AI Demos

See AI at work on real business questions

Three interactive demos: an operations agent that waits for your approval, a store admin dashboard you can instruct in plain English, and a learning-analytics dashboard.

Simulated demonstrations. They use fictional sample data and are not connected to any real store, LMS, ERP, or AI model.

Demo 1 · Operations agent

An agent that works across your systems

Pick a request and watch the agent call tools, explain what it found, and propose an action that waits for your approval. Every step is written to the audit log.

Simulated · sample data

Which products will sell out before the next delivery arrives?

Tool calls

  • inventory.get_levels{ warehouses: ["TOR", "THB", "VAN"] }→ 182 SKUs across 3 warehouses
  • sales.get_velocity{ days: 30 }→ Average daily units per SKU
  • purchasing.list_open_pos{ status: "open" }→ 12 open purchase orders

6 products will run out before their purchase orders arrive. Here are the three most urgent, based on 30-day sales velocity and current stock across all warehouses:

SKUProductStock leftPO arrives
TRL-2041Trail Pack 32L4 daysin 9 days
CMP-0187Camp Stove Mini6 daysin 11 days
HYD-0932Hydration Bladder 2L8 daysin 14 days

Proposed action: Draft expedite requests for 3 purchase orders

Emails to 2 suppliers asking them to ship PO-4471, PO-4480, and PO-4492 earlier. Requires approval.

Audit log

Every request and tool call is recorded

  1. Run the demo to see entries appear here.

Demo 2 · AI store admin

An online-store dashboard that takes instructions

Ask about orders, stock, reviews, and customers — or tell it what to do: “Add 20 units to all hydration products”, “Show products without images”, “Which products should we add next?”.

  • A fictional outdoor-gear store: 43 products, 60 days of orders
  • Review sentiment, repeat-purchase odds, and product ideas
  • Every change is previewed and waits for your approval
Simulated · sample store

Overview

Outdoor-gear store · Nov 14, 2026 · last 30 days

Store · ERP · Reviews · Search

Net sales

$47,423

+18.7% vs prior 30 days

Orders

401

+25.3% · last 30 days

Avg order value

$118

−5.3% vs prior 30 days

Open orders

89

17 need attention

Low stock

13

3 out of stock

Positive reviews

69%

30 negative to answer

Ask or instruct the store assistant

read · write with approval

What should I focus on today?

Tool calls

  • analytics.saleslast 30 days vs previous 30
  • orders.query89 open orders
  • catalog.audit43 products
  • reviews.sentiment177 reviews
  • customers.score203 customers

Today’s briefing · Nov 14

Net sales were $47,423 over the last 30 days (+18.7% vs the previous 30) from 401 orders.

Here is what needs you today, in priority order:

  • 17 orders need attention — 10 delayed in transit, 3 waiting for stock.
  • 3 products are out of stock and 10 more will run out within two weeks.
  • 30 negative reviews — most about “Durability”.
  • 7 products have no images — they sell 1.8× less than products with photos.
  • Ultralight 1-person tent — 198 searches for “ultralight tent” — only the 2-Person Tent matches.
  • 14 high-value customers (over $300 lifetime) are unlikely to buy again.

Net sales · last 30 days

$0$2.5k$5kOct 16Oct 23Oct 30Nov 6Nov 14
Last 30 days Previous 30 days

Avg order value

$118

Return rate

3%

Positive reviews

69%

Repeat customers

81%

Send me this briefing every morning at 8:00

Requires approval · email to you only

Computed from the simulated store data · Nov 14, 2026 · changes apply to this sandbox only

Net sales

Last 30 days vs previous 30 days

$0$2.5k$5kOct 16Oct 23Oct 30Nov 6Nov 14
Last 30 days Previous 30 days
View as table
Net sales by day
DayNet salesSame day, previous period
Oct 16$798$1,060
Oct 17$1,279$1,396
Oct 18$1,739$1,380
Oct 19$1,036$1,307
Oct 20$1,557$1,222
Oct 21$659$1,369
Oct 22$1,476$1,082
Oct 23$1,762$1,220
Oct 24$2,450$1,056
Oct 25$2,253$1,158
Oct 26$1,364$1,917
Oct 27$1,045$1,508
Oct 28$697$863
Oct 29$1,832$1,536
Oct 30$1,811$1,450
Oct 31$1,499$2,143
Nov 1$1,984$1,142
Nov 2$2,147$2,005
Nov 3$2,138$1,797
Nov 4$1,594$711
Nov 5$1,185$1,345
Nov 6$1,391$1,428
Nov 7$2,159$1,307
Nov 8$2,194$1,300
Nov 9$1,579$1,673
Nov 10$2,735$1,422
Nov 11$1,374$892
Nov 12$710$1,636
Nov 13$1,687$353
Nov 14$1,289$1,288

