Anthology Student consulting
Hands-on help with Anthology Student (formerly CampusNexus and CampusVue) from someone who has worked in it since 2007, first as a user and then building on it.
Applied AI and agent automation for higher education: AI that does real work, with a paper trail
I put AI agents to work, under controls I build, on real student systems with real compliance stakes. Not a chatbot bolted on, and not a pilot that never leaves the lab.
The problem
Many schools are being told to “do something with AI” and are right to be wary. The risks are specific: an agent that changes a record it should not, a confident answer that is wrong, student data sent somewhere it should not go, and no way to show an auditor what happened.
Those risks are manageable. They are engineering problems, and they have engineering answers.
What I put in place
- Agents put to work on operational tasks. The reading, comparing and checking that swamps a small team. Reconciling records between systems, triaging support cases, validating a migration, drafting reports for review.
- Plain code where the answer must never vary. A step that has to give the same result on every run is written as ordinary code. The agent runs it and explains what it found; it does not improvise it.
- The controls around them. Rules that are enforced automatically, not requested politely. What an agent may read, what it may never write, and when it must stop and ask.
- Independent verification. A second agent that re-checks the first one’s results against the source systems and produces the audit trail, in place of a person doing spot checks.
- Receipts. A record of every action an agent took, so its work can be reviewed afterward.
- Tooling for your own team. I have built a suite of agent skills that I use daily for case triage, sprint reporting, API discovery, environment smoke tests and pre-flight code checks. I can set your technical staff up the same way.
Where to start
Pick a task that is tedious, frequent and easy to check. A reconciliation. A pre-flight check before a deployment. A weekly exception report. I will tell you honestly if a task is a poor fit, including when plain code would do the job better than AI.
What it will not do
It will not make decisions about students. It will not change production data: a person approves and runs every change. It will not be given data we have not agreed to in writing. Those limits are in the design from the first day.
Common questions
Questions people ask
How is AI used on our project, and can we say no?
AI agents speed up the reading, comparing and checking around the work: studying your systems, drafting code and reports for review, and running the same checks on every record. Whether an AI tool sees your records at all, and which ones, is your decision, made in writing before work starts. Where the answer is no, the agent works on structure, rules, code and counts, and the record-by-record work is ordinary code inside your environment. More on the approach page.
Will an AI agent make changes to our production system?
Not directly. Where you have agreed to it, an agent reads a live system, and the tools it is given there are read tools. Changes are rehearsed in test and applied by a person. The full method is on the approach page.
Can you train our staff?
Every handoff includes a walkthrough for your team along with the documentation and runbooks, and I build reports and validator checks that your staff run for themselves rather than tools only I can use. For technical staff, I can set them up with the same agent tooling I use for my own work. If your team needs more than a walkthrough, say so on the first call. See applied AI.
Other services
SIS integration
Connect your student system to everything around it, as changes happen, with a record of every change.
Reporting and dashboards
SSRS reports and Power BI dashboards your staff trust, including Crystal Reports rebuilt for the web client, by someone who has used those reports.
Data migration
Move student and financial aid data between systems with a rehearsal first and proof for every record.
Financial aid automation
Automation around financial aid data, built by a former aid administrator, with the evidence an auditor asks for.
Workflow automation
Take the manual steps out of admissions, registrar, aid and finance work with Workflow Composer and FormsBuilder, inside the system your staff already use.
Managed hosting
Nowhere to run an integration? I host and maintain it on Azure, in your subscription or a dedicated one.
All services
The full list, and the three ways an engagement can run.
Have a system that needs untangling?
Tell me what you run and where it hurts. The first call is 30 minutes, free, and you will leave with a clear next step whether or not we work together.