Key takeaways
- Return on technology in a practice is mostly time, access and risk, not just cash; define which benefits you expect before you buy.
- Measure a baseline for at least four weeks before go-live, using the same measures you will use afterwards.
- Count the full cost: licences, set-up, training, staff time, integration and governance work such as a DPIA.
- Review at three, six and twelve months, and be willing to change or stop a tool that is not delivering.
Why practices struggle to judge technology
Practices buy technology for good reasons: to make phones easier to get through, to cut admin, to give patients more ways to make contact. Yet many cannot say a year later whether a new system was worth it. The usual cause is not the technology itself but the lack of a baseline, a clear statement of the expected benefits, and a planned review.
Technology can make things better or worse. In NHS England trials, moving GP practices to high-quality digital telephony raised patients' ease of getting through scores by 30 percentage points[1]. But in a BMA survey of more than 1,300 practices, 74% saw workload rise after online access changes, 68% reported more stress and 55% reported negative effects on patient care[2]. The same category of change can help or hurt depending on how it is implemented, which is exactly why measurement matters.
30 points
rise in patients' ease of getting through after practices moved to high-quality digital telephony in NHS England trials[1]
Define the return before you buy
In a practice, return rarely means profit alone. Most benefits fall into five groups, and a business case should state which ones it expects, how each will be measured and by how much it should change.
| Benefit type | Example measures | How to collect |
|---|---|---|
| Time saved | Minutes per task, hours per week on a process | Task log before and after |
| Access | Call wait from 8am to 10am, abandoned calls, same-day urgent access | Telephony and clinical system reports |
| Income | Claims submitted on time, missed appointments refilled, QOF points | Claims records, clinical system |
| Risk | Incidents, errors in coding or results handling, information governance breaches | Incident log, audits |
| Experience | Patient and staff feedback, complaints on the relevant theme | Friends and Family Test, complaints log, staff survey |
Pick two or three primary measures and set a target for each. A telephony upgrade might aim to halve abandoned calls in the morning peak; a document workflow tool might aim to cut letter handling time by a third. Targets keep the review honest.
A one-page business case
Partners and owners need a short, consistent format for technology decisions. One page is enough if it answers the questions that matter, and using the same format every time makes options easier to compare.
- Problem: what is not working today, with the baseline figure.
- Proposal: the tool and the process change that goes with it.
- Expected benefits: two or three measures with targets and a date.
- Full cost: first-year and ongoing, including staff time and governance work.
- Risks: clinical safety, data protection, supplier dependence, staff workload during change.
- Alternatives: including doing nothing and changing the process without new technology.
- Review plan: who will review, when, and what would make the practice stop.
The alternatives line is the one most often skipped and the most useful. Many problems that look like technology problems turn out to be process problems, and a rota change or a clearer script can deliver much of the benefit for almost no cost.
Count the full cost
Licence fees are the visible part of the cost. The rest is often larger in the first year and is easy to forget when comparing options.
- Set-up and migration, including any data transfer from the old system.
- Training time for every user, and backfill while they are trained.
- Process redesign: the hours spent deciding how work will change.
- Integration with the clinical system and other tools.
- Governance: data protection impact assessment, information sharing agreements, clinical safety work and supplier assurance.
- Ongoing administration: user management, configuration changes and supplier meetings.
- Exit costs: notice periods and the cost of switching if the tool does not work out.
Central funding can offset some costs. NHS England re-targeted £240 million to help practices still on analogue phones move to digital telephony, worth about £60,000 of support per practice over two years[3]. Ask your ICB what support exists before committing practice money.
Baseline, then go live, then measure
Measure the baseline
Collect your primary measures for at least four normal weeks before go-live, avoiding bank holidays and unusual events.
Write down the plan
Record the expected benefit, the measure and the target, and who will review it and when.
Go live with support
Allow for a dip in performance during the first weeks while staff learn the new system; do not judge it then.
Measure at three months
Compare with the baseline, ask users what has changed and fix early problems.
Measure at six and twelve months
Confirm whether benefits are sustained and whether any unexpected costs or risks have appeared.
Decide
Keep, change or stop, and record the reason so the next decision is better informed.
