Legacy load: Medium · Track: AI Adoption
Tax
A practice where the research is voluminous and the position is everything. adnah's read: research assembly is ceremony, the position taken on ambiguous authority is not.
Ceremony, safe to remove
- Manually pulling and organizing the code sections, regulations, rulings, and cases relevant to a recurring fact pattern
- Re-summarizing the same statutory framework in every memo that touches it
- Building transaction step-plans from a blank page when the underlying structure is a known variant
- Cross-checking a return position against prior years' filings by hand
Irreducible, stays with the lawyer
- Judging how aggressive a position can be taken given the client's risk tolerance and the actual state of authority
- Structuring a transaction around a genuinely novel fact pattern that does not map cleanly to precedent
- Advising on the practical odds and consequences of an audit challenge, not just the technical merits
- Negotiating with a taxing authority where judgment about the examiner and the process matters as much as the law
Primitives most in play
The usual failure mode
The typical failure is trusting a research summary as if it were the position itself. Tax research tools can assemble authority quickly, but the actual work of tax practice is deciding how that authority applies to a specific, often unclean fact pattern, and where the client's facts sit relative to the line. Teams that let assembled research stand in for that judgment end up with well-cited memos that answer a slightly different question than the one the client actually asked.
What changes
- Before: an associate spends a day pulling and organizing authority for a memo. After: the authority is assembled and organized automatically, and the associate's day goes to applying it to the client's specific facts.
- Before: transaction step-plans for known structures are redrafted from scratch each time. After: the standard structure is a starting template, and the attorney's work is adapting it to what is actually unusual about this deal.
- Before: comparing a current-year position against prior years is a manual file review. After: the comparison is automatic, and any drift is flagged for a reason before it becomes an inconsistency an examiner finds first.
- Before: client memos explain the law from first principles every time. After: the law is explained once and reused, and the memo's real content is the application to this client's facts and risk tolerance.
What a working transformation looks like here
A working transformation looks like memos that get shorter on background and longer on judgment, because the authority no longer needs to be re-explained from scratch each time. The practical signal is a team spending noticeably more of its billable time on 'here is what we recommend and why' rather than 'here is what the code says,' with position quality holding steady or improving because more attention goes to the facts.
Next practice area
Workers' Compensation →