# Windchill AI Assistant Sizing and Value Calculator

> Source: https://netray.co/tools/windchill-ai-assistant-sizing-and-value-calculator
> Publisher: Netray (netray.co), AI and integration engineering for ERP and manufacturing systems.
> Section: Free Tools

## Short answer

Size GPUs and test the business case for a private AI assistant on Windchill using your users, questions per day and document volume.

## Category

engineering-plm

## Erp

PTC Windchill

## Audience

- PLM administrator
- IT architect
- CIO

## Tool

**Kind:** calculator

**Currency:** $

**Inputs:**

#### Named users

**Key:** users

**Type:** number

**Default:** 250

**Min:** 1

**Step:** 10

**Unit:** users

#### Daily active share

**Key:** active_pct

**Type:** slider

**Default:** 25

**Min:** 5

**Max:** 100

**Step:** 5

**Unit:** %

#### Questions per active user per day

**Key:** q_per_user

**Type:** number

**Default:** 6

**Min:** 1

**Step:** 1

#### Prompt plus answer tokens per question

**Key:** tokens_per_q

**Type:** number

**Default:** 3000

**Min:** 200

**Step:** 100

#### Documents to index

**Key:** docs

**Type:** number

**Default:** 180000

**Min:** 0

**Step:** 10000

#### Average pages per document

**Key:** pages_per_doc

**Type:** number

**Default:** 5

**Min:** 1

**Step:** 1

#### Minutes saved per question

**Key:** min_saved

**Type:** number

**Default:** 5

**Min:** 0

**Step:** 1

#### Loaded hourly cost

**Key:** hourly_cost

**Type:** number

**Default:** 70

**Min:** 0

**Step:** 5

**Unit:** $/h

#### GPU monthly cost

**Key:** gpu_cost

**Type:** number

**Default:** 2000

**Min:** 0

**Step:** 100

**Unit:** $/GPU/month

#### Aggregate tokens per second per GPU

**Key:** gpu_tps

**Type:** number

**Default:** 1000

**Min:** 100

**Step:** 100

#### GPUs needed just to hold the model

**Key:** min_gpus

**Type:** number

**Default:** 1

**Min:** 1

**Step:** 1

#### One-off integration cost

**Key:** integ_cost

**Type:** number

**Default:** 45000

**Min:** 0

**Step:** 5000

**Unit:** $

**Outputs:**

#### Daily active users

**Key:** active_users

**Expr:** users * active_pct / 100

**Format:** integer

#### Questions per day

**Key:** q_day

**Expr:** active_users * q_per_user

**Format:** integer

#### Tokens per day

**Key:** tokens_day

**Expr:** q_day * tokens_per_q

**Format:** number

#### Peak concurrent requests

**Key:** peak_conc

**Expr:** q_day * 0.25 / 3600 * 20

**Format:** decimal

**Description:** 25% of daily questions in the busiest hour, about 20 seconds per answer.

#### GPU estimate

**Key:** gpus

**Expr:** max(min_gpus, ceil(peak_conc * 25 / gpu_tps))

**Format:** integer

**Primary:** true

**Description:** About 25 tokens per second per concurrent stream, with a model memory floor.

#### Vector index size

**Key:** idx_gb

**Expr:** docs * pages_per_doc * 2 * 6000 / 1000000000

**Format:** decimal

**Description:** 2 chunks per page, about 6 KB per chunk with vector and metadata.

#### Hours saved per month

**Key:** hours_month

**Expr:** q_day * min_saved / 60 * 21

**Format:** hours

#### Monthly value

**Key:** value_month

**Expr:** hours_month * hourly_cost

**Format:** currency

**Primary:** true

#### Monthly GPU cost

**Key:** cost_month

**Expr:** gpus * gpu_cost

**Format:** currency

#### Simple payback

**Key:** payback_months

**Expr:** if(gt(value_month - cost_month, 0), integ_cost / (value_month - cost_month), 0)

**Format:** months

**Chart:**

#### GPU estimate by questions per active user

**Type:** bar

**X:** q_per_user

**From:** 2

**To:** 60

**Steps:** 15

**Series:**

- gpus

**Disclaimer:** Planning estimate. Replace defaults with pilot measurements, especially throughput and minutes saved.

## Content

**Intro:** This calculator sizes the GPUs and index for a private AI assistant on Windchill and compares monthly value with monthly hardware cost. Defaults suit a mid-sized engineering group.

**Sections:**

#### How to use it

**Body:** Enter named users and the share who use it daily, then questions per user per day. Use the tokens per question your pilot logs report. Set GPU cost to the lease or depreciation per GPU per month. The chart sweeps questions per active user to show where a second or third GPU appears.

#### Formulas and assumptions

**Body:** Questions per day equal users times active share times questions per user. Peak concurrency assumes 25 percent of questions fall in the busiest hour and an answer takes about 20 seconds. GPUs are the larger of the model floor and peak concurrency times 25 tokens per second divided by per-GPU throughput.

**Bullets:**

- Index GB = documents x pages x 2 chunks x 6 KB
- Hours saved = questions x minutes saved / 60 x 21
- Payback = one-off cost / net monthly value

#### Reading the result

**Body:** A flat GPU count across a range of usage means memory, not throughput, is your constraint. Plan redundancy separately: one GPU has no failover. If monthly value is below GPU cost, the case rests on quality or risk benefits, not time saved.

#### Common mistakes

**Body:** Treating named users as daily users, leaving out re-indexing and embedding load, and forgetting that long retrieved context multiplies tokens per question.

**Cta:** Netray can sanity-check these numbers against your Windchill data volumes and run a measured pilot. The calculator needs no contact.

**Faq:**

#### Why is there a model memory floor?

**Answer:** A model must fit in GPU memory with room for context. A 70B class model in 8 bit needs far more memory than one card offers, so the floor can exceed what throughput alone would require.

#### How large will the index be?

**Answer:** For hundreds of thousands of documents, usually tens of gigabytes. Embedding dimension, chunk size and metadata change this, so treat the output as an order of magnitude.

#### Does the estimate include redundancy?

**Answer:** No. Add a spare GPU or a second node if downtime is unacceptable, and include it in the GPU monthly cost.

#### Can I trust the payback figure?

**Answer:** Treat it as a conditional statement: if minutes saved hold, payback is about this long. Verify minutes saved with timed tasks in a pilot.

## Search terms

Windchill AI sizing, Windchill AI ROI, on-prem LLM sizing calculator, PLM assistant payback, RAG index size calculator, GPU count estimate LLM

---
Talk to an engineer: https://netray.co/contact?intent=erp_ai | Machine catalog: https://netray.co/agent/catalog.json