Oakhaven Care Ltd

That gap matters, and this roadmap addresses it alongside the operational priorities.

• The rostering problem is the single biggest daily cost and the one most likely to spill into CQC risk. Matching 180 carers to roughly 4,000 weekly visits by geography, continuity and travel time, mostly by hand, is a structural inefficiency rather than a people problem. Until there is a proper care-management system handling scheduling logic, every other operational improvement competes for the same limited management attention.

• Care-note and medication-record duplication is a downstream symptom of the same missing infrastructure. When carers record something in one place and coordinators re-enter it somewhere else, you are not just wasting time, you are creating the conditions for errors that regulators and families will notice.

• The recruitment process is a bottleneck with a direct commercial cost: slow DBS clearance, manual reference chasing and paper-based training tracking mean carers take longer to become billable and some drop out of the process before they start. You are already running paid recruitment advertising on Indeed and Facebook, which is a reasonable starting point, but the advertising is doing work that falls apart at the application stage if the process behind it is slow.

• Key-person dependency is identified as significant, and in domiciliary care that is not an abstract risk. The registered manager personally holds most commissioner relationships. If that person left, the referral pipeline that sustains the business would be at immediate risk, because those relationships are not documented, replicated or embedded in any system that survives a personnel change.

• The business has no marketing plan, no content strategy and no SEO activity, which is a deliberate choice given that the primary referral route is social services and hospital discharge teams rather than private-pay search. That is an honest and sensible read of the current funding model. However, the private-pay market for domiciliary care is growing and the business has no infrastructure to reach it if the commissioning environment shifts. There is no urgency here, but it is worth naming now so the decision to stay absent from that channel is a deliberate one rather than an oversight.

• The three priorities the leadership team named, which are company foundations and team, managing risk and compliance, and measuring performance and reporting, all point in the same direction: the business needs operational systems before it needs marketing campaigns. This roadmap reflects that sequencing. Getting the right management platform in place, fixing the data it produces, and then using Armadello (Anicca's reporting product, which brings channel and commercial performance into one view) to make sense of what the business is actually doing, is the right order.

• The AI questions in the discovery session describe a set of genuinely AI-shaped problems: rota optimisation, care-note summarisation, CQC evidence gathering, recruitment screening, and routine query handling. The team's AI literacy is described as basic and there is no technical resource for AI projects. That combination means the AI goals need a separate, structured adoption process before any implementation. Anicca's AI Adoption Roadmap covers exactly that territory, and it is the appropriate next step to run alongside or immediately after the work in this report. This roadmap does not attempt to deliver AI implementation guidance.

Where time disappears

• Rostering and last-minute rota changes consume your coordinators at the pace the care itself runs, which is continuous. Every gap in cover triggers a chain of phone calls and manual fixes. That is not a process problem you can solve with a spreadsheet tweak; it is a structural constraint that limits how much growth your eight coordinators can absorb before the function breaks.

• Care-note checking, medication administration record (MAR) reconciliation and CQC evidence-gathering are done reactively rather than systematically. The work happens, but the time it takes is disproportionate because there is no single place where that evidence accumulates in a form that is ready to present. When an inspection comes, or a local-authority contract KPI review, your team has to reconstruct what already happened.

• Local-authority invoicing and call-monitoring reconciliation sit at the end of a process that is already manual from start to finish. Errors or mismatches at this stage are expensive to unpick, and with thin margins on hourly rates, disputed hours or delayed payments have a direct effect on cash flow.

• Recruitment screening, including Disclosure and Barring Service checks, references, right-to-work documentation and mandatory training, is your highest-pressure ongoing process, by your own account far more demanding than any marketing concern. You are running this through whatever your current HR system supports, but without a CRM or applicant-tracking tool in place, the status of each candidate at each stage of the process almost certainly lives in someone's head or inbox.

What this means for growth

• Adding a fourth local-authority area is not just a sales problem. It requires another layer of coordinator capacity, another set of contract KPIs to evidence, and a recruitment pipeline that can be stood up in a new geography. If the existing processes are not documented and repeatable, opening a fourth area simply copies the current friction into a new location.

• You win contracts on CQC rating, carer retention and reliability of cover. Those are genuine competitive advantages over the national chains and the local independents bidding on price alone. But they are only advantages if you can demonstrate them consistently, which means your quality-evidence processes need to produce records that are accurate and retrievable, not assembled at speed before a review.

