The reset_index() is your friend in this case. See the example below:
Creating Test Dataframe
>>> import pandas as pd
>>> df = pd.DataFrame({"Escritorios": ["A", "A","A","A","A"], "Municipios": ["Aruja", "Poa", "Poa", "Aruja", "Poa"], "Acessos": [1, 2, 2, 1, 2]})
>>> df
  Escritorios Municipios  Acessos
0           A      Aruja        1
1           A        Poa        2
2           A        Poa        2
3           A      Aruja        1
4           A        Poa        2
Grouping data
>>> df.groupby(["Escritorios", "Municipios"])["Acessos"].sum()
Escritorios  Municipios
A            Aruja         2
             Poa           6
Name: Acessos, dtype: int64
Grouping data and using reset_index()
>>> df.groupby(["Escritorios", "Municipios"])["Acessos"].sum().reset_index()
  Escritorios Municipios  Acessos
0           A      Aruja        2
1           A        Poa        6
That is, just assign to a new dataframe and save to Excel
							
							
						 
Could you please share a little more of the code? Like the
planta_dfis structured, what data is in it?– Matheus Delazeri