AI Transformation Lab
Glimmer Bot Glints AI agent

Meet Glimmer. The AI agent we run our company on.

It works with your company's data in plain language, answers your team's questions with the source shown, and builds the tools you ask for. Glints achieved 4x output with Glimmer.

What it is

An AI teammate that knows your business.

It sits in the group chat your team already opens every morning, so there's nothing new to install. What makes it worth having is what it's been taught.

It knows your business, not the internet

Taught on your handbook, your policies and your own data. That's why it quotes your leave policy, not a general answer off the web.

Every answer shows its source

It names where the number or the rule came from, and tells you when it can't confirm something. You can check it.

One good question becomes everyone's

Asked in the shared chat, answered in the open. The team copies each other's best asks instead of each starting from zero.

#team-ops · 14 members
HM
HR manager
@Glimmer which industries get the most applicants per job in Indonesia?
Glimmer Bot
Here's the ranked table — with the data source and how each number is defined.Ranked table · source stated
SL
Sales lead
Stealing this for my pitch deck 🙏👍 4 · ⚡ 2

Condensed from a real thread in our workspace — read the originals.

What it can do

Three things your team will use it for on day one.

Each of these is something Glimmer already does inside Glints every day. The exchanges below are condensed from real business asks and cases, and each row links to a real thread.

01 Analyst AI

Your numbers analysed, without chasing anyone

Ask in plain English and get real numbers and analysis back in minutes. Our CEO asked why revenue was dropping for a client; Glimmer pulled the account's full history, found the root cause, and drafted a ready-to-send reply.

Read the actual thread
Ask Glimmer
CEO
CEO
@Glimmer this client says our applicant quality is poor and turned down the renewal. Can you pull up their account?
Glimmer Bot
Their sales roles rejected 75–100% of applicants — the requirements are set stricter than the market. Job-by-job stats below, and here's a reply you can send them.Draft reply attached
02 Operator AI

Your team's questions, off your plate

It answers from your own documents: handbook, SOPs on everything from price lists to warranty terms. Your team no longer waits for you before operating. No hallucination: every answer shows its source.

Read the actual thread
Ask Glimmer
HM
HR manager
@Glimmer for our Taiwan employees — what happens to unused annual leave at year-end?
Glimmer Bot
It can be carried forward by agreement, and must be paid out on termination. One thing I can't confirm from here: your contract's own carry-over clause — worth checking.Taiwan Labor Standards Act, Art. 38
03 Builder AI

Your own tools, without hiring a software company

Need a quote calculator, a stock tracker, a new page on your website? Normally that means hiring a software company: get a quote, wait weeks, pay again for every change. With Glimmer, you describe what you need in plain English and it's built the same day. Even this website was built this way, not by an agency.

Read a real build thread
Ask Glimmer
ML
Marketing lead
@Glimmer the new workshop date is confirmed. Can you update the website? New date, new price, and put it live today.
Glimmer Bot
Done. The page shows the new date and price. It's live on the website now, here's the link.Live · same afternoon
Proven inside Glints first

We didn't build this to sell it. We built it to run Glints.

Before offering Glimmer to anyone else, we ran our own company on it: the CEO's account reviews, HR's policy questions, finance data pulls, new product features. What we deploy for you is the same software, with the lessons already learned.

AI Agents in Glints

Not a demo — the actual threads from our workspace.

The agent in these threads is Glimmer — the same software we run our own company on, and build and deploy for clients.

Same agent · non-engineers building the tools

And it doesn't stop at analysis — it puts building in the hands of the people closest to the work, not just engineers. Below, a business leader describes a tool he needs in plain English:

Dated, on purpose. On 20 August 2026, Slack announced a new kind of channel where someone who doesn't write code can ask an AI for a feature and the team ships it together. The thread below is us doing exactly that in our own chat on 19 May, three months earlier, on the recruitment software our own team uses every day.

