For years, "AI in ERP" was a vendor slide deck rather than a workflow you could actually deploy. NetSuite's Text Enhance changed that conversation. With generative AI now embedded directly in record-level interactions, finance and operations teams across East Africa are seeing measurable time savings on workflows that were previously the slowest part of the working day.
What Text Enhance actually does
Text Enhance is a context-aware AI assistant baked into NetSuite forms. From a record — a purchase order, a customer, an item, a sales order — it can summarise context, draft narrative fields, generate communication, classify free-text, and translate. It's not a chatbot bolted on the side; it's a button on the field you're already editing.
The five workflows where it pays for itself first
1. Purchase order drafting. Buyers describe what they need in their own words; Text Enhance generates a clean PO description from the vendor and item context. Buyers we work with have cut PO creation time by 40-60% on indirect spend.
2. Customer communication. Statement reminders, invoice cover notes, and follow-ups can be generated on-tone from the customer record, then reviewed before sending. AR clerks who used to draft each email by hand now review and approve.
3. Item description normalisation. Imported product catalogues are notoriously inconsistent. Text Enhance can rewrite SKU descriptions against a house style, which matters when the same item flows to invoice, marketplace, and POS.
4. Vendor matching from OCR. Paired with the OCR pipeline, Text Enhance can resolve fuzzy vendor names ("Kenya Power", "KPLC", "Kenya Power & Lighting Co.") to the right master record on first capture.
5. Field summarisation. Long Notes fields — service records, claims, case histories — can be reduced to a one-line snapshot for dashboards and reports without losing the underlying detail.
What we're seeing in practice
Across HealthActiv, Harleys, and a handful of other early adopters in the INNOVIOR portfolio, the pattern is consistent: Text Enhance does not eliminate steps, it accelerates the slow parts of existing ones. The biggest gains aren't in transactions per hour — they're in reduction of context-switching. A buyer who used to leave NetSuite to draft a PO email no longer does.
"The first week we turned it on for purchase orders, our procurement team picked up 30 hours back. By month two, the auditors stopped flagging item description inconsistencies. We didn't expect that to be the bigger win." — Finance Director, INNOVIOR client (pharma distribution)
Getting it right
Three things separate the implementations that stick from the ones that get switched off after two months:
- Define a house style first. Generative AI without a style guide produces fluent but inconsistent text. Lock down tone, abbreviations, and required fields before rollout.
- Audit, don't auto-post. Keep a human in the loop for outbound customer communication at least until you have 4-6 weeks of validation data.
- Train on your own data. NetSuite's base model is competent; on-tenant fine-tuning against your own item master and customer correspondence is what makes it feel native.
What it isn't
Text Enhance is not autonomous. It does not initiate transactions, post journal entries, or make procurement decisions. It is a productivity layer that compresses the manual work surrounding decisions humans still own. CFOs who position it that way internally see better adoption than those who pitch it as a replacement.
Where to start
The fastest pilot is one Text Enhance use case in one team for two months. Procurement description drafting is our usual recommendation: low risk, high frequency, easy to measure. Once you have a baseline, expanding to AR communication or item normalisation is straightforward.
If your NetSuite licence already includes AI capabilities, we can scope a Text Enhance pilot with you in a single working session. Reach out at info@innoviorafrica.co.ke or book a demo and we'll walk through what your first 90 days could look like.
