing company data with AI tools is managed through approved tools and formal review, while 43% say high concern is materially restricting or delaying their AI use. One leader with mature controls reports, “We are concerned, but we have proper guardrails and governance in place. It hasn't slowed or stopped any programme.”
Governance and quantified ROI travel together
The small group of organisations reporting quantified AI returns nearly all have established governance frameworks in place, whereas no organisation without a framework could do so. The organisations reporting quantified returns defined baselines and cost-per-unit before scaling, a practice most organisations describing “unmeasured” value have not yet adopted. The pattern is directional among small groups, but it points the same way.
Fewer than half have reached established AI governance
All organisations that discussed governance, placed along the maturity pathway from no framework to an established one.
No framework23%No established AI governance framework
In development33%Framework in development or partially implemented
Established44%Established framework that is actively evolving or enforced
Source: G2 AI Custom Research, prepared for OneAdvanced. Based on 97 respondents who discussed AI governance, including organisations not yet using AI.
| Governance status | Share |
|---|---|
| Established framework that is actively evolving or enforced | 44% |
| Framework in development or partially implemented | 33% |
| No established AI governance framework | 23% |
Voice of the respondent
“We see on average a thirty-minute saving per case, with all cases investigated fully at the first level when brought in by AI. Thirty minutes per case, at around two to three thousand tickets per month, at a blended rate, our return on investment is currently being tracked at just over 300k annually.”
Voice of the respondent
“The AI governance framework is owned by IT and the teams, as they are the ones who would define the usage guidelines and also review it year on year based on the policy changes that need to happen with the evolving AI in the market. It does get enforced regularly, and there are regular trainings and regular participation required.”
Voice of the respondent
“We don't have a very formal established governance framework, but we do have a draft framework in place, which is being continuously revised. It's been in place for roughly the last one to two years, although it is still evolving as our AI usage becomes more mature.”
Voice of the respondent
“It doesn't have a governance framework. It has a compliance framework for evaluating AI tools, but not a governance framework for organisation-wide AI usage, though certain departments do.”
What this means
Put a measurement plan in place before scaling AI further, and put governance alongside deployment rather than after it. Before expanding AI use, organisations should define what a measured outcome looks like for their highest-volume use cases, the way the support function in this study tracked time-per-case savings.
OneAdvanced perspective
Amanda GrantChief of Strategic Ventures, Data & AI, OneAdvanced
“Many organisations are already seeing meaningful productivity gains from AI, but long-term value depends on trust, governance and quality data. The businesses that will realise the greatest return are those treating AI as a strategic capability, supported by clear controls and connected information foundations.”
Chapter 4: The Connected Mandate
Leaders treat integration as critical to growth but are underinformed on what sovereignty and governance require
Leaders see the urgency of reducing fragmentation. Nearly three-quarters (73%) call becoming a connected organisation extremely important or business-critical, and 58% of those who assessed the importance of connection over the next two to three years say integrated systems will be essential or critical to their ability to compete, grow market share, and launch products. Integration has moved out of the back office in leaders' minds: they describe it as growth infrastructure, with regulatory expectations expected to raise the bar further.
What does a truly connected system look like?
One trusted source of truth. Asked to define a connected organisation in their own words, 73% describe integrated systems and a single, trusted source of truth that product, customer and enterprise information can move through reliably. A smaller group (14%) defines connectedness through automated end-to-end workflows that enable AI and faster decisions, and 10% emphasise governed, secure and role-appropriate access. The ordering is telling: most organisations are still trying to make data move reliably at all, a prerequisite they have not yet solved, before they can credibly aim for AI-enabled automation on top of it.
Connectedness is a business-critical priority for most leaders
Respondents who rated the importance of becoming a connected organisation.
73%
19%
7%
1%
Source: G2 AI Custom Research, prepared for OneAdvanced. Based on 100 respondents. Each percentage point is one respondent.
| Rating | Respondents |
|---|---|
| Extremely important or business-critical | 73 |
| Very important strategic priority | 19 |
| Moderate or conditional priority | 7 |
| Other response | 1 |
Trusted integration is the dominant definition
73% describe a connected organisation as integrated systems with a trusted single source of truth, far ahead of AI-enabled automated workflows at 14%. Only 10% define connectedness through secure, role-appropriate data access, which foreshadows the gap the final chapter examines.
There are revenue stakes behind connection
A retail architecture head traced the line from integration to revenue in one step: if product data from the product-information system is not available to the website, “that product can't be sold.” Even the 7% who describe connectedness as a moderate or conditional priority acknowledge its criticality: “it will become more and more important as new software packages get released all the time.”