• There is no CRM in use. That matters less on the client side in a B2B2C model where referral relationships with social workers, hospital discharge teams and local-authority commissioners drive most placement decisions. But it matters significantly on the recruitment side, where candidate relationships go cold between application and DBS clearance, and on the family-liaison side, where handling queries by phone with no shared record means the same question gets answered differently by different people.

The practical first step

• Before any marketing or channel investment makes sense, the priority is to map which of these processes, rostering, recruitment screening, care-record reconciliation or invoicing, is creating the most downstream cost and risk right now, then fix that one process systematically before touching the others. Trying to standardise everything at once in a 180-carer operation adds its own disruption.

• Armadello would give your leadership team one view of what is actually happening across the business, combining contract performance, recruitment pipeline activity and operational output rather than pulling reports from separate places. That kind of visibility is what makes a fourth-area expansion manageable rather than speculative.

• There is a live question here about whether AI tooling could help with specific internal tasks, policy drafting, care-note summarising, scheduling support. Your team is already using ChatGPT informally for policy drafting, which suggests appetite without structure. Closing that gap properly, with a clear usage policy and the right tools for the right tasks, is the territory of Anicca's separate AI Adoption Roadmap rather than this report, but it is worth naming because the operational pressure you are describing is exactly where that kind of work pays for itself.

All of that is real and hard-won. The problem is that none of it is captured, measured or actively managed anywhere.

Who you are winning and how

• Your two customer groups, local-authority and NHS commissioners on one side and private-pay families on the other, have almost nothing in common in terms of how they buy, what they need to hear, and how you keep them. Roughly 80% of your revenue comes through council frameworks and tenders, where the relationship is institutional and the decision-maker is a commissioner, not the person receiving care. The remaining 20% comes from private families who choose you personally and can leave you personally. Treating these two groups as one audience, or not treating them as an audience at all, means you are not actively speaking to either.

• Your cost of winning a new contract is largely invisible to you. You told us it is "hard to put a figure on", which is understandable given that acquisition happens through tender processes and word of mouth rather than paid campaigns. But invisible costs are still real costs: the management time spent on tender responses, relationship maintenance with commissioners, and handling inbound family enquiries all have a value. Without knowing roughly what it costs to bring on a new care package, you cannot make a sensible decision about whether to invest more in that activity or redirect effort toward growing hours within your existing base.

• There is no CRM in place. That means contact records, package histories, relationship notes and family communications exist, if they are recorded at all, in people's heads, email inboxes and spreadsheets. When a care coordinator leaves, that institutional knowledge leaves with them. It also means you have no straightforward way to see which clients have been with you longest, which families might be ready to top up local-authority hours with private-pay hours, or which commissioner relationships need attention ahead of a tender renewal.

The private-pay opportunity you are not acting on

• You identified this yourself: families whose relative is funded by the local authority often reach out when needs increase and the funded hours no longer cover everything. That moment, where a family considers paying privately for additional care, is one of the few genuine growth opportunities available to you without winning a new contract. At the moment, there is no process behind capturing it. It happens informally, if it happens at all.

• A modest, structured approach to this does not require a sales team or a marketing budget. It requires knowing which clients are approaching review points, having a simple way to record that a family has expressed interest in additional hours, and being ready with a clear, honest explanation of what private-pay top-up looks like and what it costs. That is a process and a contact record, not a campaign.

Reputation, reviews and referrals

• Your CQC rating is your most credible marketing asset. Commissioners consult it. Families Google it before they call. You told us it matters far more than online reviews, and that is true for commissioner decisions. For private families discovering you independently, however, Google Reviews are often the first thing they read, and yours are not actively managed. That does not require manufacturing responses or chasing five-star ratings. It means acknowledging reviews promptly, thanking families who leave them, and having a simple, repeatable way of inviting satisfied families to share their experience when the moment is right.

• Word of mouth among families is your primary private-pay acquisition channel, and you have no formal structure around it. That is not unusual in care, but it does mean the channel is entirely dependent on individuals remembering to recommend you at the right moment. A light-touch referral prompt, whether that is a note from the care coordinator at a review, or a follow-up from the office after a positive conversation, would cost very little and would formalise something that is already happening sporadically.

