AI Use Case Evaluation
When you have a bunch of ideas for AI tools that could help your organization, it can be hard choosing what to begin with. In this tool you can fill in some information about each of your ideas and at the end we'll help you with your decision.
Fill out the form below to the best of your abilities. Try to give the scores as consistent as possible between all the solutions. We've foreseen a maximum of 7 tabs, each for a solution. Fill out as many as you like.
When you're ready, we'll aks you for your email address so we can get back to you with a few questions about your experience with this tool and how we can help you. After receiving your email address we'll send you an access code to unlock the end report.
The end report gives an overview of the ideas you filled in, plotted on a decision matrix. This tool will also give you a short list of "next steps" you can take: with which ideas do you go forward and which ones do you postpone to later.
Remember, this is an automated tool. It will give an output based on the data you entered. Use these results at your own risk. Also keep in mind that the conclusions are generic and not tailored to your organization so some suggestions might not fit the size of your organization or the budget you have.
help
help
help
help
help
Taking all this into account, what is your best guess at the business value?
help
help
Please leave your email address so we can get back to you with a few questions about your experience with this tool and how we can help you. After receiving your email address we'll send you an access code to unlock the end report.
We've sent you a mail with your access code, please check your inbox, copy the access code from this email and paste it in the box below.
The code you entered is incorrectReport
Remember, this is an automated tool. It will give an output based on the data you entered. Use these results at your own risk. Also keep in mind that the conclusions are generic and not tailored to your organisation so some suggestions might not fit the size of your organisation or the budget you have.
Based on the information you entered, these are the conclusions:
Decision matrix
Below is a chart that shows all of the solutions in one go. Every bubble on this graph represents a use case, the size indicates the maturity.
Prioritization
Champions
These solutions score the highest on both feasibility and business value. You should consider these cases first when allocating budgets. They'll have the highest chance of giving you a good return on investment.
Research cases
The research cases will deliver a lot of business values if you have a chance to overcome the technical challanges ahead. This is a high-risk high-reward area. You might want to choose one of these cases to invest some limited budget in, just to see how far you can get. And maybe you're in a good position when the technology catches up to your idea and the feasibility shoots up.
Incremental improvements
Here are some potential quick wins, if the cost to implement is low enough. But don't expect much more then break-even. Consider doing just one or two of these cases.
Reasses later
Both business value and feasibility are low. For now there is no sense in investing in these cases. But, conditions can change. The technology could move forward or the market could change. So don't forget about these ideas.
Next steps
Based on the maturity score, there's a different next step to take.Idea
Everything starts with an idea. Next is working out the business value in more detail. You've made a high level assesment of the business value, but you can work this out better, put some numbers behind it, map out exactly how the tasks and procedures that you want to automate work.
When you're using the bottom-up approach: let the business sponsor of this solution do this analysis but give them support from a professional business analyst. You can timebox the effort of both the business sponsor and the business analyst (let's say 5 days and 2 days respectively). Also consider doing the business analysis for all cases at this level, the investment is limited and you might discover errors in your initial estimations.
Analysed
Those ideas that have been analysed further are now candidates for (self-)experimentation. This means that the technical principles behind this solution should be tested using off the shelf software. You can - or rather should - use fictitious data to see if the AI system can reach some reasonable and acceptable result.
For both bottom-up and top-down approaches, it's a good idea to do this experimentation in-house. You have a better insight in the data, the procedures and what constitutes an acceptable result. Do consider getting some help by an AI engineer or data scientist that can give you some practical information, tips and tricks on how to move forward. Such a specialist can also give some insights in what behaviour can be fixed in a custom implementation and what can not. Time can again be timeboxed for both your organisation and the AI specialist (let's say 5 days and 3 days respectively). Also consider doing the business analysis for all cases at this level, the investment is limited and you might discover errors in your initial estimations.
Build
When the analysis and experimentation phases are done, you can consider starting to build somthing. Here it is imperative that you select the cases to move forward with using the prioritization lists above. Your investment will be significantly higher so do choose the case(s) to move forward with carefully.
PoC
First start with a PoC (or Proof of Concept). This is a rudimentary implementation of the solution you envision. It is stand alone (cloud) application that is not integrated with your own IT systems. Both the AI model and the surrounding code can be tailored to your use case so you can expect it to perform better then the experimentation that you did earlier.
Pilot
A PoC was created and the assumptions on business value and feasibility still hold (or perhaps, have improved)? The you can consider building a pilot (or MVP - Minimal Viable Product). The pilot will be integrated with some of your systems and once it's done, you should see some return on investment.
Industrialization or iteration
The pilot is finished. Feasibility and business value should be clear now. Based on the business value you can now choose to continue with rolling out the solution on a larger scale. You can also start planning to iteratively add features or improve performance.