Infor VISUAL AI Assistant Sizing and Value Calculator
This calculator sizes the hardware and estimates the value of a private AI assistant running on your own servers in front of Infor VISUAL. Enter user numbers, usage and your own assumptions to get GPU count, index size and payback.
Your numbers
Share of named users who ask at least one question per day.
Includes retrieved context. Document questions run higher than record lookups.
Assumption: depends on model size, quantisation and GPU. Measure on your hardware.
Amortised hardware, power and hosting.
Your results
Planning estimate only. Tokens per second, minutes saved and active share are assumptions you must replace with measured or surveyed numbers.
Monthly value vs GPU cost by named users
Sweeps Named users from 50 to 2000, holding your other inputs fixed.
Validate my Infor VISUAL sizing
Optional: share your numbers and get a short review of GPU count, index size and pilot plan.
Optional. The tool and every export work without this. No spam.
How to use it
Start with named users and the share who ask a question on a given day, then the number of questions and the average tokens for a prompt plus answer. Enter the documents you want searchable and the minutes a good answer saves. The outputs update live, and the chart sweeps named users.
- Use a pilot or survey to set the active share; 10 to 25 percent is common for early rollouts.
- Document questions use more tokens than record lookups, so raise the average if retrieval is heavy.
- Replace the tokens-per-second default with a measured figure for your model and GPU.
The formulas
Questions per day equals named users times active share times questions per user. Tokens per day is questions times tokens per question. Peak tokens per second is busiest-hour questions divided by 3,600 times tokens per question; GPUs are that figure divided by tokens per second per GPU, rounded up, plus spares. Index size is documents times pages times 3 chunks times 1,024 floats at 4 bytes with 50 percent overhead. Monthly value is questions times working days times minutes saved divided by 60, times loaded hourly cost.
How to read the result
If monthly value is well above GPU cost, payback is short and the sizing is not the constraint. If it is close, the assumptions on minutes saved and active share matter most; test them in a pilot before buying hardware. Index size is usually small compared with the model weights; the GPU count is driven by peak concurrency, not by user count alone.
Common mistakes and Infor VISUAL notes
Typical errors: using average load instead of the busiest hour, ignoring spares, and counting every user as active. For Infor VISUAL the data comes from your own system and documents, so no per-seat cloud fee applies; hardware, power, setup and people are the cost. Vendor tools may be enough if you only need standard reports.
- Assume nothing about throughput until you measure.
- Keep a spare GPU or a failover plan for working hours.
- Check licence terms of the open-weight model you choose.
Frequently Asked Questions
What tokens per second should I use?
The default is a planning figure for batched serving of a mid-size open-weight model on one modern GPU. Real throughput depends on model, quantisation, context length and software, so benchmark vLLM or Ollama on your hardware.
Why does peak matter more than daily volume?
GPUs must serve the busiest hour, not the daily average. If a quarter of questions land in one hour, capacity must cover that burst or users wait.
Is the value figure reliable?
It is only as good as minutes saved per question and the active share. Use a small pilot with timed tasks to set them, then treat the result as a range.
Does the index size include the model?
No. Index size covers vectors for your documents. Model weights add roughly several GB to tens of GB per model depending on size and quantisation.
If you want these numbers validated against your Infor VISUAL data volumes and a pilot, Netray can help you measure real usage and size the hardware.
Related Tools
Infor VISUAL Private AI Layer Blueprint Generator
Fill in your Infor VISUAL environment and get a ready-to-share architecture plan for a private, on-prem AI assistant.
On-Prem AIGPU Sizing Calculator for LLM Inference
Work out how many GPUs you need to serve a given open-weight model to your user base, based on memory footprint and token throughput.
On-Prem AIOn-Prem AI ROI Calculator
Turn hours saved per employee into annual net benefit, payback months, and 3-year ROI for an on-prem AI investment.
On-Prem AIOn-Prem AI Deployment Checklist
A 30-point pre-deployment checklist covering use cases, hardware, security, model operations, and rollout for self-hosted enterprise LLMs.
More Other ERPs tools
See all- Infor VISUAL Work Order Status and Structure Reference
- Infor VISUAL Manufacturing Scheduling Checklist
- Infor XA (MAPICS) Period End Close Checklist
- Infor VISUAL Database Table Reference
- Infor XA (MAPICS) Application Module Reference
- IBM i Commands for Infor XA Administrators Reference
- Acumatica Generic Inquiry Design Template
- Acumatica Manufacturing Edition Setup Checklist
- Sage X3 Import Template Worksheet
Keep exploring
Free tools
- Infor VISUAL Private AI Layer Blueprint GeneratorFill in your Infor VISUAL environment and get a ready-to-share architecture plan for a private, on-prem AI assistant.
- Infor XA AI Assistant Sizing and Value CalculatorSize the GPUs and estimate the monthly value and payback of an on-prem AI assistant for Infor XA users.
- Infor VISUAL Manufacturing Scheduling ChecklistA practical checklist for finding why VISUAL schedules do not match the floor: calendars, resource data, routings, lead times and release habits.
ERP fixes and how-tos
- Private AI Layer for Infor VISUAL On-Prem: Reference ArchitectureReference architecture to add a private AI assistant to Infor VISUAL on-prem: data access paths, retrieval, permissions, sizing and rollout, no cloud needed.
- Infor VISUAL system requirements: server, SQL Server, client and networkWhat to check before installing or upgrading Infor VISUAL: Windows and SQL Server support, client and ODBC needs, network and sizing, and where to get exact specs.
- Infor VISUAL report builder: build a custom work order reportHow to build and customise reports for Infor VISUAL using Crystal Reports and SQL, with a worked work order report example and the common tables involved.
Runs with your ERP
- 1 EDI Source EDI Integration with Infor VISUALHow 1 EDI Source connects Infor VISUAL to trading partners for EDI orders, invoices and ship notices via ION or direct API: setup, pitfalls and private AI.
- Infor VISUAL Shop Floor Mobile for Labor TrackingHow Infor VISUAL Shop Floor Mobile captures clock-ins and labor data from shop-floor devices directly into VISUAL: data flow, setup, pitfalls and private AI.
- VISUAL Time & Attendance and Infor VISUAL: labor, job costing and payrollHow VISUAL Time & Attendance captures clock-ins and job labor, feeds VISUAL job costing and exports earnings to payroll, with setup, pitfalls and AI ideas.
Integrations
- How to use ChatGPT with Infor VISUALHow to use ChatGPT with Infor VISUAL data: exports, custom GPT actions over a small API, enterprise data controls, and when a private model is the better fit.
- How to add an AI chatbot to Infor VISUALHow to build an AI chatbot on Infor VISUAL data for planners, buyers and customers, grounded in the SQL Server database and able to run fully on-prem.