Why does enterprise data integration matter?
58% of leaders who assessed the next two to three years say integrated systems will be essential or critical to their ability to compete, and a further 34% call them a very important strategic enabler. The language is the language of growth: holding market position in crowded markets, launching more products, expanding what the organisation can offer.
Two pressures sharpen the deadline
The first is that integration is upstream of AI, and leaders know it: several describe integrated systems as what makes AI-enabled insight and faster decisions possible at all. The second is regulatory. A digital strategy lead in the utility sector expects external demands to force the issue: “the expectations from regulators on us are becoming much more demanding.”
Integration is essential to competing over the next two to three years
Respondents who assessed the importance of integrated systems to competitiveness and growth over the next two to three years.
Source: G2 AI Custom Research, prepared for OneAdvanced. Based on 97 respondents who assessed integration strategy and future urgency. Percentages are shares of the 97 respondents (56, 33, 7 and 1 respondents respectively), rounded.
| Rating | Respondents |
|---|---|
| Essential or critical for competitiveness and growth | 56 |
| Very important strategic enabler | 33 |
| Moderately important or conditional priority | 7 |
| Other response | 1 |
Voice of the respondent
“I think it will become more and more important as new software packages get released all the time. But at the moment, the priority is controlling costs.”
Voice of the respondent
“As a retailer, we sell effectively on the data of our products. So if our data from our product information management system isn't available to our website, then that product can't be sold. It's as crucial as, if the systems aren't connected, we don't sell anything.”
Voice of the respondent
“It will become increasingly important. As I mentioned previously, the lack of integration slows down decisions, and the expectations from regulators on us are becoming much more demanding. So integration will be much more important in the next two to three years than it is right now.”
Voice of the respondent
“I think integrated systems will continue to be essential, even before you go AI. In the last three, four, five years, the company had started to try and integrate these systems. So we still have a long way to go, but we've already started that journey, and it's essential for us to compete, grow and also develop more services for customers.”
What this means
Frame integration as a growth and competitiveness agenda. Leaders already describe integrated systems as growth infrastructure, so build the business case around the capabilities they want to improve: faster decisions, expanded offerings and stronger operational performance, with connected people, data and systems as the foundation.
OneAdvanced perspective
Xavier LangloisGeneral Counsel, OneAdvanced
“Connected operations have become a business imperative. The organisations that thrive over the next decade will be those that combine integrated technology, trusted data and responsible AI governance to make faster decisions, adapt with confidence and unlock sustainable growth.”
Chapter 5: The Sovereignty Blind Spot
Sovereign AI is a blind spot inside an otherwise clear strategy
But there is a gap between conviction and fully understanding the bigger picture around fragmentation, AI, and governance. More than half of the leaders who tried to define sovereign AI in their own words showed limited awareness or misunderstanding of the term, even though three-quarters say control over where data resides and which AI models can touch it already factors into their investment decisions.
What is sovereign AI, and why does the misunderstanding matter?
Sovereign AI is the ability of an organisation or nation to develop, run and govern AI using infrastructure, data, models and oversight that remain under its own jurisdiction and control. No outside party, whether a foreign government, an external vendor or another company's cloud, can access, constrain or exert legal authority over the AI system or the data flowing through it. At the organisational level it means controlling where data resides, which models can touch that data, who operates the infrastructure and under what framework.
It can be achieved through self-hosting, private cloud deployments in a specified region, or contractual controls such as zero data retention with model providers. Crucially, the concept goes beyond hosting: a model running on-premises still is not sovereign in the full sense if its training pipeline, updates or telemetry route through an external party, or if governance and accountability for its use are undefined. Leaders equate sovereignty with where AI runs and who operates it, so an organisation can conclude that on-premises hosting has made it sovereign while its residency obligations and governance accountability remain unresolved.
Put simply, sovereign AI describes a system of deploying AI within a fully realised governance framework, with security infrastructure designed to prevent the leaks and doubt leaders currently describe. A misunderstanding as to what sovereign AI means may predict blind spots around how to move away from fragmentation towards a more secure, connected, automated tech stack.
Half of leaders cannot define sovereign AI
Respondents who attempted a definition of sovereign AI, by how they framed it.