Measuring what matters

• You do not measure Net Promoter Score or any equivalent, and your service-user turnover is driven by circumstances rather than dissatisfaction, which makes standard retention metrics a poor fit. That is an honest and accurate read of your situation. What is worth measuring, and what you currently cannot measure easily, is the health of your commissioner relationships ahead of tender renewals, the proportion of families who move to private-pay top-up hours, and the rate at which families refer others. None of those require a sophisticated tool; they require a contact record and the habit of noting what happens.

• Armadello would give you a single view of where private enquiries come from, how they move through to becoming active packages, and what the cost of winning each one looks like in terms of time and effort rather than media spend. That kind of visibility is what makes the difference between knowing your reputation is good and knowing which parts of your operation are actually generating new business.

• If you introduce even a basic CRM, the priority should be private-pay families and the contacts inside your commissioner organisations, not necessarily every service-user record. Starting narrow and building the habit is more useful than attempting to migrate everything at once and abandoning it after three months.

Coordinators are fielding a high volume of phone calls from families asking questions that your own visit-monitoring data could already answer: who is coming, what time, did the visit happen. That is time pulled away from care coordination to handle enquiries that a simple, proactive notification would never have generated in the first place.

Customer communications

• Your carer app captures check-in and check-out data in real time, but nothing currently pushes that information to the families who most want it. A basic alerting layer, whether a text message or an automated email triggered by a missed or completed visit, would cut a meaningful share of inbound calls without requiring a new system from scratch. This is the single highest-value communications fix available to you right now.

• You mentioned you are considering a chatbot on the website. For a domiciliary care audience, that is worth thinking through carefully before committing. A chatbot works well for answering predictable, low-stakes questions at any hour, but families in difficult situations often want to speak to a person. If you do move forward with one, it should route anything emotionally charged directly to a coordinator rather than trying to resolve it automatically.

• Email marketing is almost entirely absent. You describe rare broadcasts and no real strategy. That is not necessarily wrong given your audience and your team's capacity, but there is a gap between "rare broadcasts" and a useful, light-touch email programme for families and prospective clients. A short monthly update, visit summaries for self-funding clients, or a simple welcome sequence for new families would build trust and reduce the number of reassurance calls your coordinators receive week to week.

• Local authority communications still go out by letter in some cases. That reflects the authority's own processes as much as yours, so there is limited leverage in pushing that channel to change. What you can control is making sure your end of those communications is consistent, clearly written and sent on time, because delays or ambiguous wording in that particular relationship carry compliance and contractual consequences.

• The sign-off process for any marketing spend or creative currently sits entirely with your registered manager, with no delegated process in place. That is understandable given how little marketing activity currently happens, but it will become a practical bottleneck if communications activity grows. Even a brief written note of what decisions can be taken without escalation would prevent delays.

• Your Facebook presence is already used for recruitment advertising. If you ever wanted to run awareness or enquiry-generating content alongside that, the channel and the account are already there. That is a simpler starting point than building a new channel from nothing, though any move in that direction would need the registered manager's sign-off and a clear sense of what you are trying to achieve, given the budget sits with one person.

Internal and team communications

• Coordinators are the communication backbone of the business, but there is no mention of any structured internal channel beyond phone and the carer app for rotas. If coordinators are each handling queries individually with no shared record of what was promised or agreed, there is a real risk of inconsistency, and no easy way to spot patterns in the questions families most often ask.

• A simple shared log of recurring family enquiries, even a basic spreadsheet to begin with, would let you identify which questions come up most often and either answer them proactively or build a short FAQ resource for families. It does not require a CRM to start; it requires agreement on the habit.

• The team's time is almost entirely committed to care delivery and compliance. That is not a criticism, it is the reality of a regulated care business, and it shapes what a working relationship with any external partner needs to look like. You said infrequent, practical check-ins work best. Any communications programme Anicca supports needs to be built around that constraint, not against it, which means low-maintenance systems, clear sign-off routes and output that does not create more internal coordination work than it saves.

• You have no content programme, no style guide and no documented brand voice. What does exist, occasional recruitment posts and policy documents, was never meant to build trust with prospective clients or referrers. That means every time someone searches for domiciliary care in your area, finds your name through a referral, or checks your website to make an assessment, the content they encounter either says very little or says it in a way that was not planned.

• The inconsistency runs deeper than the website. You have told us that uniforms, vehicle branding and paperwork do not follow a consistent visual standard. A potential client or their adult child comparing two care providers will notice that kind of inconsistency, even if they cannot name exactly what feels unresolved. A one-page brand guide covering logo use, colour, tone of voice and how to describe your services would cost very little to produce and would immediately tighten how Oakhaven presents itself across every touchpoint.

• You have said you are not comfortable providing spokespeople, testimonials or case-study content for marketing use. That is a common and entirely understandable position in domiciliary care, where client confidentiality and dignity are rightly the priority. It does, however, close off the most persuasive content type in this sector. The practical workaround is to build credibility through expertise rather than through individual stories: articles and guides written from the perspective of a knowledgeable, trustworthy care provider, covering topics families actually search for, can carry real weight without requiring anyone to be named or photographed.

• The Facebook activity you already run is focused on recruitment. That is fine for its purpose, but it means the channel is doing one job when it could quietly do two. Short, informative posts aimed at families, without needing case studies or spokespeople, would cost nothing extra to produce and would begin building a different audience alongside your recruitment following. The question is whether you have the time and consistency to run both, which brings the planning problem back to the centre.

• A simple content calendar, even one covering six weeks at a time, would impose enough structure to move from reactive posting to intentional publishing. Agree three or four topics you want Oakhaven to be associated with, such as navigating care assessments, what good domiciliary care looks like in practice, or how to spot when a parent needs more support, and produce one piece per topic per month. That is a manageable output and it compounds: a well-written article published today keeps appearing in search results for years.

• There is also a compliance dimension worth noting now rather than later. The EU AI Act's transparency duty under Article 50 requires AI-generated content to be clearly labelled as such. If your content ever reaches an EU audience, that requirement applies directly. If you remain entirely UK-facing, it is still the direction that transparency expectations are heading and it is far easier to build a simple habit now than to audit old content under pressure later. That means keeping a straightforward record of which pieces were drafted with AI assistance and which were written from scratch, so that disclosure, if and when it is required, is a small administrative step rather than a scramble.

• You mentioned that staff are cautious about AI being used near care records, which is a sensible instinct. Content production is a completely separate context from care records, and there are straightforward ways to use AI writing tools only for marketing copy, with clear boundaries, without those tools ever touching clinical or operational data. Whether to introduce a sanctioned tool, set a usage policy and decide where the boundary sits is a question for Anicca's AI Adoption Roadmap rather than this report, but the gap is real and worth closing deliberately rather than leaving it to whoever has a spare moment and a browser window open.

Your website and social presence exist almost entirely to attract carers, not clients. That shapes every recommendation below.

Frameworks and contract revenue

• The framework and tender route is not a marketing channel in the way paid search or email is, but it is the single largest source of your income, and it carries real concentration risk. Three local-authority contracts and one NHS contract account for the overwhelming majority of your turnover. If one framework is retendered and you are unsuccessful, or a commissioner restructures its supply chain, the financial effect is immediate and significant.

• Your digital marketing capability has almost no role in winning or protecting framework business at present, and for many tender processes that is the correct position. However, your online reputation, your CQC rating visibility and the quality of your public-facing content do matter when commissioners do due diligence on existing and prospective providers. A weak or outdated web presence is not invisible to a commissioning team.

• Expanding onto additional frameworks, whether with neighbouring local authorities or through NHS or integrated care board routes, is a genuine growth lever that sits partly in your operations and partly in how you present your organisation in tender submissions. That is not paid media work; it is a question of bid quality and provable track record.

Carer recruitment channels

• You are already spending a modest amount, a few hundred pounds a month, on Indeed and Facebook for carer recruitment. That is the right instinct given where job seekers actually look, but there are no targets set for either channel, no cost-per-application figure tracked, and no way of knowing whether that spend is performing well or being wasted. A care business running on tight margins and depending on carer supply cannot afford to run recruitment spend blind.

• The first practical step is to set a simple benchmark: how much are you currently spending per completed application, and how does that compare with how many of those applicants reach interview or start? Neither platform requires complex setup to answer that question, but it does require someone to record the numbers consistently rather than checking in occasionally.