Ask Glimmer
Thread
ZP
Zheng Peng LimMay 19, 1:35 PM

@Glimmer — for the Superpowered recruitment platform: when we find a good candidate today, the only option also recommends them to the client straight away. I'd like a way to just add them to a job's shortlist first — marked as sourced — without recommending them yet.

ZP
Zheng Peng LimMay 19, 1:35 PM

Same as what we do today, just stopping one step earlier. cc: @Francis Goh help pls 🙏

GlimmerBotGlints AI agentMay 19, 1:38 PM

Done. Candidates can now be added to a job's shortlist as sourced without being recommended to the client. If someone's already on the shortlist it leaves them as-is, and it works the same way as the rest of the platform.

ZP
Zheng Peng LimMay 19, 1:40 PM

Does this mean it's done LOL what do I do next?? @Francis Goh @Yong Jie Wong send help 😅

FG
Francis GohMay 19, 1:43 PM

Code's done — it just needs review + deploy. @Glimmer create an MR.

Ask Glimmer continued
Thread
GlimmerBotGlints AI agentMay 19, 1:45 PM

Merge request opened — the new feature plus tests, ready for review.

Yong Jie WongMay 19, 1:46 PM

@Glimmer babysit the MR.

GlimmerBotGlints AI agentMay 19, 1:53 PM

A check failed on the way through — I fixed it and re-ran. all checks green

Yong Jie WongMay 19, 2:58 PM

Merged into production 🎉 deploy in ~15 min. Thanks @Glimmer — pushing to prod!!

ZP
Zheng Peng LimMay 19, 2:58 PM
thanks!! 🙌

Operations analysis, on demand. A product lead briefs it out; Glimmer works through 281k signups, isolates the root cause, and quantifies the fix.

Ask Glimmer Operations · BI
Thread
JT
Jenny TjhioApr 10, 11:50 AM

@Glimmer analyse the cohort of ID candidates who signed up but never applied — root causes, behaviour patterns, and product fixes. Run the full funnel breakdown.

GlimmerBotGlints AI agentApr 10, 11:53 AM

Analysed 281,779 signups. 54.4% never applied — despite clear intent (83% set job preferences, 100% "actively looking"). The drop-off is one step:

Funnel stepAppliedNever appliedGap
Has resume87.8%4.6%83pp
Has experience75.0%11.9%63pp

Blocker = resume-upload friction, worsened by marketplace oversupply (200–300:1 in admin/data-entry roles). Recommended quick win: 1-click apply without a resume for entry-level jobs.

GlimmerBotGlints AI agentApr 10, 11:53 AM

Opportunity: converting just 15–20% of the 32,718 active non-appliers ≈ 5,000–6,500 extra applications / month from this cohort alone.

An account review from one complaint. The CEO forwards a client gripe; Glimmer validates it against the data and even drafts the customer reply.

Ask Glimmer Sales · Support
Thread
DH
Dylan HuangMay 5, 8:17 PM

@Glimmer a client — XXXXXXXX — says applicant quality was poor across their sales & account-management roles, free and paid alike. Pull their account performance.

GlimmerBotGlints AI agentMay 5, 8:18 PM

Complaint validated verified client. Their 2025 sales roles each drew only:

2025 sales roleApplicantsRejected
Spesialis IT Sales2075%
Sales Specialist7100%

Root cause: strict 3–5 yr + degree filters on niche roles, and no paid promotion ever used to widen reach. All 31 jobs now closed.

GlimmerBotGlints AI agentMay 5, 8:18 PM
✍️ Drafted customer reply
"Thank you for the feedback. Your recent sales roles received 7–20 applicants each, with many below the 3–5 yr IT-sales experience bar. To lift quality we'd recommend Hot Job promotion, CV Finder for niche roles, and easing screening to widen the qualified pool…"
Real thread from our Glints workspace — client name redacted, everything else is real.