Source: G2 AI Custom Research, prepared for OneAdvanced. Based on 97 respondents who discussed sovereign AI; 94 offered a definition. Values are respondent counts; three responses that addressed vendor credentials or investment criteria rather than a definition are shown as other.
| How sovereign AI was framed | Respondents |
|---|---|
| Limited awareness or misunderstanding of sovereign AI | 49 |
| Organisation-controlled, private or self-hosted AI | 33 |
| National or regional data residency and jurisdictional control | 12 |
| Other responses | 3 |
Sovereign AI lacks a shared meaning
52% of leaders who defined it showed limited awareness or misunderstanding, and fewer than one in eight connected it to data residency and jurisdictional control. The pattern is a conflation with consequences: leaders equate sovereignty with where AI runs rather than the fuller combination of infrastructure, data, models and governance under a jurisdiction.
How much does AI sovereignty factor into UK enterprise AI investment?
The gap matters because sovereignty is already steering money. 76% of leaders say control over where data resides and which AI models can access it factors into AI investment decisions, including 35% for whom it is deciding or mandatory. And when evaluating any technology vendor, 90% say cybersecurity credentials are a deciding factor or threshold requirement. Buying pressure, in other words, is fully formed; shared understanding is not.
The blind spot sits inside the pro-integration majority
Among leaders with a limited or mistaken understanding of sovereign AI, nearly nine in ten still rate integrated systems as essential or very important to future competitiveness. These are the same leaders driving the integration agenda: committed to the destination and underinformed about the terrain, which leaves risks such as ungoverned AI use and security exposure across fragmented estates unpriced in the very investment decisions meant to fix them.
Voice of the respondent
“I would say that is either owned by a business or a state, such that the data is held within a territory or within a business unit or organisation and not accessible by those outside either our organisation or that state.”
Voice of the respondent
“Sovereign AI is an AI running within the four walls of something, be that a company or be that a country. It is a guardrailed, highly governed, private AI solution.”
What this means
Build a plain-language definition of sovereign AI before it becomes a procurement blocker. 52% of leaders who defined it showed limited or mistaken understanding, even though 76% say data residency and model access already factor into AI investment decisions. A shared internal definition, tied to actual data-residency and governance requirements, will prevent decisions being made on an incomplete idea of what sovereignty covers. Closing the knowledge gap around AI sovereignty is part of closing the readiness gap around AI adoption and connected systems.
Strategic takeaways
4 moves that turn connection into a business capability
- Critical
Prioritise high-risk cross-system workflows
Map the workflows that require the most hand-offs between applications, starting with those handling sensitive data or distributed access. Use the mapping to focus integration, identity and access-control improvements where fragmentation creates both staff-time costs and security or compliance exposure.
- Critical
Put AI governance alongside AI deployment
Make governance an operational workstream rather than a policy exercise: assign ownership, embed controls into systems and establish monitoring as AI use expands. Build shared literacy on sovereign AI so leaders can make more informed decisions about control, operation and data residency.
- High
Turn early AI gains into measurable business value
Retain productivity and work-quality measures, but add defined financial and operational baselines for priority AI use cases. This will help distinguish promising pilots from repeatable initiatives that can demonstrate cost, revenue or profitability impact.
- High
Frame integration as a growth and competitiveness agenda
Develop an integration roadmap around the business capabilities leaders want to improve, including faster decisions, expanded offerings and operational performance. Position connected people, data and systems as the foundation for growth as well as modernisation.
Frequently asked questions
It is a practical operating condition for many organisations: 59% of respondents who discussed application use said routine workflows span five to nine applications. That level of stack breadth means teams must repeatedly move work, data and context across specialised tools to complete end-to-end processes.
Security and compliance exposure was flagged by 90% of respondents. The leading concerns were data exposure, access-control and identity-management risks, cited by 47% of those who discussed security risk, while 39% pointed to expanded attack surfaces, inconsistent controls and weak monitoring. Fragmentation also carries operational costs: 64% of those discussing costs identified staff time, lost productivity and rework.
Yes, but the value is primarily operational and not yet consistently quantified in financial terms. Among respondents discussing outcomes, 61% cited time savings, productivity or work-quality gains; only 7% reported quantified cost, revenue or profitability effects, and 30% described ROI as limited, absent or unmeasured.
Readiness is uneven, and governance is trailing adoption. 68% of organisations actively use AI, yet 45% of those active users lack a fully established governance framework. In addition, 52% of those who defined sovereign AI showed limited awareness or misunderstanding, signalling an important knowledge gap around secure control and operational ownership.
Connectedness is viewed as extremely important or business-critical by 73% of organisations, and 58% of those who assessed the next two to three years say integration is essential or critical for competitiveness and growth. Leaders link integration with AI-enabled insight, faster decisions, stronger operational performance and the ability to expand offerings.
Connect it, govern it, understand it
The central transformation this research points to is the shift from treating integration as a technology improvement to treating connected operations as a business capability. The cost of standing still is already visible: workflows spanning five to nine applications drain staff time and dilute payroll, fragmented estates expose organisations to security risk that 90% of leaders recognise, and roughly one in five leaders volunteered, unprompted, that fragmented work is wearing down their people.
AI sharpens the timeline. Most organisations are already active users, and the productivity gains are real, but only 7% can quantify a financial return and 45% of active users are deploying AI without a fully established governance framework. The early evidence favours the disciplined: organisations that manage AI security concern through approved tools and formal review keep adopting while others stall, and the few quantified returns in this study sit almost entirely inside governed organisations.
Leaders already agree on a connected future: 73% view connectedness as extremely important or business-critical, and 58% say integration is essential or critical to competitiveness over the next two to three years. What remains is closing the gap between conviction and understanding, because more than half of the leaders who tried to define sovereign AI, a concept already steering their AI investments, showed limited or mistaken understanding of it. The organisations that connect high-value workflows first, put governance alongside deployment rather than after it, and learn what sovereignty actually requires will turn operational connection into a platform for resilience, faster decisions and growth.
Research methodology
How was this research conducted?
This research draws on 100 in-depth interviews with UK business decision-makers and business professionals, all working at organisations with 250 or more employees or system users. The sample spans a mixed buying group: senior IT, finance, transformation, cybersecurity, procurement, platform, operations, risk, compliance, AI and data & analytics roles, plus C-suite leaders including CTOs, CEOs, CISOs and CFOs. All participants confirmed they own or significantly influence decisions about their organisation's technology systems, integration or AI investments.
Who we interviewed
Interviews were conducted as AI-moderated, conversational voice interviews, following a structured flow covering workflow fragmentation, the operational and financial cost of disconnected systems, cybersecurity and compliance exposure, what a connected organisation means, AI adoption and readiness, AI governance, sovereign AI, and future integration priorities. The open-ended format let respondents describe their organisation's actual practices in their own words rather than select from preset answers.
How we analysed it
Every transcript was processed through AI-enabled thematic analysis, then independently reviewed by G2's AI Custom Research team to verify coding, resolve ambiguous responses, and confirm the narrative accurately reflects what respondents said. Question-level bases vary because percentages reflect the interviews that addressed each topic; bases are stated alongside each chart. No company-size or industry breakdown is reported for individual respondents to preserve anonymity.
Limits
Question-level bases vary because percentages reflect the interviews that addressed each topic; bar charts show respondent counts against the stated base, exactly as in the source analysis. The link between governance and quantified AI returns is directional among small groups. Emotional language was volunteered by respondents, not prompted. No company-size or industry segmentation was possible.
Presented in partnership with
Research prepared for OneAdvanced
This report was prepared for OneAdvanced by G2 AI Custom Research. Visit OneAdvanced to explore its software, products and resources.
Visit OneAdvanced ↗
Presented in partnership with IQ by OneAdvanced
Introducing IQ: The Intelligent System of Work
As organisations across the UK accelerate their adoption of AI, the challenge is no longer simply having access to the technology. The challenge is creating the foundations needed for people, data and AI to work together effectively. IQ brings these elements together in a connected, trusted and intelligent system of work, powering the world of work and helping organisations move faster, make clearer decisions and innovate without unnecessary risk or disruption.
- ConnectedUnifying workflows, teams and data, with business and sector context carried across every interaction.
- TrustedSecure, sovereign and resilient, with enterprise-grade cyber security and sector-aligned compliance.
- IntelligentEmbedding AI-driven insight and automation directly into the flow of work, delivering real-time insight wherever it is needed.
59%of leaders believe their organisation matches or outperforms peers in AI maturity
49%say AI supports less than a quarter of their operations
Source: OneAdvanced IQ research. These statistics are from OneAdvanced’s own research, separate from the G2 study above.
“Month-end would’ve take me a good day to report on previously with my old solutions; now with IQ, it's just one click”
Maxine Rendell, Chief Operating Officer, Dexter Montague
“It will have an instrumental impact in improving the organisation’s ability to deliver critical services when it’s most needed”
Jonathan Stones, CEO, John Barkers Solicitors
“The one thing that excites me is the added value that it will bring to the users experience and our interaction with the information that we already know exists within the systems that we use”
Gillian Day, Group Director of Finance and Corporate Services, JTL Training Ltd
“It will revolutionise the way we use our data and all the business knowledge that we have and help us very quickly deliver the outcomes we’re seeking”
Tony Van Der Hoorn, Chief Financial Officer, Channel Islands Coop
Sponsored content supplied by OneAdvanced. Customer quotes are drawn from OneAdvanced customer stories, not from the anonymised G2 interviews in this report.