Name one person as the benefits owner for each new system. They do not need to be technical; they need to collect the measures, gather feedback and bring the review to partners on time. Without a named owner, reviews slip and the system becomes part of the furniture whether or not it earns its place.
National programmes show what is possible at scale: the NHS App saved an estimated 4.1 million hours of staff time in 2025/26[4]. Local tools deserve the same discipline of measuring time saved rather than assuming it.
AI tools need extra care
AI tools are spreading quickly. In a 2025 Nuffield Trust and RCGP survey of 2,108 GPs, 28% said they currently use AI tools in clinical practice[5], and 11% of all GPs use tools they obtained independently rather than through their practice[6]. GPs who do not use AI report serious concerns: 89% worry about professional liability and medico-legal issues, 88% about a lack of regulatory oversight, and 82% about patient privacy and data security[7][8].
Those concerns should shape the business case. The MHRA says many software and AI products used in health and social care are regulated as medical devices[9], so check a tool's regulatory status, ask for its clinical safety documentation, complete a data protection impact assessment and make sure staff use only approved tools. Self-obtained tools that nobody has assessed are a governance risk, whatever time they appear to save.
Example: a fictional telephony review
Example: Kingsway Surgery, a fictional practice, moves to cloud telephony with queue position messages and a call-back option. Its four-week baseline shows an average wait of nine minutes between 8am and 10am and about a fifth of morning calls abandoned. The business case targets a wait under five minutes and abandoned calls halved within six months.
Example: at three months the wait is six minutes and abandoned calls have fallen by a third, but staff report that call-backs bunch at 11am and clash with prescription work. The practice moves prescription processing to the afternoon. At six months both targets are met, and the twelve-month review confirms the benefit held through winter. The review also records an unexpected gain: fewer complaints about access. The decision to keep the system is written up with the numbers so the next technology decision starts from evidence.
Questions to ask suppliers
- Which practices like ours use this, and can we speak to one without you on the call?
- What measures did they track, and what changed?
- How does it integrate with our clinical system, and who supports that integration?
- What data does it hold, where, for how long, and under what agreement?
- Is it regulated as a medical device, and what clinical safety documentation do you provide?
- What does it cost to leave, and how do we get our data back?
Write the answers into the business case. They become the checklist for the three-month review and protect the practice if the supplier's promises and the reality differ.
Honest reviews
- Compare with the baseline, not with memory.
- Include staff time spent on the system, not just time saved.
- Ask patients and staff what changed, not only whether they like it.
- Look for work that moved elsewhere rather than disappeared.
- Record the decision and the reason, including when a tool is stopped.
Stopping a tool that does not work is a good outcome of a review, not a failure. Keeping a system nobody uses costs money every month and teaches staff that reviews do not matter.
Frequently asked questions
How do you measure the return on practice technology?
Define the expected benefits and measures before buying, collect a baseline for at least four weeks, count the full cost including staff time and governance, and review at three, six and twelve months.
What costs are usually missed in a technology business case?
Training and backfill, process redesign, integration, governance work such as a DPIA and clinical safety checks, ongoing administration and exit costs.
What extra checks do AI tools need?
Check whether the tool is regulated as a medical device, ask for clinical safety documentation, complete a DPIA and make sure staff only use approved tools.
References
NHS England and DHSC (2023). Delivery plan for recovering access to primary care (opens in a new tab) ↑
British Medical Association (2026). Pressures in general practice data analysis (opens in a new tab) ↑
NHS England and DHSC (2023). Delivery plan for recovering access to primary care (opens in a new tab) ↑
NHS England (via GOV.UK) (2026). NHS England annual report and accounts 2025 to 2026 (opens in a new tab) ↑
Nuffield Trust (2025). How are GPs using AI? Insights from the front line (opens in a new tab) ↑
Nuffield Trust (2025). How are GPs using AI? Insights from the front line (opens in a new tab) ↑
Nuffield Trust (2025). How are GPs using AI? Insights from the front line (opens in a new tab) ↑
Nuffield Trust (2025). How are GPs using AI? Insights from the front line (opens in a new tab) ↑
GOV.UK (Medicines and Healthcare products Regulatory Agency) (2025). Software and artificial intelligence (AI) as a medical device (opens in a new tab) ↑