A bit similar to what osnowyo proposed
Heraclius wins the battle of Antioch in 613.
It's a bit difficult because Heraclius is still a new commander, Sharbaraz is one of the veterans of the Persian army, apart from being a rather gifted commander and Heraclius in the 620s is very different from Heraclius in the early 610s, but not impossible.
Now, if Heraclius wins decisively enough to force the Persians to withdraw east of the Euphrates or at least force Sharbaraz to halt his advance and fall back temporarily to reorganise his army, then the Byzantines have more time to organise their defence and counterattacks. Philippicus would also benefit from this, as he might be able to exploit this temporary problem of the Persians to advance even deeper into Armenia, thus forcing Shahin to cover an even greater distance through the rugged terrain of western Armenia and more quickly, resulting in his army being even more disorganised and exhausted when they finally reach the Byzantine army. Perhaps Philippicus, with a little bit of luck, manages to crush the Persian army.
Now Khosrau II is in trouble, as his northern flank is exposed. He could get into "house cleaning" mode and demote/banish/execute Shahin for his defeat, which would deprive the Persians of one of their more competent commanders. I could also see him ordering Sharbaraz to redeploy his army to cover Armenia and Media as quickly as possible, especially if Philippicus repeats Heraclius's OTL campaigns in the area to a lesser extent and seems to be poised to strike at Mesopotamia. If Khosrau has been insisting on the speed part, perhaps Sharbaraz might resort to a forced march, which, given the longer distance involved, would most likely end up with his army in worse shape than that of Shahen's, thus allowing Philippicus to score a second victory.
Now Khosrau is in a dire spot, as he has to face two Roman armies while his own army is rather disorganised. Heraclius crosses the Euphrates and starts reconquering the cities and forts the Persians had captured between 604 and 611. Khosrau is forced to agree to a peace treaty that restores the pre-603/604 status quo.
Now, this is the best case scenario for the Byzantines; a more likely one is perhaps the following: Philippicus defeats Shahen, Sharbaraz is recalled and ordered to defend Mesopotamia, Khosrau raises another army to cover the passes from Armenia and Media. Heraclius has some trouble evicting the Persians from northern Syria and fails to cross the Euphrates. Meanwhile, Philippicus fails to do much more in Armenia and is forced to retreat westwards due to logistical troubles. In the end, the two sides conclude a peace treaty (or, more likely, a mere truce), in which the Persians make some territorial gains.
Post war: in the best-case scenario, the Romans have basically recovered their ground. They still suffer from the problems from the reign of Phocas (an empty treasury, divided populations etc) but things are slowly improving. The army's morale has been boosted by these recent victories and Heraclius can proceed with certain military reforms of his more easily. The empire's richest provinces are left unscathed by the war (except for some parts of Syria), which means that restoring the empire's financial health is easier, while trade routes that IOTL were closed by the Persian occupation are still open, meaning that it l will be easier for the general economy to recover. Some minor details of these changes;
- Continued control of Egypt allows Heraclius to resume the distribution of free bread to the population of Constantinople (IOTL, he was forced to levy a special tax of 3 gold coins per recipient in order to import the needed grain in 618 and ultimately ended the whole practice).
- The supply of raw materials, particularly of silk (which was either imported or produced in Syria is continues, meaning that a large part of the empire's manufacturing and trade doesn't enter into steep decline and instead manages to to recover, improving the overall state of the economy
- The import of precious metals, particularly gold, is much easier than OTL, meaning that the emperor doesn't have to debase the silver coins and limit the issuance of solidi
- the aristocracy, particularly the Senate, don't gain ground at the expense of the emperor's power and prestige, as it happened during the mid 610s, meaning that the emperor is in a better position to enact reforms and revitalise the Roman state.
If the Persians are busy doing some house cleaning in the late 610s and early 620s, and then have to face Khazar and Turkic raids, then Heraclius may be able to turn his attention to the Balkan peninsula and return to Maurice's policy of actively engaging the Avars. With a better and more powerful imperial army and much more money at his disposal, he may be able to create trouble for the Avars, inciting more Slavic tribes to revolt and pressing both Slavs and Avars back to the Danube. Sooner or later, the Avars would find themselves under growing pressure, as the Empire would continue to foment unrest among their vassals and undermine their influence. They might decide to try to strongarm the empire to back down, but, given the better position of the empire ITTL, such an attempt would most likely end in disaster for the Khanate. Therefore, by the mid 620s, the empire is in a rather good state for the first time in almost 20 years, with one of its two major foes neutralised and the other busy elsewhere.
In the more realistic scenario, the situation is more akin to what happened during the reigns of Justin II and Tiberius II, with constant low-level warfare between the two sides happening a few years after the truce, as both empires would have to face trouble elsewhere (the Romans in the Balkans and secondarily in Italy and North Africa, the Persians in the Caucasus and Khorasan). However, the empire is still in a much better place than IOTL.
