Answers / FAQ
Every Question, Answered in Full
Buyers ask us roughly the same set of questions before a pilot. Every one of them is already answered somewhere on this site, which is not much help when you need a single link. This is that link: what it costs, how it starts, whether it reaches your machines, whether your IT team will approve it, whether your operators will use it, what it actually does, and whether you should believe any of it.
How this page works
Every answer below is complete. Nothing is teased to make you click, and no answer stops at the point where it gets uncomfortable. Where a page goes further than a paragraph reasonably can, the link sits at the end of the answer rather than in place of it.
The seven groups run in the order the questions usually arrive: price first, then the practical mechanics, then the two gates that actually kill projects (IT and the floor), then what the product is, then whether the numbers mean anything. If your question is not here, that is worth telling us, and we will add it.
One thing to know before you read: KaizenFlow is at design-partner stage. We have no customer case studies, no logos, and no achieved results to quote. The percentage ranges you see anywhere on this site are modeled targets used for planning. Several answers below say so directly, because the alternative is a page that reads well and misleads you.
Pricing and contract
What does KaizenFlow cost? Three engagements, all published. The pilot is a fixed $25,000 to $75,000 for eight weeks, scoped by how many lines and connectors we baseline. Pro is $5,000 to $15,000 per month, from $5k for a single facility. Enterprise is custom and covers unlimited facilities, private VPC deployment, custom connectors and SSO, a dedicated CSM, and an SLA. The full breakdown is on the pricing page.
Is it priced per machine or per site? By site and scope, not by machine. Pro starts at $5,000 per month for a single facility, and the pilot fee is set by the number of lines and connectors we baseline, agreed before kickoff with no usage-based surprises. Because KaizenFlow installs no hardware, there is no per-sensor line item that grows with your machine count.
Do you have a free trial? No. A meaningful evaluation has to connect to your MES, SCADA, ERP, or historian and baseline real production, which is real work for both sides. The paid pilot keeps both sides serious, and it credits in full if you continue. If you stop, you keep the verified report.
What happens to the pilot fee if we continue? It credits in full against your first-year agreement. If you continue, the pilot effectively costs nothing beyond that first year. If you stop, you keep the verified report.
Is there a contract lock-in? No. Either party can end the agreement with 30 days' written notice. On termination you get a 30-day window to export everything, after which your data is permanently deleted. That is written into our terms rather than promised in a sales call.
What payback period should we expect? We do not quote a payback number - at design-partner stage we will not dress a model up as a track record. What we will do is show you the model. The ROI calculator takes four inputs: facilities, revenue per facility, unplanned downtime hours per month, and scrap rate. It applies the modeled ranges (8-18% downtime, 5-12% scrap, 4-11% throughput, 3-7% energy), infers an hourly cost by dividing revenue by an assumed 6,000 production hours per facility per year, and credits 35% of a recovered hour as margin. It does not subtract subscription cost, so what it prints is gross modeled savings, not payback. Put Pro pricing at $5k-15k per month against that figure yourself, and correct it if 6,000 hours a year is wrong for your plant, because it often is. Your pilot then replaces the whole model with measured numbers in eight weeks.
Will the price go up after we sign? Design-partner rates are locked for your agreement term. We may adjust pricing at renewal or for new agreements, with at least 30 days' written notice before a renewal takes effect. Nothing moves mid-term.
Getting started and implementation
How long until we are live? Most single-site connections are live in one to two weeks. There is no hardware to install and no historian to migrate, so the timeline is driven by credentialing and tag mapping rather than infrastructure. The full pilot runs eight weeks and ends in a verified before and after savings report.
How much of our team's time does it take? Plan for three things: an executive sponsor in a one-hour weekly ROI review, someone who can grant read-only access to your systems in week one, and floor time in weeks 6 to 8 to execute the opportunities you select. We do the connecting, calibrating, and analysis - there is nothing for your team to install or maintain, and no dedicated hire is needed.
Who does the work? We do. Every plan includes onboarding and a dedicated success engineer who stands up the connections, maps the tags, and runs the analysis. Your team executes the improvements it selects, because those are your machines and your people. Enterprise adds a dedicated CSM and SLA commitments.
Do we have to replace our MES, SCADA, ERP, or historian? No. KaizenFlow connects on top of the systems you already run and reads from them in place. Nothing is ripped out, and your operators keep their existing HMIs and dashboards.
What actually happens during the eight weeks? Three phases, each with a clear exit. Weeks 1 to 2, connect and calibrate: read-only connections come up and you agree the baseline. Weeks 3 to 5, surface and prioritize: the specialists rank every opportunity by dollar impact and confidence, and your team picks what to act on. Weeks 6 to 8, execute and verify: you act, we reconcile against the week-one baseline, and the numbers go into a report finance can check line by line. The pilot program page covers each phase.
