And lately I’m seeing roles asking for ‘AI-driven interpretation’ with 48‑hour turnarounds on 3D, but nothing about data conditioning, well ties, or phase check. On a 220 km² PSDM I just finished in Petrel, horizon picking took five days after two days of wavelet/phase QC to avoid spurious amplitude leads. Are others pushing back on these timelines in interviews, and is it costing offers?
In interviews I scope it as two deliverables: a ‘48-hour’ AI first-pass on the 3D, then a gated phase/wavelet QC + well-tie cycle (3–5 days on a 200+ km² PSDM) before any amplitude-driven decisions. If they push back, I point to SEG’s interpretation QC notes (https://wiki.seg.org/wiki/Seismic_interpretation) and offer a small pilot; AI can sprint, but you still need to tie your shoes. I only compress when wells are dense and phase is already verified.
On my last 3D, I got buy-in by carving out a non‑negotiable 4‑hour micro‑QC: a quick phase‑rotation sweep and one blind well tie before I show any amplitude. It catches polarity/wavelet mismatches fast, and if they still want the 48‑hour drop, I stamp the maps “amplitude provisional” and keep the first pass strictly structural, @OP.