Fourteen interactive tools for deciding whether an AI or automation project is worth doing. No email address, no PDF in your inbox, no follow-up sequence. Everything calculates in your browser and nothing is sent anywhere.
What most tools like these do
These are free interactive calculators, scorecards and assessments for planning AI and automation work: readiness scoring, payback estimates, RPA suitability, risk tiering and integration routing. All fourteen run entirely in your browser, require no email address, and send nothing to a server.
Two questions. Fourteen tools is more than anyone wants to browse, and the right one depends entirely on which decision is in front of you this week.
What are you trying to decide?
How long have you got?
Start here
A first-pass estimate of how many hours across your organisation are genuinely in play. Rough by design, and enough to know whether this is a conversation worth having at all.
Open the toolThen
Follow it with the payback calculator once you have a specific process in mind.
Each one opens on the page that explains what it models and what it deliberately leaves out, because a number without that context is worse than no number.
Showing all 14 tools.
Twelve questions across data, process, integration and governance, scored per dimension with live bars. It names your weakest dimension and says what that constrains, rather than averaging four scores into one comfortable number.
Open the assessmentHow long until the build repays the effort of building it, expressed in weeks of running rather than a currency figure. Takes volume, handling time and exception rate, and shows why a high exception rate destroys payback.
Open the calculatorHow much agent handling time is genuinely available to a chatbot, based on ticket volume, repeat-question share and average handling time. It models what a bounded assistant can reach, not a headcount reduction.
Open the calculatorPrices the waiting between steps rather than the hands-on work, which is where most process cost actually sits. Frequently the number that changes which project gets funded.
Open the calculatorThe broadest of the five cost tools: how many hours across an organisation are plausibly in play. Deliberately rough, and the right starting point when you have no specific process in mind yet.
Open the calculatorScores whether a process should be a robot at all. It is built to return "use the API instead" when that is the truthful answer, which is the recommendation an RPA vendor has no incentive to make.
Open the scorecardFour questions about one system that return the most durable route in: API, database, file exchange or robot bridge. It also flags the things to ask a vendor for in week one rather than week six.
Open the pickerAn exercise rather than a calculator. You pick the bottleneck in a sample process, then the measured data appears. The busiest-looking step is almost never the constraint, and being wrong once tends to stick.
Try the exerciseFive questions about consequence, returning an indicative band and the obligations to plan for. It states on the tool that classification is a legal determination for your counsel, because it is.
Open the classifierFive levels from suggest-only to fully autonomous, each showing the controls it requires and what it costs to operate. Most teams find they want level two and were being sold level five.
Open the dialMove the document confidence threshold and watch review hours trade against errors reaching your systems. There is no setting where both are zero, which is the honest shape of the decision.
Open the simulatorFour weight sliders over five sample leads, with the queue re-ranking live. Built for two teams to argue over in a room, because the disagreement it surfaces is worth more than the ranking.
Open the builderDrag the forecast horizon and watch the confidence band widen. Useful for anyone being shown a clean twelve-month projection line by a supplier, because that line is hiding this.
Open the explorerA decision tree across twelve outcomes that routes a described situation to the right service line. It will sometimes tell you the answer is a process change rather than AI, and that is a real outcome.
Open the routerMost tools in this category ask for an address before showing you a result. We decided against it, and the reasoning is worth stating rather than just claiming a virtue.
If you have to give an address to see your own score, the tool is a lead form wearing a calculator's clothes, and its output gets designed to justify a follow-up call. Removing the gate removes that incentive.
We hold ISO 27001 certification and the simplest way to protect data is not to have it. Everything computes client-side, so there is no record of what you typed and no retention question to answer under the DPDP Act.
Several of these tools are designed to return an answer that costs us work: use an API instead of a robot, the process is not ready, the model will not beat your planner. That is more useful to you and, over time, to us.
If you would rather talk it through than click through, that route is a consultation, and it is a separate decision from using anything on this page. The full service detail sits in the AI and automation practice.
Four limits worth knowing before you take any output of these into a meeting.
Each tool encodes a decision we make repeatedly during delivery. The page it sits on explains the reasoning behind it.
Home of the 12-point readiness assessment and the roadmap phases behind it.
Two tools live here: the payback calculator and the bottleneck exercise.
The suitability scorecard, plus the fragility discussion that explains its scoring.
The risk tier classifier, with control mapping across three frameworks.
The deflection calculator, alongside a grounding trace that shows a refusal.
The handoff cost calculator and the node-level failure inspector.
The autonomy dial, and a run that halts at a human approval gate.
The confidence threshold simulator and the review queue design behind it.
The lead scoring builder, with the case for readable weights over learned models.
The uncertainty explorer, and why a baseline matters more than an accuracy figure.
The route picker, and the four-route hierarchy it applies.
Ticket triage and voice call flow walkthroughs, if support is your starting point.
Also worth a look: generative AI development for the evaluation methodology these tools inherit, our IT consulting practice for wider technology strategy, and the blog or help centre for written detail.
Including the ones about what happens to your data, which is none of it.
Send us what a tool gave you and what you put into it. We will tell you whether it holds up, including when it does not. We reply in under 3 minutes.
