AI for land surveyors: what it can do in the office, and what stays with the PLS
A plain guide to which survey-office tasks AI can draft today, which must stay with the PLS, and a six-step way to start.
Last checked 13 October 2026
Why this guide exists
Search "AI for land surveyors" and most of what comes up is either a 2023-era chatbot primer or an article about point clouds, drones and field gear. That leaves out where a small firm loses a lot of its week: the office. Reading title commitments, pulling exception documents, typing deed calls, writing proposals.
This guide covers that office work. What AI can draft today, what it should never touch, what the state boards have said, and how to try it on your own files without putting a plat at risk.
Will AI replace land surveyors?
Not the part that carries your seal. The U.S. Bureau of Labor Statistics (BLS) projects surveyor employment to grow 6% from 2025 to 2035, with about 3,100 openings a year, many of them to replace people retiring or leaving the field (BLS Occupational Outlook Handbook, May 2025 data).
Michael Clifford, a professional land surveyor (PLS), argues that boundary surveying is poorly suited to AI because it rests on professional judgment about where ownership lines are (LidarNews, updated Feb 2026). Practitioners on the RPLS.com forum say the same thing in plainer words. One wrote that AI "has been known to dream up data which it thinks is missing." Another said he would not recommend it "for anything requiring an educated/experienced determination" (RPLS.com, Apr 2026).
What AI does change is how much office prep a smaller crew can get through before the truck leaves.
What AI does well in a survey office
These are drafting and sorting jobs. A person still checks every item against the source document.
- Title commitment review. List each Schedule B-II exception (the items the title policy will not cover) with its recording reference, and build a checklist of documents to pull.
- Deed call extraction. Pull bearings, distances, monuments and adjoiners out of a PDF or scan into a call table you can check line by line.
- Closure checks. Report whether a description closes and by how much. Do the math in plain code or your COGO (coordinate geometry) software, not in a chatbot's head. A language model can extract the calls; it should not be trusted to do the arithmetic.
- Research digests. Summarize adjoining deeds, old plats and prior surveys into a short brief for the party chief: what each record says and where records disagree.
- Draft plat notes in your firm's own wording, from your past jobs.
- Proposals and scope letters drafted from the client's request and your similar past jobs.
- Office admin. Safety plans, policies, operating documents.
Some surveyors already do pieces of this. On RPLS.com in April 2026, one practitioner described using Claude to check whether deeds close, Python scripts for level-loop adjustments and cut sheets, and turning Trimble job files into field reports. He planned to try summarizing ALTA title reports into Schedule B reviews next (RPLS.com, Apr 2026). In the same thread, a poster mentioned a Facebook poll on AI use that came back mostly "no". So the profession is split, and careful.
What stays with the licensed surveyor
Keep these off the AI list entirely:
- Weighing boundary evidence
- Deciding which corners to accept
- Resolving conflicts between records
- Final legal descriptions, unless verified line by line against the sources
- Anything you sign and seal
The "dream up data" worry is the right one. Language models fill gaps with confident guesses, and old or badly scanned documents make that worse. One RPLS.com poster pointed out that older, poorly scanned title documents are hard for AI to parse. The fix is not to trust the tool more. It is a process where each AI-drafted item is checked against the source page and the check is written down.
What state boards have said
Three boards have addressed AI so far, and they line up. Texas (Policy Advisory Opinion No. 71, Nov 2024), North Carolina (AI guidelines, Jul 2025) and Idaho, which relayed the NCEES (National Council of Examiners for Engineering and Surveying) Position Statement 6.10 (Jan 2026), all treat AI as an allowed tool (Idaho has not adopted 6.10 into its rules). All three keep the licensee fully responsible. Texas stresses competence and client-data confidentiality; North Carolina and NCEES 6.10 call for verifying outputs and documenting inputs and checks. We found no Utah guidance as of October 2026. This is a summary, not legal advice. The details, and a policy outline, are in our board-rules post.
The tools already exist
You do not need custom software to start. Deed Reader Pro transcribes deeds (including handwritten ones, per the vendor), builds a call table and closure report, and plots into CAD (vendor site, viewed Oct 2026). AECFirmAI's Title Reader classifies Schedule B exceptions from an uploaded commitment and plots easements (vendor site, viewed Oct 2026). The big CAD vendors are putting their AI effort into point clouds and modeling, not record research (GIM International, Nov 2025; Amerisurv, Sep 2025). We compare the options, with prices and weak spots, in the tools post.
How to start without risking a plat
- Pick one task you do every week. Title review on ALTA/NSPS jobs (the American Land Title Association and National Society of Professional Surveyors standard survey) or deed calls on boundary jobs are good first picks.
- Use a business account. Read the data terms and confirm in writing that client and title documents are not used to train models. Texas's opinion specifically says licensees should understand how an AI tool uses and stores data.
- Test on finished jobs. Run the tool on two to ten surveys you have already sealed, where you know the right answer. Count every miss.
- Log the check. For each AI-drafted item, record who verified it against which source page.
- Write a one-page policy so staff know what is allowed. The title workflow post includes a QC checklist you can adapt.
- Measure hours before and after. The cost post walks through the math with real tool prices.
A free starting point
If you want a second set of eyes, Lunan offers a free conversation: we walk through one recent job from request to plat and send back a short written list of which office steps could be drafted by AI and which should stay with your surveyor. Start it here.
Sources
- Surveyors, Occupational Outlook Handbook (U.S. Bureau of Labor Statistics) (May 2025)
- Can AI and Land Surveyors Play Nicely? (LidarNews, Michael Clifford, PLS) (Feb 2026)
- Using AI in your survey workflow (RPLS.com forum thread) (Apr 2026)
- TBPELS Policy Advisory Opinion No. 71 on AI (Nov 2024)
- NCBELS Guidelines on the Use of Artificial Intelligence in Engineering and Surveying Services (Jul 2025)
- Idaho IPELS AI Statement (relaying NCEES Position Statement 6.10) (Jan 2026)
- Deed Reader Pro (viewed Oct 2026)
- AECFirmAI Title Reader (viewed Oct 2026)
- How AI is changing the role of the surveyor in AEC (GIM International) (Nov 2025)
- Carlson 2026 delivers performance and power (Amerisurv) (Sep 2025)