Home › Curriculum › AI across Odoo roles › AI features inside Odoo 20: configuration and adoption
AI in every Odoo role · 14 min · Odoo 20AI features inside Odoo 20: configuration and adoption
Configuring Odoo's AI features, agents and skills for users, budgeting IAP credits and driving adoption.
In this lesson
- What Odoo 20 AI can do and what it needs
- Designing agents around a use case
- Adoption, budgets and measurement
Key points
- Agents read files, create/update records with approval, take dictation; topics → skills
- IAP credits for all AI; edition/plan limits; pgvector on-prem
- One use case per agent: skills, sources, acting user, write approval, reviewer
- Train, budget credits, measure four weeks, document in the SDD
Real-world example
An Abuja distributor's first agent drafted vendor bills from emailed PDFs. AP clerks approved each one; after four weeks the error rate on drafts was 4% and typing time dropped by two-thirds. The second agent (Helpdesk replies) launched with the same playbook.
Questions and answers
Which PostgreSQL extension do on-prem AI features need?
pgvector.
What were 'topics' renamed to in Odoo 20?
Skills.