Pulse · sentiment intelligence

Pulse — know how your customers feel, in real time.

Pulse turns raw feedback — reviews, support tickets, NPS comments, social posts and survey text — into a live pulse score, the themes driving it, and the at-risk accounts you should call today. Paste a stream of feedback below and watch the engine read it, line by line.

Live demo · runs entirely in your browser No data leaves this page · same engine as @artai/services
Interactive analyzer

Paste feedback. Get a pulse.

One feedback item per line — a review, ticket, NPS comment or survey response. We pre-loaded a realistic SaaS mix so you can hit Analyze right away.

Feedback stream

reviews · tickets · NPS · social
Mix the sources freely. Pulse handles negation (“not bad”), intensifiers (“very slow”) and dampeners (“kind of helpful”) — it isn't a keyword count.
0 items

Pulse score

0–100 composite
awaiting input
NPS
Avg score
Items

Distribution

promoters · passives · detractors

Sentiment of every item

What's driving the pulse

theme clustering

What customers love

What to fix

Top themes overall

At-risk · call these today

most negative

Highlights · your champions

most positive

Every item, scored

0 analyzed
How it works

One engine. Client-side here — server-side in production.

1 Ingest 2 Tokenize 3 Score 4 Theme-cluster 5 Pulse

Pulse ingests each feedback line, tokenizes it, and scores sentiment with a real lexicon model — signed cue words, negation handling (a negator flips the next cue), intensifiers (very, extremely) and dampeners (slightly, kind of). Each item gets a score in −1…1, labeled positive / neutral / negative by threshold. It then theme-clusters the corpus against domain keyword groups (pricing, performance, reliability, support, usability, onboarding, features, docs) and rolls everything up into one pulse 0–100 plus an NPS-style number, promoter / passive / detractor split, and ranked at-risk and highlight lists.

The exact same logic runs in production. The @artai/services-api service runs this identical engine server-side over your live feedback streams — review platforms, your help desk, NPS surveys, social and webhooks — so the numbers you see in this booth demo are the numbers you'd get at scale. Promoter when score ≥ 0.4, detractor when score ≤ −0.2; pulse blends mean sentiment with the promoter/detractor skew. Deterministic, explainable, no black box.

Support

Surface the angriest tickets first and route by theme before they escalate.

Product

See which features customers love and which bugs are dragging the pulse down.

CX / CS

Watch account health and get a daily list of at-risk customers to call.

Brand / Social

Track how sentiment moves across reviews and social in near real time.

Markets

Read the mood of a market or ticker from news and chatter, the same way.

One company, three engines.

Pulse reads sentiment, ARTAI reads markets, and the Helpdesk acts on it. Explore the rest of the Smritixs stack — or get pricing and the investor story.