Orders by status

Last 30 days

  • Pending payment 1
  • Processing 23
  • Awaiting stock 3
  • Shipped 52
  • Delayed 10
  • Delivered 299
  • Returned 13
  • Cancelled 7

Sales by category

Net sales · last 30 days

  • Packs $12,336
  • Apparel $8,861
  • Footwear $8,675
  • Shelter $6,660
  • Camp kitchen $3,417
  • Accessories $2,806
  • Hydration $2,679
  • Lighting $1,989

Low stock

13 products · fewest days left first

  • Trail Runner FTW-2207 · 0 in stock · sells 0.9/day Out of stock
  • Multi-Tool ACC-3677 · 0 in stock · sells 0.5/day Out of stock
  • Soft Flask 500 ml HYD-8077 · 0 in stock · sells 0.4/day Out of stock
  • Folding Spork Set CMP-3546 · 2 in stock · sells 0.6/day 3 days left
  • Camp Sandal FTW-6763 · 1 in stock · sells 0.3/day 3 days left
  • Dry Bag 10L ACC-7669 · 2 in stock · sells 0.5/day 4 days left

Review sentiment

177 reviews · last 60 days · AI-classified

Positive 69%Neutral 14%Negative 17%

Top complaints

  • Durability 12
  • Shipping delays 8
  • Price / value 4

Customer outlook

Likelihood to buy again

  • Likely 98
  • Possible 45
  • Unlikely 60
  • Zoe Walker 16 orders · $2,745 lifetime Likely · 90%
  • Ava Côté 19 orders · $2,598 lifetime Likely · 90%
  • Olivia Clark 25 orders · $2,574 lifetime Likely · 90%

In production, a language model would use scoped tools on your Shopify, Magento, or WooCommerce store and your ERP — reading freely, and writing only what you approve.E-commerce automation

Demo 3 · AI learning analytics

An LMS dashboard you can talk to

Type a question — who needs help, who is excelling, who stopped logging in, how courses compare — and get a detailed report in seconds, with charts, tables, and a CSV export.

  • About 120 fictional students across four courses
  • Reports computed from the data, not pre-written
  • Messages to students always need instructor approval
Simulated · fictional students

Overview

Fall 2026 · week 9 of 14

Students

119

all courses

Active this week

110

0 vs last week

Average grade

67.8%

−0.5 pts this week

On time

86%

of work · weeks 1–9

Need support

18

27 more on watch

Excelling

8

85%+ · nothing missing

Ask the analytics assistant

read-only · LMS + SIS

Who needs help this week?

Tool calls

  • lms.grades.queryall courses · weeks 1–9
  • lms.activity.querylogins · submissions
  • sis.enrolment.lookup119 active enrolments

Students who need support · all courses

18 of 119 students need support this week, ranked by combined risk (grades, missing work, inactivity, and falling marks).

The least-engaged quarter averages −28.1 pts compared with the most-engaged quarter.

StudentStatusGradeTrendMissingLast login
Carter MendesITS 220Needs support36.4%0 pts411d ago
Rosa BouchardHLT 140Needs support42.4%+8 pts39d ago
Marcus OkaforHLT 140Needs support35.8%−1.2 pts214d ago
Aisha SinghBUS 101Needs support43.8%−25.8 pts2today
Wyatt ClarkeENG 110Needs support46.6%+15 pts214d ago
Omar MartinITS 220Needs support36.6%0 pts42d ago
Gabriel ThompsonBUS 101Needs support44.8%0 pts47d ago
Adam MensahHLT 140Needs support44.8%−4 pts33d ago
Lucas WongBUS 101Needs support54.0%−22 pts27d ago
Hunter BianchiITS 220Needs support43.3%+8.8 pts2today
Layla ClarkeENG 110Needs support47.9%+5.2 pts2today
Layla BouchardHLT 140Needs support52.7%+2 pts34d ago

+ 6 more in the CSV export.