Policy answers, straight from the source. An HR lead checks a statutory-leave rule; Glimmer cites the actual labour law and our internal wiki — and is honest about what it can't confirm.

Ask Glimmer Knowledge base · HR
Thread
TP
Tiara Sevrilia PutriJun 11, 2:59 PM

@Glimmer for our Taiwan-based employees — how is unused statutory annual leave treated? Is it carried forward to the next year, then paid out if still unused by year-end or on termination?

GlimmerBotGlints AI agentJun 11, 3:01 PM

Correct — with one condition. Under the Taiwan Labor Standards Act, Article 38, unused statutory annual leave may be carried to the following year only if the employer and worker agree. If it's still unused by the end of that year — or on termination — it must be paid out.

Cited sources
Taiwan Labor Standards Act, Art. 38 (law.moj.gov.tw) · Glints internal HR wiki — leave-request policy
GlimmerBotGlints AI agentJun 11, 3:01 PM

One caveat: I couldn't confirm from our wiki that the Glints TW contract spells out carry-forward — worth checking the local HRMS policy. General guidance; verify with Legal for specific cases.

Self-serve BI, in plain English. No SQL and no analyst queue — a question becomes a sourced, ranked answer in seconds.

Ask Glimmer BI · Data
Thread
YP
Yanto Hans PratamaJun 8, 11:30 AM

@Glimmer what's the average applicants per job by industry in Indonesia — logistics, construction, IT services, financial services, outsourcing?

GlimmerBotGlints AI agentJun 8, 11:33 AM

Indonesia · jobs created in the last 6 months. Applicants ÷ jobs, by industry:

IndustryJobsAvg / job
Logistics & Supply Chain7,828229.0
Construction7,928207.2
IT & Services17,661181.5
Financial Services7,482173.7
Outsourcing18,684157.4

Logistics and construction pull the most applicants per role; outsourcing has the most jobs (18.7k) but the lowest pull. source · production warehouse

Triaging user reports, instantly. Is the new release stable? Glimmer reads 48 hours of in-app feedback and pinpoints exactly where the problem lives.

Ask Glimmer Support · Product
Thread
WY
Weisheng YipJun 27, 4:37 PM

@Glimmer check the last 48h of user feedback on the new 1.109.1 release — any freeze or hang reports? We'd had some on 1.109.0 on Infinix devices.

GlimmerBotGlints AI agentJun 27, 4:39 PM

Zero freeze/hang reports on 1.109.1. Pulled every bug report in the window — 5 total, all unrelated (apply-flow and profile edits):

DeviceReportFreeze?
OPPO CPH2269Phone-number changeNo
realme RMX2151Can't update CVNo
Xiaomi 23124RA7EOApply needs 3 tapsNo

The only lag complaints are on the older 1.108.4 (including one Infinix). The 1.109.0 freeze issue did not carry into 1.109.1. release looks clean

How a deployment works

Up and running in three steps.

Step 01

We teach it your business

We set Glimmer up on your handbook, your policies and your data, so its answers are about your company, not the internet's.

Step 02

It joins your team chat

One shared deployment in the chat tool your staff already use, managed centrally. Your data stays yours, and so does what we build for you.

Step 03

You see the evidence

What people asked, what got answered, and the time saved. You can see whether it's working, not take our word for it.

“Do we have to replace our POS or accounting software?”

No. Glimmer layers over what you already run — an old desktop accounting package, a POS built just for you, spreadsheets everywhere. Nothing gets ripped out, and nobody has to relearn their job. We confirm the fit for your specific systems during scoping.

A typical deployment: Glimmer sits between your team chat and the systems you already use — nothing above it changes, nothing below it gets replaced.

“What if the AI provider raises prices or changes the rules?”

Then we switch the engine underneath. Glimmer isn't tied to one AI company — the leading models keep leapfrogging each other, and your deployment can follow the best one. Your data, your setup and your team's habits all carry over. A durable investment, not a bet on one vendor.
Your Glimmer deploymentStays the same
Your company data
Your team chat & history
Your team's habits
Engine todayOpenAI
Engine todayClaude
Engine todayGemini
Engine todayDeepSeek
Engine todayQwen
Engine todayZhipu

Engines shown for illustration — we pick what fits your deployment, and can swap later without starting over.Logos belong to their respective owners.

“Why not just buy everyone a ChatGPT account?”

Because ten separate subscriptions means ten copies of a tool that knows nothing about you. None of them has read your handbook or seen your sales figures, so ten people spend their day pasting the same background in by hand. At the end of the month there's no way to see what the company got for the money. One shared deployment is connected to your business once, works in the open so the team copies the best asks, and usage is measurable.

“Slack and Teams are adding their own AI. Why not just use that?

Do it. It's a good sign you're on the right track, and it makes our job easier.

What they give you is the room: a channel with an AI sitting in it. What they don't give you is an AI that knows your business. Out of the box it knows what your staff have typed in the chat, and nothing else. It hasn't read your handbook, it can't see your sales figures, and it's never heard of the system you've run for the last ten years. Connecting it up is the work, and it's what decides whether anyone still uses it in month three.

Glimmer also isn't tied to one chat company. Whatever your team is on today (Slack, Microsoft Teams, Telegram, Google Chat), it joins there, and if you move next year it moves with you. If you don't have a shared chat tool yet, we help you pick one that fits during scoping.
Today Slack
Today Teams
Today Telegram
Today Google Chat
Glimmer
Same data, same answers

Chat tools come and go, and each one is adding its own AI. Glimmer is the part that knows your business, so it joins whichever app your team is on and moves with you to the next one.Logos belong to their respective owners.

Questions leaders ask

The honest answers, before you book a call.

Do we have to replace our POS or accounting software?

No. Glimmer layers over what you already run — an old desktop accounting package, a POS built just for you, spreadsheets everywhere. Nothing gets ripped out, and nobody has to relearn their job. We confirm the fit for your specific systems during scoping.

What if the AI provider raises prices or changes the rules?

Then we switch the engine underneath. Glimmer isn't tied to one AI company — your data, your setup and your team's habits all carry over. A durable investment, not a bet on one vendor.

Why not just buy everyone a ChatGPT account?

Because ten separate subscriptions means ten copies of a tool that knows nothing about your business. None of them has read your handbook or seen your sales figures, and there's no way to see what the company got for the money. One shared deployment is connected to your business once, works in the open so the whole team copies the best asks, and usage is measurable.

Slack and Teams are adding their own AI. Why not just use that?

Do it. It's a good sign you're on the right track. What they give you is the room: a channel with an AI in it. What they don't give you is an AI that knows your business. It hasn't read your handbook, can't see your sales figures, and has never heard of the system you've run for ten years. Connecting it up is the work, and it's what decides whether anyone still uses it in month three. Glimmer isn't tied to one chat company either, so whatever your team is on today, it joins there and moves with you if you switch. If you don't have a shared chat tool yet, we help you pick one during scoping.

What about our data and confidentiality?

Your data stays yours, deployed on your infrastructure or private infrastructure we set up for you. We cover exactly where it goes during scoping, and nothing is shared beyond your deployment. We don't sit in the middle holding your data, and what we build for you belongs to you.

What does it cost?

It depends on the size of the build, so we price it after scoping your workflow on a call. Most builds qualify for government grants like EDG: the government pays part, you pay the rest, and we help you with the application. You'll know the full numbers before any build starts.

I'd rather try it hands-on before committing to a build — can I?

Yes. Our leadership workshop puts an AI agent in your hands for an afternoon, no coding, seats from S$779 — a lower-commitment way to feel it work before you deploy one company-wide.

Can we run a private session for just our company?

Yes. If you'd rather bring your own leadership team or staff into a closed session, we'll tailor it to your business. Talk to us to scope it.