Free · 12 questions
Staff AI Risk Scorecard
Are your staff using AI safely? Score how your team uses ChatGPT, Copilot and other AI tools, get a prioritised fix list and a free AI use policy template.
Free tool · AI Automation ROI Calculator
How much could AI automation save your team?
Estimate the hours and money AI automation could save your marketing team each year, how quickly it pays back and the return on investment. Adjust the tasks, time and costs to match your team and the results update instantly.
The calculator
Your team, your numbers, your return.
The defaults describe a typical five-person marketing team. Change anything and the estimate updates as you type.
Want a costed automation plan for your team?
Send us your scenario and we'll reply personally with the three automations we'd build first, and what they'd realistically cost. Free, with no obligation.
The guide
How do you calculate the ROI of AI automation?
The ROI of AI automation is the value of the time it saves, minus what it costs, divided by what it costs. To work it out, you need four numbers: the hours your team spends on a task, the share AI can take on, your hourly cost and the cost of building and running the automation.
- Hours saved per year = hours on the task per person per week × share automated × people × adoption rate × working weeks
- Value of time saved = hours saved per year × hourly cost
- Net benefit = value of time saved − tool costs − setup cost
- ROI = net benefit ÷ (setup cost + tool costs) × 100
- Payback period = setup cost ÷ (monthly value of time saved − monthly tool costs)
A worked example
The calculator's default scenario is a team of 5 people costing £45 an hour, working 46 weeks a year, with a 70% adoption rate.
| Task | Hours per person | Share automated | Team hours saved |
|---|---|---|---|
| Content drafting and repurposing | 6 hrs | 30% | 6.3 hrs |
| Reporting and dashboards | 3 hrs | 50% | 5.2 hrs |
| Research and briefs | 4 hrs | 30% | 4.2 hrs |
| Email and lead follow-up | 3 hrs | 40% | 4.2 hrs |
| Admin, tagging and data entry | 3 hrs | 50% | 5.2 hrs |
That's 25.2 hours a week, or 1,159 hours a year, worth £52,164. With a £8,000 setup cost and £400 a month in tools, the first-year net benefit is £39,364, a first-year ROI of 308%, and the setup cost pays back in about 2 months.
Which marketing tasks can AI automate?
The best candidates are repetitive, high-volume and low risk if something goes wrong. In almost every case, AI should do the heavy lifting while people keep the judgement.
| Task | What AI can handle | What people should keep |
|---|---|---|
| Content drafting and repurposing | First drafts, turning one piece into many formats, summaries, metadata | Strategy, original insight, final edit and sign-off |
| Reporting and dashboards | Pulling data from platforms, building weekly reports, writing commentary on changes | Interpreting results and deciding what to do next |
| Research and briefs | Competitor and topic research, collating sources, drafting briefs | Choosing the angle and checking facts |
| Email and lead follow-up | Enriching and scoring leads, personalising follow-ups, routing to the right person | Relationship-led conversations and judgement calls |
| Admin, tagging and data entry | Categorising content, updating CRM fields, tidying spreadsheets | Spot checks and exception handling |
| Internal knowledge | Answering “where's the latest brand guideline?” questions from your own documents | Keeping the source documents accurate |
How much time does AI actually save?
It varies widely by task and by how well the workflow is built. In Microsoft and LinkedIn's 2024 Work Trend Index, a survey of 31,000 people across 31 countries, AI power users reported saving over 30 minutes a day, and 75% of knowledge workers said they already use AI at work.
Individual use is not the same as automation, though. The same research found 78% of AI users bring their own tools to work, which usually means ad hoc prompting rather than shared workflows. The bigger savings come when a task runs end to end, with people reviewing the output instead of producing it.
The most reliable figure is your own. Time the tasks you plan to automate for two weeks before you start, and measure again after launch.
What does AI automation cost?
Most of the cost sits in five places:
- Design and build: mapping the workflow, building it and testing it with real data. Usually the largest one-off cost.
- Tool subscriptions: the automation platform and any AI tools, usually charged monthly per user or by volume.
- AI model usage: pay-as-you-go costs for the AI models an automation calls, which grow with volume.
- Training and handover: time for your team to learn, trust and adapt the new workflow.
- Maintenance: occasional fixes when a connected tool changes or a prompt needs updating.
If an automation handles personal data, for example to score leads, UK GDPR applies. Build it in from the start using the Information Commissioner's Office guidance on AI and data protection.
Why do AI automation projects fail to deliver ROI?
- Automating the wrong task. Low-volume or judgement-heavy work saves little and risks a lot.
- No baseline. Without timing tasks first, nobody can prove the saving afterwards.
- Low adoption. If people work around the automation, the saving never arrives.
- No human review. Errors that slip through create rework that cancels out the time saved.
- Tool sprawl. Overlapping subscriptions eat into the return. Our AI tool implementation work fixes this first.
How the calculator works
Results update live from the formulas above. A few assumptions sit underneath:
- One full-time person's capacity is taken as 37.5 hours a week.
- Tool costs are counted from month 1. The setup cost is counted once, in year 1.
- The conservative case halves the time saved and keeps every cost the same.
- Everything runs in your browser. Nothing you enter is stored.
Frequently asked questions
What is a good ROI for AI automation?
It depends on your alternatives, but an automation that pays back within 12 months is usually a strong investment, and many simple workflows pay back far sooner. Always compare against a conservative case: if the numbers still work at half the expected saving, the project is low risk.
How do I work out my team's hourly cost?
Take annual salary, add on-costs such as employer National Insurance, pension and benefits (often 20 to 30% for UK employers), then divide by working hours in a year, roughly 1,650 for a full-time employee. A £60,000 salary works out at around £45 an hour on that basis.
What does the adoption rate mean?
It's the share of the theoretical time saving you'll actually realise. Not every task runs perfectly and not everyone uses a new workflow straight away. 70% is a sensible planning figure; raise it once you've measured real usage.
Does time saved mean we can reduce headcount?
Usually the saving is realised as capacity rather than cuts: more campaigns, faster turnaround or time back for strategic work. The value is real either way, which is why the calculator expresses it as the cost of the hours you get back.
Which tasks should we automate first?
Start with work that is high volume, repetitive and low risk if something goes wrong, such as reporting, data entry and repurposing content. Leave judgement-heavy work, like strategy and sensitive client communication, with people.
How accurate is this calculator?
It's an estimate based entirely on your inputs. The most reliable way to improve it is to time the tasks you plan to automate for two weeks before you start, then measure again after launch.
Is anything I enter stored?
No. The calculator runs in your browser and nothing is sent or saved. If you request a costed plan, we receive your scenario summary and the details you enter, nothing else.
Keep going
Related free tools.
Free · Up to 28 prompts
AI Visibility Prompt Generator
What are your buyers asking AI? Generate the prompts to test in ChatGPT, Perplexity and Gemini, with a free sheet to track whether AI mentions you.
Free · 10 checks
AI Prompt Injection Scanner
Is your website hiding instructions for AI? Paste a page's HTML to find invisible text, comments, attributes, schema and Unicode tricks aimed at ChatGPT, Claude and AI agents.