It sounds like you want to put a thermistor into the fluid, and look at rate of heat loss as indicating flow. This can be tricky, as a self-heating thermistor will make its own flow in still conditions. It may be good for on-off sensing.
The interesting part is that any change in the fluid's temperature makes the thermistor behave differently. So, if it has a resistance of 10 ohms @ 20C, it will be significantly lower with the temperature up at 25C.
Getting this to work requires a bridge, with a compensating thermistor held at the fluid temperature, but not self heating. It's a real pain to try to work out.