Machines, hardware and integrations
Do we need to install sensors? No. KaizenFlow installs no hardware and adds no new signals. It reads what your plant already emits through 43+ pre-built connectors to your MES, SCADA, ERP, PLCs, and historians. New instrumentation is only worth it later, and only if a specific opportunity depends on it.
Will it work with our older machines? Often, yes. Modbus runs over TCP and serial and reaches equipment that predates OPC-UA, MTConnect and FANUC FOCAS cover CNC and machine tools, and where a control is too old to reach directly a lightweight edge adapter sits alongside it and reads. The test is whether the machine emits a signal something can read. Our guide to monitoring old machines walks through each case.
Which systems do you connect to? ERP: SAP, Microsoft Dynamics 365 Business Central, Oracle NetSuite, Oracle E-Business Suite, Infor, Epicor. SCADA and historians: OSIsoft PI, Ignition, Kepware, Rockwell FactoryTalk, AVEVA Wonderware, GE Proficy. PLC: Rockwell (Allen-Bradley), Siemens TIA Portal, Mitsubishi, Beckhoff, Omron. Protocols: OPC-UA, MQTT, Modbus TCP and RTU, REST, ODBC/SQL, and CSV over SFTP. The integrations page goes deeper on how each layer connects.
What if our system is not on the connector list? The generic OPC-UA and REST connectors cover most of the long tail. For genuinely bespoke systems, building a net-new connector is part of the design-partner onboarding process.
Our tags are a mess. Is that a problem? No, that is normal. You do not need a data-science team, a new historian, or a finished data lake. Cleaning enough tags to measure reliably is part of the work we do in the first two weeks, and we confirm the numbers reconcile against what your team already trusts before anything is surfaced as an opportunity.
When is KaizenFlow the wrong choice? When there is nothing to read. If your plant has almost no digital infrastructure - older machines, no PLCs worth reading, no ERP data - clip-on sensing is the faster start, and Guidewheel is a credible choice. KaizenFlow needs signals to read. We are also the wrong tool if what you want is a CMMS, an MES, or employee surveillance.
Security, IT and data
What data actually leaves the plant? Only what each connector is scoped to read: process tags, event and state data, counts, timestamps, and the cost rates you provide for dollar math. KaizenFlow does not read control logic or recipe IP, and collects no employee personal data by default. Every connector's scope is visible to your admins, and data moves over TLS 1.3 with AES-256 encryption at rest.
Can we run this on-premise or air-gapped? Not today. The standard platform is multi-tenant cloud with strict logical isolation, where every query is scoped by tenant ID. Enterprise deployments support private VPC isolation and data-residency requirements, and that is as close to your own four walls as we go. If your policy requires a fully air-gapped install, we are the wrong vendor right now, and we would rather say so than sell you a six-month discovery.
Can KaizenFlow write to our PLCs? No. Connections are read-only by default and KaizenFlow does not send commands to the floor. Recommendations are routed to people, who act through your existing systems and procedures.
What happens if the internet drops? Nothing, on your floor. KaizenFlow observes - it does not control - so a connectivity outage never stops a line. Analytics resume when the connection does; where an edge adapter is deployed, reads continue locally and sync when the link returns. The exact behavior for your topology is covered in the security brief.
Is KaizenFlow SOC 2 certified? No, and we will not imply otherwise. We operate to SOC 2 Type II and ISO 27001 control frameworks, but alignment is not certification, no audit report exists, and we are not going to dress the first thing up as the second. If a signed SOC 2 report is a hard gate in your procurement, we do not clear it today, and you should know that before you spend a meeting on us. What we can share now: control documentation, architecture diagrams, the subprocessor list, and data-flow documentation - request the security brief.
Who owns the data? You do. You retain full ownership of everything you upload or connect, and we do not sell, share, or use your production data for anything other than providing the service. Manufacturing metrics are retained per your organization settings, 12 months by default, and audit logs for 24 months.
Is our data used to train AI models? We do not sell your data or use it to train third-party models. AI analysis runs through vetted providers (OpenAI, Anthropic) under contractual data-protection obligations. Aggregated, anonymized benchmarks may be used to improve the platform, as set out in our terms.
Can we export our data? Yes, on every plan. Your data is yours: metrics, events, and the savings ledger can be exported or read over the API for use in your own BI tools. On exit, you get a full data export - that is written into our terms, not buried in a sales deck.
Who can see what inside the platform? Access is role-based across four roles: Admin, Manager, Engineer, and Viewer. Authentication is JWT-based with configurable token lifetimes, sensitive endpoints are rate limited, and every recommendation and action is attributable and audit-logged. Your security team can work through the whole model in the IT review pack.
