Comparing Magic Notes and Servent Notes for AI-assisted case recording, including deployment, pricing, social-care evidence, integration and the questions councils should test before choosing a platform.
01
Established supplier-hosted product with strong social-care adoption.
02
Servent Notes runs in the customer's own Microsoft Azure subscription.
03
Magic Notes has stronger published evidence.
04
Compare the complete workflow, not just the AI-generated draft.
Magic Notes and Servent Notes at a glance
AI case recording can reduce the time frontline teams spend turning visits, assessments and conversations into formal records.
Magic Notes and Servent Notes both address that problem. They take different approaches to deployment, ownership, commercial structure and how the technology fits into the organisation’s wider systems.
Magic Notes is an established supplier-hosted product with strong adoption across social care and public services. Servent Notes is deployed into the customer’s own Microsoft Azure subscription and configured around the organisation’s workflows and technology environment.
For councils comparing the 2, the useful test is the complete process:
conversation → structured draft → practitioner review → approved record → system of record
The amount of work, risk and cost remaining across that process matters more than the quality of a product demonstration.
Area | Magic Notes | Servent Notes |
|---|---|---|
Deployment | Supplier-hosted public cloud service | Deployed into the customer’s Microsoft Azure subscription |
Data environment | Beam-hosted service with UK data processing stated in its current G-Cloud listing | Customer’s Microsoft Azure environment, using agreed Azure services and regions |
Identity | Product-managed authentication including 2FA | Can use the customer’s Microsoft Entra ID |
Offline recording | Yes | Yes |
Translation | Available | Available |
Case-management workflow | Current G-Cloud listing describes copy-and-paste/export and lists no API | Route into existing systems can be scoped as part of deployment |
Commercial model | Annual subscription based primarily on recording volume | Implementation, platform subscription and customer Azure consumption |
Adoption evidence | Extensive public-sector and social-care use | Earlier-stage product with public-sector Azure and social-care delivery experience |
The deployment model changes the buying decision
A single recorded visit can create audio, a transcript, AI inputs, generated drafts, the approved record, logs and backups.
Buyers need to know where each of those sits, who can access it and who is responsible for the technology around it.
Magic Notes is operated by Beam as a hosted public-cloud service. Its current Digital Marketplace listing says the service uses AWS and Azure, with data stored and processed in the UK. Beam manages the application and underlying service.
Servent Notes takes a different approach.
The application is deployed into the customer’s Microsoft Azure subscription. It can use the organisation’s Entra identity, Azure governance and monitoring controls, with the customer retaining ownership of the Azure application environment. The precise location of AI processing depends on the Microsoft services and regions selected during deployment, so the data flow and responsibility model need to be documented during architecture and information-governance review.
The deployment choice gives the council a different control model.
A supplier-hosted service may suit an organisation that wants the supplier to operate the application and infrastructure. A council with an established Azure estate or desire may place more weight on having the application deployed within cloud infrastructure it owns and governs.
Magic Notes has more social-care evidence today
Magic Notes has a clear advantage in published adoption evidence.
Kingston Council used the product across care reviews, occupational therapy assessments, supervisions and care-needs assessments. Beam reports a 68% reduction in administrative time during the pilot, with an average of 1.33 hours saved per assessment and 96% accuracy across transcripts and summaries. Practitioners still reviewed and edited the generated records before adding them to the case-management system.
The Local Government Association’s account of the same work reports initial time reductions of 50% to 60% for case notes and assessments. It also records limitations found during testing, including poorer results where speech was less clear, reduced verbal communication was involved or background noise affected the recording.
The product has also moved beyond early pilots in several councils. Hackney awarded Beam a £61,200 extended social-care pilot covering 2026, while Devon awarded a £92,000 contract following a children’s services pilot. Ealing awarded a £150,000 contract for Magic Notes in statutory social care during 2026.
That body of evidence should carry weight in a procurement exercise.
Servent Notes is earlier in its product adoption. Servent has public-sector Azure experience and has worked on an Azure application for social-care recording, including mobile capture, transcription and AI-generated draft summaries, but the published Servent Notes evidence base is not yet comparable with Magic Notes.
A council considering Servent Notes should therefore use a controlled pilot to establish its own evidence before deciding on wider deployment.
The commercial models work differently
Magic Notes publishes its pricing.