• Organic Facebook posts are carrying most of the recruitment load alongside the paid spend. That is not unreasonable, but organic reach on Facebook for employer pages is limited and declining. The paid budget, however small, is doing more of the work than it may appear to be.

Private-pay client acquisition

• Private enquiries currently arrive entirely through word of mouth, and there are no channel targets set, no cost-per-enquiry figure tracked, and no paid activity aimed at private families. That is a coherent position for a business whose capacity is dominated by framework commitments, but it also means that if you want to grow the private-pay proportion of your income, there is no mechanism in place to do so deliberately.

• Families searching for domiciliary care increasingly ask AI assistants, including ChatGPT, Perplexity and Google's AI Overview feature, for local recommendations. You noted this directly. Whether Oakhaven appears in those results depends largely on the quality, structure and authority of your web presence and the consistency of your information across directories and review platforms. This is an organic search and content problem as much as it is a new technology one, and it is addressable without paid media.

• If you wanted to test private-pay paid media, the sensible starting point would be a tightly scoped Google search campaign targeting families in your geographic patch who are actively searching for home care. That would be a small, time-limited test with a clear cost-per-enquiry target, not a standing commitment. Given that you have no client-acquisition paid activity running at all, the evidence base for or against it does not yet exist. Anicca's paid media work is built around cost-per-result targets rather than platform-reported numbers, which is particularly important in a sector where a single private client represents meaningful monthly revenue.

Measurement across all channels

• There are no performance targets set for any channel, paid or organic. That means there is no shared view of what a carer application costs, what a private enquiry costs, or which activity is producing results. The registered manager is handling this alongside running care operations, which makes consistent measurement practically impossible rather than just unlikely.

• Armadello would bring channel activity, website traffic, enquiry volumes and spend into one place, so you would have a single view of what each channel is producing without manually pulling reports from Indeed, Facebook and your website separately. For a business with limited marketing resource, that matters more, not less, because the registered manager needs to be able to see the picture in a few minutes rather than assembling it from scratch each time.

• The honest position is that channel investment decisions at Oakhaven are currently made without data, which means the channels that get attention are the ones that are most visible or most familiar rather than the ones that are performing. Fixing the measurement is the prerequisite for everything else, including any decision about whether to put more resource behind recruitment, private-pay acquisition, or framework tender support.

• Birdie (or your equivalent care-management platform) handles care plans, electronic medication records and call monitoring. Your rostering tool manages staff scheduling. Xero handles your finances. These are three operationally critical systems, and right now they run as separate islands. That means any change in one, a shift pattern, a care package amendment, a client ending a contract, has to be manually reflected in the others. That is where errors creep in and where admin time disappears.

• Your recruitment process sits entirely in spreadsheets and email. There is no applicant-tracking system, so candidate progress, interview notes and offer status live in inboxes and files that only the person who created them can easily read. When someone leaves or moves role, that knowledge goes with them. A lightweight applicant-tracking tool would give the whole recruitment process a single visible home, reduce the time spent chasing updates, and make it much easier to see where candidates are dropping out.

• No CRM is in use anywhere in the business. For a domiciliary care provider, the most immediate gap is on the private client side: enquiries from families, referral conversations with social workers, follow-ups with discharge teams. Without a CRM, those conversations live in people's heads or inboxes, and there is no way to see at a glance how many active enquiries you have, how long they have been sitting, or whether anyone has followed up. A basic CRM does not need to be expensive or complex to fix this.

• Your API maturity (that is, your systems' ability to share data with each other automatically, without anyone having to copy information across by hand) is low across the board. Most of your tools are used through their own screens, with no connections built between them. The practical implication is that joining anything up will require some deliberate work, not just flicking a switch. The right starting point is to check what your care-management system and rostering tool can actually export or connect to, because most platforms in this sector have grown their integration options considerably in recent years. You may find more is possible without custom development than you expect.

• You have never run a formal software audit and have no way of tracking which tools the team is actually using day to day. That matters because you are paying for things you may not be using fully, and people may have found their own workarounds (personal spreadsheets, WhatsApp threads, informal workarounds) that are invisible to leadership. A one-off audit, nothing more than a structured conversation with each team lead and a look at what licences you are actually paying for, gives you the baseline to make sensible decisions about what to keep, what to consolidate and what to replace.