Operators and adoption
How much data entry do operators have to do? None to get OEE. KaizenFlow reads run, stop, rate, and quality signals from the systems you already have - MES, SCADA, PLCs, historians - so the numbers do not depend on operators logging codes at a terminal. Operator input is optional context, never the source of truth.
Is this monitoring machines, or monitoring people? Machines and process. KaizenFlow prices losses by machine, line, shift pattern, and cause - it is not a time-and-motion tool, and it does not score individual operators. If you want employee surveillance, we are the wrong vendor.
Our last system died because nobody used it. Why is this different? Two structural reasons, not a claim about our interface. First, the numbers do not depend on adoption: OEE is read from your MES, SCADA, PLCs, and historians, so it stays correct whether or not anyone logs a code at a terminal. Second, the pilot does not end in a dashboard, it ends in a before and after number your finance team signs. A system nobody opens cannot produce that, which is why we would rather be judged on the report than on login counts. The operator adoption page goes deeper.
Do operators need training? Not in order to be measured, because the data comes from your systems rather than from them. For teams that want it, the AI Academy is free to learn: seven role paths - operators, supervisors, quality and CI, engineers, plant managers, executives, and IT - plus timed practice exams, no login and no paywall. You can send people through all of it before you sign anything. Certification is the paid half: operators certify free with any platform license, the other paths are $649 each, and a plant-wide license caps a whole site.
What the product does
What is OEE, and what is TEEP? OEE is Availability multiplied by Performance multiplied by Quality, measured against planned production time. TEEP measures the same effectiveness against all the time there is by adding a Utilization factor, so TEEP is OEE multiplied by Utilization. OEE is the metric for shop-floor improvement; TEEP is the metric for capacity and capital-expansion decisions. The OEE and TEEP guide has the formulas and a worked example.
Is this a CMMS? No. KaizenFlow does not dispatch maintenance, manage work orders, or run production. It ranks losses by dollar and carbon impact and verifies the savings once you act, then hands its findings to the systems that run the work. It is not a CMMS or MES replacement.
Does it replace our BI tools? No. KaizenFlow does not replace a BI stack. General BI tools remain the right choice for ad-hoc reporting and executive dashboards. KaizenFlow adds a purpose-built loop that ranks losses by dollar impact and produces a verified savings ledger, and the two can feed the same dashboards.
What do the nine AI specialists actually do? Each owns a loss type and scores it in dollars. Anomaly Sentry votes across Z-score, MAD, IQR, and Isolation Forest to catch drift; Throughput Analyst finds bottlenecks; Quality Sentry runs SPC and Cpk; Energy Optimizer detects idle load and peak draw; Reliability Forecaster predicts MTBF; Schedule Strategist compresses changeovers; Yield Modeler attributes scrap; Maintenance Planner replaces fixed intervals with predicted ones; and Savings Auditor reconciles the before and after for finance. They feed one queue ranked by dollar impact and confidence.
Will it create alarm fatigue? The goal is the opposite. Instead of a raw alarm stream, findings are voted on for confidence and scored for dollar impact, then presented as a short ranked queue. Low-confidence, low-cost blips stay out of the way so the team's attention goes to the few anomalies actually worth acting on.
Does the AI decide anything on its own? No. AI output is advisory. Connections are read-only, so the platform executes nothing on the floor; recommendations are routed to people who act through your existing systems and procedures. Decisions, and responsibility for them, remain with your team.
Proof and verification
Will you share other customers' results? No. We are at design-partner stage and do not publish customer numbers, case studies, or logos. The only results we will show you are your own.
Are the modeled percentages real results? No. KaizenFlow is at design-partner stage, and the ranges are modeled targets used for planning, not achieved outcomes. The ledger exists precisely so that any real saving is measured and signed rather than assumed from a range.
What makes a saving verified rather than estimated? A verified saving has cleared three bars: a baseline normalized for shift calendar, product mix, and planned downtime; a before-and-after delta scored for confidence rather than asserted; and a sign-off from your finance team against actuals they can trace. An estimate has cleared none of them.
Who signs the savings report? It is built for your finance team to review and sign. Every figure traces back to source data and a stated method, so it can be checked line by line.
How accurate are the forecasts? We validate forecasts against real outcomes rather than quoting a fixed accuracy number. Anomaly findings require several independent methods to agree and carry a confidence score, and results that do not clear the confidence bar stay in a measured-but-unverified state and never reach the signed ledger.
The mechanics behind all five of those answers live on the savings ledger page.
Ask us something that is not here
If your question is missing, it is probably the interesting one. Send it over and you will get a straight answer, including when the answer is that we are the wrong fit for your plant.