The latest G-Cloud pricing document, updated July 2026, lists the following annual prices for Magic Notes Core:
Annual recording allowance | Annual fee |
|---|---|
Up to 6,000 hours | £30,000 |
Up to 12,000 hours | £48,000 |
Up to 25,000 hours | £75,000 |
Up to 50,000 hours | £112,500 |
Up to 100,000 hours | £175,000 |
Up to 150,000 hours | £225,000 |
Those subscriptions include unlimited users, unlimited templates and unlimited in-use support. The Core product includes recording, transcription, summaries, offline recording and related case-recording functionality.
Servent Notes uses a different cost structure because it is deployed into the customer’s Azure environment.
The customer needs to consider:
initial deployment and configuration;
the Servent Notes platform subscription;
Microsoft Azure consumption;
any additional integration or bespoke development required.
The useful comparison is therefore total production cost at the expected level of use.
A council expecting thousands of hours of annual recording should model both options against realistic adoption, rather than comparing the starting subscription price alone.
For the purposes of this comparison, a Servent Notes service supporting around 50,000 recording hours a year would cost approximately £87,500 annually, based on a £75,000 platform subscription plus around £12,500 of Azure consumption.
The subscription covers platform access, users, templates, product updates and standard support. Azure consumption is paid through the organisation’s own Microsoft Azure subscription, so the customer can see and control the underlying cloud cost directly.
Unlike Magic Notes, the current Servent Notes model does not increase the platform subscription according to recording volume. Higher usage primarily affects the customer’s Azure consumption.
Integration with case-management and other line-of-business systems can also be scoped where required.
Templates should be tested with real records
Template quality and practitioner editing effort are more useful measures during a pilot.
Magic Notes’ current pricing includes unlimited templates, and Beam states that templates can be configured around customer requirements.
A better evaluation is to give each supplier the same real examples.
That could include:
a Care Act assessment
a safeguarding record
a home-visit note
a supervision record
another document already used by the service
Then measure what comes back.
How complete is the draft? How much editing does the practitioner need to do? Are important statements easy to check against the source? Does the language fit the organisation’s existing records?
A council learns more from 20 practitioners using their own documents than it does from seeing 150 templates in a product demonstration.
Language support may matter for some services
Magic Notes currently supports translation for sessions held in languages other than English, including Welsh.
Servent Notes includes multilingual translation at a transcription or record level within its standard production scope.
For a council where multilingual case recording is a material requirement, that difference should be included in the pilot criteria.
Accuracy should also be tested using the languages, accents, environments and service-user groups practitioners encounter in practice.
Follow the record all the way into the case-management system
Producing a draft case note removes part of the administrative task.
The approved information still has to reach Liquidlogic, Mosaic, Eclipse or another system of record.
Magic Notes’ current G-Cloud listing says users copy summaries and transcripts into internal case-management systems. Administrators can export written information as CSV or PDF and download audio, but the current listing states that the product does not provide an API.
That may be entirely adequate for some teams.
For others, copying an AI-generated record from one application into another leaves a manual step that becomes significant at scale.
Servent Notes deployments can take the required downstream workflow into account when the solution is designed. The integration requirement and associated cost should be agreed during the deployment scope rather than assumed.
How many manual steps remain between finishing the conversation and putting the approved record into the system that the organisation relies on?
Test the complete workflow
A fair pilot should give competing products comparable users, scenarios and measures.
Measure:
time spent creating records before and after;
time spent reviewing and correcting generated drafts;
the number and type of corrections;
completeness against the required record format;
practitioner confidence;
manager assessment of record quality;
recording failures and offline performance;
time until the approved record reaches the case-management system;
adoption and support requirements;
information-governance findings;
expected production cost at the required scale.
Human review belongs in those measures.
Kingston’s experience shows why. Practitioners saved substantial time using Magic Notes, while still changing transcripts and summaries to add detail and reflect their professional writing.
The useful measure is therefore net practitioner time saved after review and correction, rather than the speed at which AI produces its first draft.
Which approach fits your organisation?
Magic Notes has a strong social-care footprint, published customer evidence and a supplier-operated SaaS model. Its commercial structure is published and primarily tied to recording volume.
Servent Notes gives organisations another deployment model. It runs within the customer’s Microsoft Azure subscription and can be configured around the organisation’s identity, Azure controls, record structures and required technology workflow.
The relative importance of those differences will vary.
A social-care team may put existing practitioner evidence first. An information-security team may care more about application ownership and data flows. An enterprise architect may concentrate on identity and integration. Finance and procurement will want to understand what happens to cost when a successful pilot becomes a council-wide service.
Those requirements should be agreed before suppliers are scored.
More from the blog:
