Question 1

What is Klaut, and what is Compass — are they the same thing?

Accepted Answer

No. They are two layers of the same stack.

Compass is the product — a workforce intelligence system that recognises the person behind every piece of work. It observes how your team operates in the systems they already use, corrects mistakes in the moment, generates personalised micro-lessons from observed gaps, and reports development progress through three privacy-bounded views: employee, manager, HR.

Klaut is the company that customises and deploys it. Compass runs standalone; Klaut configures it for your stack, writes the integrations, builds the dashboards your executives will actually read, and remains on call as the deployment matures. The structural reference is Palantir's deployed-engineer model — adapted for mid-market scale.

Question 2

Who is Compass for, and who should not buy it?

Accepted Answer

Compass is for organisations that need AI to deliver productivity without displacing the workforce. That increasingly means: Indonesian enterprises navigating the social cost of layoffs, EU companies under the AI Act's worker-impact obligations, and US companies whose change-resistance comes from organised labour or internal culture rather than from technology.

If your AI strategy is "reduce the team by 30%", Klaut is the wrong firm. We'd rather tell you that on the first call than spend a quarter pretending otherwise. We've structurally designed the product so that revenue does not depend on headcount reduction.

Question 3

We just want an AI chatbot, not a workforce thing. Are we a fit?

Accepted Answer

Probably not, and we'd tell you on the first call.

Plenty of vendors will sell you a chatbot. Compass is for the company that already has — or is about to deploy — AI on top of people, and wants the people to compound in capability as the AI gets smarter. If you only need a customer-facing chatbot, a generic implementation partner will be faster and cheaper than Klaut.

Question 4

We already use Lattice / Culture Amp / Cresta / Observe.AI. What's different?

Accepted Answer

Each of those products owns one slice of the loop. Lattice and Culture Amp handle performance management. Cresta and Observe.AI handle conversation intelligence. LMS platforms handle formal training.

None of them close the loop from observed mistake → personalised correction → micro-lesson → re-assessment. None of them are architected with the privacy-bounded three-buyer alignment (employee owns their mistakes; manager sees aggregate; HR sees role-level capability) — because their existing enterprise sales motion depends on giving managers per-individual visibility, which Compass deliberately does not.

The gap is positional, not feature-level. They are structurally less able to copy the position than to copy any single feature.

Question 5

How does Compass actually work, end to end?

Accepted Answer

Five steps, running continuously.

1. Observe — Compass connects to the work systems your team already uses (helpdesk, CRM, ticket queue) and reads the work product: the ticket, the response, the resolution path. Not screens. Not keystrokes.

2. Correct — when a mistake is identified, Compass intervenes in the moment, privately, in a tone calibrated for development rather than reprimand.

3. Educate — from accumulated correction data, Compass generates 5-to-10-minute micro-lessons targeting that individual's specific failure modes.

4. Assess — periodically, Compass tests whether lessons have been internalised, not via quizzes but by watching real work over time.

5. Report — monthly, three separate views: a personal report card for the employee, an aggregate view for the manager, a role/function view for HR.

The longer the loop runs, the better the personalisation. The longer it runs, the richer the per-role mistake taxonomy Compass develops becomes — and that taxonomy is the proprietary asset competitors cannot easily replicate.

Question 6

Does Compass record screens, capture keystrokes, or read private messages?

Accepted Answer

No, no, and no.

Compass reads the work product through your existing system APIs — the ticket in your helpdesk, the response in your CRM, the conversation in your service platform. We do not install screen recorders. We do not capture keystrokes. We do not have access to private channels (DMs, personal email, anything outside the work systems you explicitly connect).

Surveillance is the opposite of what Compass is engineered for. If your evaluation criteria require monitoring at that depth, we're not the right product.

Question 7

Which work systems does Compass plug into today?

Accepted Answer

Initial release targets the customer-service stack: Zendesk, Freshdesk, Intercom, Salesforce Service Cloud, plus the Indonesian CS platforms our anchor partner operates inside. CRM and ticketing integrations expand from there.

If your stack isn't on the list, the Foundation engagement maps integration cost before you commit to anything — you won't discover a missing connector in week 5.

Question 8

Customer service is your first vertical. When do sales, engineering, operations land?

Accepted Answer

Order is set by where work product is structured enough for Compass to observe reliably — not by where the demand is loudest.

Customer Service is first because performance signals are unambiguous (handle time, resolution, CSAT, escalation rate) and the cultural friction to observing individual work is lowest. MVP target: August 2026. First pilot client: October 2026.

Sales is the natural second vertical (pipeline conversion, deal cycle, win rate — equally clean metrics). Support engineering and operations follow as the per-role mistake taxonomy in those domains matures. Roadmap dates beyond CS we publish only once we have closed a pilot — we'd rather under-promise than over-promise on a multi-vertical roadmap before the first one ships.

Question 9

A better foundation model lands mid-deployment. Then what?

Accepted Answer

We swap it.

Every Compass deployment runs through a provider-abstraction layer — Anthropic today, OpenAI tomorrow, Gemini next quarter, an open-weights model running in your VPC if that's where the eval lands. Switching is a config change, not a rewrite.

The taxonomy of mistakes Compass has learned about your team is the proprietary asset. The underlying model is replaceable infrastructure.

Question 10

Can my manager see every mistake I make?

Accepted Answer

No. The privacy architecture is the product, not a side feature.

Personal mistakes belong to the employee. The manager sees aggregate team capability — strengths, common gaps, training velocity — with no drill-down to any named individual. The HR view operates at role-and-function level, not person level. There is no drill-down from a manager to an individual without explicit, recorded consent from that individual.

This isn't a configuration setting we let admins toggle off. It is wired into the data model. It is also why Compass aligns three buyers — CEO, HR, employee — that competing products typically split.

Question 11

Will our data train anyone's model — yours, or a foundation model vendor's?

Accepted Answer

No.

We sign zero-retention agreements with every foundation-model provider before any data flows. If a provider cannot offer zero-retention for your account, we do not use them on your account — there are always alternatives.

Klaut, the company, does not operate a model of our own that learns from your data. The per-role mistake taxonomy Compass develops is scoped to your tenant — it is never pooled across clients, sold, or used to train a shared model.

Question 12

EU AI Act, Indonesian PDP Law, SOC 2 — where do you stand on compliance?

Accepted Answer

Honestly: two of those are architectural compliance, one is operational and not yet finished.

EU AI Act: the privacy-bounded three-view design directly addresses worker-impact obligations. We provide the worker-impact assessment artifact the Act requires for high-risk AI systems.

Indonesian PDP Law (UU PDP): data residency, purpose limitation, and consent are designed into the data model — your data stays in your environment, scoped to the stated purpose.

SOC 2 Type II: on the operational moat list, targeted Q4 2027 as part of Series A readiness. If SOC 2 attestation is a hard procurement gate today, we are not yet a fit and we'll say so on the first call.

Question 13

What does a Klaut engagement cost end-to-end?

Accepted Answer

Three layers, every price agreed in writing before any work starts.

Foundation (paid discovery and assessment): $8,000–$25,000 depending on scope and client size. Fixed price.

Implementation (converting Foundation findings into a deployed system): $25,000–$120,000 depending on integration complexity and how many functions get Compass-deployed. Fixed price.

Compass subscription (the recurring layer): $5,000–$18,000 per month at mid-market scale, per-seat per-month, tied to the number of employees Compass observes.

No retainers. No hourly billing. No surprise line items at the end. If we miss the metric we agreed in Foundation, we keep working at our cost until we hit it — or refund the final payment and walk.

Question 14

Why is the Foundation discovery paid, not free?

Accepted Answer

Free discovery doesn't work for either of us.

Free discovery selects for clients who haven't yet decided whether AI is real for their business — we end up doing pre-sales work for companies who never close. Paying for Foundation pre-qualifies you for the engagement that follows, and we use the money to do good diagnostic work rather than gloss over hard answers to keep you happy.