• The practical first move is not to connect everything at once. Map the single handoff causing the most repeated manual work, most likely the gap between your rostering or care-management system and Xero, where hours worked need to translate into payroll and invoicing, and establish whether a direct connection or a simple automated export can remove the re-keying step. That one fix tends to have the most visible time saving and builds the case for doing more.

• Armadello would give leadership a single reporting view across channels and commercial performance rather than pulling figures from separate platform exports. That becomes particularly useful once you have more than one data source feeding into business decisions, which is where Oakhaven is heading as soon as any two systems are connected.

• Your registered manager tracks care delivery: missed visits, hours delivered against contracted hours, safeguarding counts, compliance status on DBS, training and medication records. Your finance officer tracks the invoicing side. These two views of the business never meet in one place, which means you cannot easily answer the question that matters most to a commissioner or a CQC inspector: are we delivering what we are contracted to deliver, and is that reflected accurately in what we charge?

• The five KPIs you track weekly are exactly the right ones for a domiciliary care business at your stage. The issue is not what you measure but how you measure it. Manual compilation from multiple sources every week is slow, leaves room for error, and means that by the time a number reaches whoever needs to act on it, it is already a few days old. In care delivery, a carer vacancy trend or a rising missed-visit percentage that is caught on Wednesday looks very different from the same trend spotted the following Monday.

• There is no single person accountable for data quality. That is a structural gap, not a technology problem. Even the best reporting tool will reflect unreliable inputs if the people entering data are not clear on who owns the standard. Before any system changes, you need one named person responsible for the integrity of the numbers that feed into your weekly view, with clear rules on how hours, incidents and compliance records get entered and when.

• You have no data strategy or governance document. For a regulated care provider, this matters beyond internal efficiency. You hold sensitive personal data about vulnerable people, which means your obligations under data protection law (GDPR) are more than a box-ticking exercise. A basic data governance document does not need to be long, but it should define what you collect, where it lives, who can access it, how long you keep it, and what happens if something goes wrong. That is a practical governance step you can take independently of any new tool.

What a single reporting view would change

• Right now, pulling your weekly numbers together is a manual task shared across two people with different focuses. Armadello, Anicca's reporting product, is built to bring channel and commercial performance into one live view: hours contracted versus delivered, recruitment spend on Indeed and Facebook set against vacancy rates, and operational compliance indicators alongside any digital marketing activity, all updated automatically rather than assembled by hand each week. The immediate effect is time saved and a consistent version of the numbers that both your registered manager and finance officer are looking at, rather than two separate ones.

• Once your reporting is in one place, patterns become visible that are genuinely difficult to see across separate spreadsheets. If carer turnover rises in a particular area at the same time as missed visits creep up in the same postcodes, that connection points directly to a staffing and scheduling problem. Spotting it in week two rather than week five is the difference between a corrective action and a CQC concern.

• Your recruitment paid media on Indeed and Facebook is currently serving a single purpose: filling vacancies. There is no reporting line that connects what you spend on those channels to your vacancy rate or your cost per hire over time. Bringing that spend data alongside your operational vacancy numbers would let you judge whether those channels are working in any meaningful sense, rather than running them on instinct.

The practical next step

• The first move is not a technology project. It is a decision about who owns the whole picture, followed by a short document that defines your data standards. That creates a foundation that any reporting tool can actually be trusted to build on.

• Once that is in place, consolidating your operational and commercial numbers into a single live view is straightforward and delivers a felt improvement quickly. The time your registered manager and finance officer currently spend compiling the weekly report is time that could go back into running the service.

• Your revenue is split roughly 80% local authority and NHS hours against 20% private pay, but those two income streams almost certainly carry very different margins. Local authority rates are set externally and compressed by contract, while private pay is priced by you. If you are not tracking the cost to deliver each hour by contract type, you cannot confirm which stream is genuinely more profitable after travel, mileage, coordinator time, and the agency cover you reach for when permanent carers are unavailable.

• Recruitment and agency cover sit in your top three cost categories, and that is significant. Every agency shift is a direct cost you could potentially reduce, but you can only justify the investment in retention, training, or faster in-house recruitment if you can show what agency cover is actually costing you per month against what you would spend on the alternative. That comparison does not exist yet in any formal way.

