You have been asked to state how AI was used in producing your tender. The buyer wants a statement that can be compared with other tenderers' statements. That is why they have named an instrument rather than asking an open question.
It takes about five minutes. It is free. You do not need an account, and nothing you enter is stored or sent to anyone.
What you do
Open the toolkit and choose Bid or submission as the medium. You are then asked about six activities, and for each you say whether a person did it, AI helped, or AI did it.
The approach taken
The substance of the response
Written drafting
Costing and commercials
Editing and compliance
Design and layout
The tool calculates a band from your answers. You do not choose the band, and there is no way to select one directly. Copy the resulting code, which looks like TM-C@2.0, and put it in your tender with the six answers beside it.
Five graded bands run from Minimal through Limited, Moderate and Considerable to Extensive. Two further declarations sit outside that scale: No AI Involvement, and AI Generated where AI led every activity. Those two are declared rather than calculated, because they are statements about the work rather than measurements of it.
TM-NONEABCDETM-GEN
No AI Involvement
Declared
Minimal
0 < r ≤ 0.10
Limited
0.10 < r ≤ 0.36
Moderate
0.36 < r ≤ 0.63
Considerable
0.63 < r ≤ 0.89
Extensive
0.89 < r < 1
AI Generated
Declared
Bid or submission
A tender is scored out of 40 points. That figure is a property of this medium and differs for other kinds of work, which is why a band is always stated with the medium it was scored under.
Answer honestly, and answer for the work as submitted
Two things catch people out.
Answer for what actually happened, not for how you would do it next time. And be careful with the activities where AI proposed something a person then approved. Approval is a human act, but the proposal came from AI, so that is AI helped rather than a person did this. This is the most common place a declaration drifts optimistically, and we know because it happened to us.
No band is a bad answer
The framework does not rank the bands and no colour or icon anywhere on the scale implies a worse result. A tender that is Considerable is not a worse tender than one that is Minimal. The buyer is asking what happened, not judging it.
An inaccurate declaration is a different matter, and buyers may ask you to substantiate it.
If you want to check a code
Any code can be validated at the framework's validation endpoint, which confirms whether it is well formed, which model and version it belongs to, and what band it resolves to. Buyers can use this to check a code a tenderer has submitted is real.
CC BY 4.0The framework is open and free to use. It requires no registration and no membership.
Analytics
We would like to measure how this site is used, so that we can improve it. Nothing is measured and nothing is contacted until you accept. Your answer is kept in this browser, so clearing this site's data asks again.