About 1 in 5 Foundation clients decide not to proceed to implementation. We think that's healthy — both sides should walk away if the diagnostic doesn't support the bigger spend.

Question 15

Compass subscription — per seat? Annual? Cancellable?

Accepted Answer

Per-seat per-month, billed monthly, tied to the number of employees Compass observes in the deployed function.

Standard term is 12 months once implementation goes live. Cancellation terms sit in the MSA — there's no auto-renew gotcha; you re-sign or you don't.

Lifetime value compounds as you deploy Compass into additional functions under the same MSA — start with customer service, add sales next year, then support engineering, then operations. Per-deployment overhead amortises, so the marginal cost per added function drops over time.

Question 16

How do we get started — what does the first 90 days look like?

Accepted Answer

Three phases.

Weeks 1–2 · Foundation kickoff. We map your current AI exposure, identify the highest-leverage workforce intelligence opportunity, agree the metric we'll measure against, and scope the implementation. Deliverable: a one-page proposal you could take to any other vendor — yours to keep.

Weeks 3–8 · Implementation. Compass deploys into the agreed function. Integrations to your systems. Privacy architecture configured for your industry and jurisdiction. Dashboards built. We're on-site for week 1 of this phase.

Weeks 9–12 · Pilot run. Compass starts observing. The five-step loop runs. The first monthly reports go out to employee, manager, and HR. By day 90 you have a live deployment generating compounding value — not a pilot ending in a slide deck.

Question 17

Who from our team needs to be in the room?

Accepted Answer

Three people, non-negotiable.

One executive (CEO, COO, or function head) who can unblock procurement and security in an afternoon, not a fortnight.

One operational owner — the actual head of the function Compass is deploying into. Head of Customer Service for the first vertical. They co-own the metric we agreed in Foundation.

One IT or security lead to handle the integration to your existing systems and the privacy review against your governance regime.

Without those three present and accountable, we don't take the engagement. We've learned this the expensive way.

Question 18

Remote or on-site? Where does Klaut work from?

Accepted Answer

Jakarta is home.

We're on-site for week 1 of implementation and for the final cutover, remote in between with a 4-hour overlap on your business hours. Indonesian and Southeast Asian clients get full on-site availability throughout the engagement if needed.

Global clients (Europe, US, LatAm) get monthly on-site visits, with travel included in the engagement price — no separate travel line items, no expense reports to approve afterwards.

Question 19

Honest question — does Klaut actually have paying clients today?

Accepted Answer

No, not yet. Pre-revenue, MVP stage. We're not going to pretend otherwise.

As of May 2026: 38 prospects identified, 9 active conversations, 4 qualified opportunities, 1 proposal sent, 0 contracts signed. First paying Foundation client targeted end of Q2 2026. Compass MVP targeted August 2026. First pilot client targeted October 2026. YC Summer 2027 application Q1 2027 with that pilot data in hand.

If "ten reference customers in your vertical" is your minimum bar, we are not your firm yet. The right step is to come back in Q1 2027 once the first pilot case study is live, or stay in the conversation now and let us earn the deployment once Compass is real.

Question 20

Klaut is small. What happens to our deployment if Klaut shuts down?

Accepted Answer

A fair question to ask any pre-Series-A vendor, and one we'd rather you ask now than discover post-signing.

Two safeguards. One, every Compass deployment ships with escrowed source access in the MSA — if Klaut becomes unable to deliver, you retain the right to access and operate the deployed system. Two, Compass runs on standard cloud infrastructure (your VPC, your keys, your foundation-model provider account) — the deployed system continues operating independently of Klaut's continued existence.

We'd rather lose a deal to a bigger competitor than have you discover a dependency you didn't sign up for.

Frequently asked,
candidly answered.

No matches

Ask a question we should have answered.

Frequently asked, candidly answered.

No matches

Ask a question we should have answered.

Frequently asked,
candidly answered.