• Your marketing budget is under £1,000 a month, and by your own description it has never been treated as a real budget line alongside care delivery costs. That framing is understandable in a staffing-intensive care business, but it creates a blind spot. Even a small spend on Indeed and Facebook for recruitment has a cost per hire attached to it, and you are not tracking that number. If one channel consistently produces joiners who stay and another produces joiners who leave within three months, you have no way to see it.

• You have under ten active software subscriptions across the whole business, which means your fixed SaaS overhead is low and unlikely to be the problem. The issue is not what you are spending on tools; it is that none of those tools are currently talking to each other or feeding any kind of commercial picture. Your back-office costs are a known category but not a monitored line against output.

• You are not yet tracking any return on the marketing activity you do run. That is honest and it is also fixable without large investment. The practical first step is agreeing what numbers matter: cost per hire from Indeed versus Facebook, the margin difference between a local authority hour and a private pay hour, and what your back-office overhead works out to per client per week. None of those require new software to start; they require someone to own the calculation and run it regularly.

• Once those baselines exist, bringing them into one reporting view becomes worth doing. Armadello would give your leadership a single place to see spend, channel performance, and the commercial result behind it, rather than pulling figures from separate platform exports and spreadsheets. At your current scale and subscription footprint, the gap is not complexity; it is that the habit of connecting marketing spend to a commercial outcome has never been established.

• One note on AI spend specifically: you have no formal AI tools in use and no AI budget to account for. The informal use of ChatGPT for drafting policies is worth acknowledging as a starting point, but any decisions about AI investment or governance belong in a separate conversation. The AI Adoption Roadmap that Anicca offers alongside this work is the right place to assess what AI tools would actually be worth paying for and what controls you need around them before costs or risks accumulate quietly.

Marketing data and consent

• Your website and any recruitment or client-enquiry forms collect personal data, almost certainly including some sensitive data from people asking about care for a family member or themselves. Whether the consent wording on those forms, your cookie banner, and the privacy notice on the site actually reflect what you collect and process is something you have probably not reviewed as a single exercise. In a sector where UK data protection law (GDPR) already applies to special-category health data, that gap is not a technicality.

• The DPAs you maintain with local authorities and your care-planning software provider are the right instinct. The question is whether every other tool touching personal data, including any form plugin, email tool, recruitment platform, or analytics setup on the website, has a corresponding agreement or has even been reviewed. Most businesses add tools one at a time and do not run a corresponding check on data flows. The exposure builds quietly and you only discover it when something goes wrong.

• The practical first move is a light-touch audit of every tool, form and tracking setup on the site and in your marketing, checking what data each one collects, what the consent wording says, and whether a DPA exists. This does not need to be a large project. It needs one named owner internally and a clear list of what to fix. Anicca can run that audit as part of setting up Armadello, which would bring your channel and website performance data into one place, because doing so means mapping what your analytics and tracking tools actually collect, and that is the same exercise.

• On AI-generated content and AI interactions: the EU AI Act's transparency requirements under Article 50 come into force in August 2026, requiring clear disclosure when a user is dealing with AI or AI-generated content, and apply directly to EU-facing businesses. Oakhaven is UK-only, so Article 50 does not bind you directly, but it signals clearly where UK practice is heading. If you use AI to draft any client-facing content, including website copy or correspondence, a brief disclosure policy costs nothing and positions you ahead of where expectations are moving.

• The incoming CQC scrutiny you mentioned is worth taking seriously in this context. A regulator reviewing your digital presence and seeing inconsistent or incomplete consent practices, particularly around sensitive enquiries from families of potential care recipients, would count that against you. A clean, well-documented approach to consent and data is easier to defend in an inspection than a set of individual tools and forms assembled over time with no common thread.

AI tools and the policy gap

You mentioned that staff occasionally use ChatGPT for drafting policies. There is no AI usage policy in place.

That gap is real, and in a regulated sector where care records, safeguarding information, and sensitive data are part of everyday work, it carries specific risk around what data might be passed into an unsanctioned tool. Naming it here is the right thing to do; closing it properly, through a usage policy, staff guidance and a governance structure, is what Anicca's separate AI Adoption Roadmap is built to deliver, not this report.

If that work is not already in motion, it should be.