Draft check-in messages for 18 students

Instructors review and send each message — nothing goes out automatically.

Computed from the simulated dataset · all courses · week 9 of 14

Grade distribution

Current course grade · all courses

18 A 80–100 34 B 70–79 37 C 60–69 19 D 50–59 11 F 0–49
View as table
Grade distribution
BandRangeStudents
A80–10018
B70–7934
C60–6937
D50–5919
F0–4911

Weekly engagement

Average logins per student, by week

02.55W1W3W5W7W9
All students Needs-support group
View as table
Weekly engagement
WeekAll studentsNeeds-support group
14.22.2
24.32.2
34.32.1
442.2
54.22.3
641.6
73.91.2
84.11.7
94.11.3

Course comparison

Average grade by course

  • BUS 101 70.7%
  • ITS 220 59.9%
  • HLT 140 67.5%
  • ENG 110 72.9%

Needs support

18 students · highest priority first

  • Carter Mendes ITS 220 · Grade below 50% · 4 missing assignments 36.4%
  • Rosa Bouchard HLT 140 · Grade below 50% · 3 missing assignments 42.4%
  • Marcus Okafor HLT 140 · Grade below 50% · 2 missing assignments 35.8%
  • Aisha Singh BUS 101 · Grade below 50% · 2 missing assignments 43.8%
  • Wyatt Clarke ENG 110 · Grade below 50% · 2 missing assignments 46.6%
  • Omar Martin ITS 220 · Grade below 50% · 4 missing assignments 36.6%

Top performers

Highest current grades

  • Sana Kowalski BUS 101 · 98.2% · 100% on time Excelling
  • Carter Brown HLT 140 · 93.0% · 100% on time Excelling
  • Priya Roy BUS 101 · 91.4% · 89% on time Excelling
  • Olivia Ali BUS 101 · 91.1% · 100% on time Watch
  • Kai Kim ENG 110 · 90.4% · 100% on time Excelling

Most improved

Submitted work: last 3 weeks vs weeks 1–5

  • Ava Kim HLT 140 · now 88.3% +24.3 pts
  • Nathan Okafor ITS 220 · now 82.0% +23.8 pts
  • Ruby Scott ITS 220 · now 81.7% +21.9 pts
  • Ethan Scott ITS 220 · now 67.3% +19.5 pts
  • Wyatt Chen HLT 140 · now 92.3% +19.5 pts

In production, a language model with read-only access to your LMS and student information system would generate these reports — within your institution’s privacy and approval rules.AI LMS integration

What the demo shows

Three everyday operations requests

Each scenario follows the same pattern we use in production: read data with narrow tools, explain the result, and propose changes rather than making them.

  • Stockout risk

    “Which products will sell out before the next delivery arrives?”

    Tools used:

    • inventory.get_levels
    • sales.get_velocity
    • purchasing.list_open_pos

    Needs approval: draft expedite requests for 3 purchase orders.

  • Catalog check

    “Which new products aren’t ready to publish?”

    Tools used:

    • catalog.search_products
    • catalog.check_completeness
    • content.draft_description

    Needs approval: create 3 pending product changes.

  • Weekly summary

    “Summarize last week’s sales by channel.”

    Tools used:

    • orders.aggregate
    • inventory.stockout_events

    Needs approval: post this summary to #operations.

From demo to production

What changes in a real deployment

The demo shows the interaction pattern. A production agent adds the security and integration work that makes it safe to use with real business data.

  • Real tools, real data

    In production, each tool call goes to an MCP server or API connected to your store, ERP, or CRM — not to sample data.

  • Sign-in and permissions

    Users sign in with your identity provider. Each person and agent can only use the tools their role allows.

  • Approvals where your team works

    Approve or dismiss proposals in Slack, Microsoft Teams, email, or an internal app — with the approver recorded.

  • A durable audit trail

    Logs are stored in your infrastructure and can be sent to the monitoring tools you already use.

One governed server, many AI clients

Let’s talk

See what this would look like on your systems

Get in touch and we’ll set up a meeting to explore which questions an agent could answer for your team, which systems it would need, and where approvals belong.

  1. 1Get in touch
  2. 2Meeting
  3. 3System analysis
  4. 4Implementation