<?xml version="1.0" encoding="UTF-8"?><rss version="2.0" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>meshin.co — FAQ</title><description>Fractional Chief AI Officer for 30–300 person companies: AI governance that passes procurement and agentic automation shipped to production.</description><link>https://meshin.co/</link><lastBuildDate>Sat, 12 Sep 2026 00:00:00 GMT</lastBuildDate><item><title>Are models trained on our customer data?</title><link>https://meshin.co/faq/are-models-trained-on-our-customer-data</link><guid isPermaLink="true">https://meshin.co/faq/are-models-trained-on-our-customer-data</guid><description>Not by default.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>Not by default. I use vendor terms that exclude your data from model training, the same enterprise or business-tier terms a compliance team would ask about. I check that clause before any tool touches your data. If a vendor will not commit to that in writing, that tool does not get used on your systems.</content:encoded><category>faq</category></item><item><title>Can you get us certified to ISO/IEC 42001, or do you just build a working system?</title><link>https://meshin.co/faq/can-you-get-us-iso-iec-42001-certified-or-just-a-working-system</link><guid isPermaLink="true">https://meshin.co/faq/can-you-get-us-iso-iec-42001-certified-or-just-a-working-system</guid><description>I build a working system, not the certification itself.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>I build a working system, not the certification itself. I work to get your management system, risk register, and evidence trail to the point an auditor would recognise as ready. An accredited body runs the independent audit that grants certification. Whether you then pursue ISO/IEC 42001 itself is a separate decision, best made once you can see the system actually running.</content:encoded><category>faq</category></item><item><title>Can you work with our existing AI vendor or platform, or do you bring your own tools?</title><link>https://meshin.co/faq/can-you-work-with-our-existing-ai-vendor-or-platform</link><guid isPermaLink="true">https://meshin.co/faq/can-you-work-with-our-existing-ai-vendor-or-platform</guid><description>I work with what you already run.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>I work with what you already run. OpenAI, Anthropic, Azure, or whatever your team picked before I arrived — the Retainer does not require a vendor switch. What changes is the governance layer around the tools you keep. That means a data flow review, contract terms checked for training exclusions, and evals wrapped around workflows already running. It is not a rebuild from scratch.</content:encoded><category>faq</category></item><item><title>How does AI risk get recorded, and who signs it off?</title><link>https://meshin.co/faq/how-does-ai-risk-get-recorded-and-who-signs-it-off</link><guid isPermaLink="true">https://meshin.co/faq/how-does-ai-risk-get-recorded-and-who-signs-it-off</guid><description>In an AI register.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>In an AI register. One line per initiative: its data flows, its risk level, and its owner. Under the Retainer, I review every entry and sign it as the named officer. Your organisation still holds final accountability for anything that reaches customers or the board, as it does for any other risk register you keep.</content:encoded><category>faq</category></item><item><title>How is a prompt or model change stopped from shipping unnoticed?</title><link>https://meshin.co/faq/how-is-a-prompt-or-model-change-stopped-from-shipping-unnoticed</link><guid isPermaLink="true">https://meshin.co/faq/how-is-a-prompt-or-model-change-stopped-from-shipping-unnoticed</guid><description>Evals in continuous integration.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>Evals in continuous integration. A test set runs on every change and fails the build when quality drops, the same gate your engineers already trust for code. A prompt edit, a model swap, a vendor&apos;s silent update: all pass the same evals as the original release before reaching production. Tracing shows what changed.</content:encoded><category>faq</category></item><item><title>Is my company too small, or too early, for a Fractional Chief AI Officer?</title><link>https://meshin.co/faq/is-my-company-too-small-or-too-early-for-a-fractional-chief-ai-officer</link><guid isPermaLink="true">https://meshin.co/faq/is-my-company-too-small-or-too-early-for-a-fractional-chief-ai-officer</guid><description>Size isn&apos;t the real test.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>Size isn&apos;t the real test. If you&apos;re running pilots that never reach production, or a customer or investor has already asked who owns your AI risk, you&apos;re ready. If you have no engineers, no pilots, and nobody outside the company has asked that question yet, you&apos;re early. Fix that first, then book a call.</content:encoded><category>faq</category></item><item><title>What counts as &apos;in production&apos; for the 90-day Sprint?</title><link>https://meshin.co/faq/what-counts-as-in-production-for-the-90-day-sprint</link><guid isPermaLink="true">https://meshin.co/faq/what-counts-as-in-production-for-the-90-day-sprint</guid><description>Real users, real transactions, a rollback path, someone on call.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>Real users, real transactions, a rollback path, someone on call. That is the bar, monitored end to end. A demo behind a feature flag does not clear it, and neither does a workflow only your team tried. The success measure gets written into the statement of work before the Sprint starts. That settles whether it actually shipped before the work even begins.</content:encoded><category>faq</category></item><item><title>What do we do about tools staff are already using without approval?</title><link>https://meshin.co/faq/what-do-we-do-about-tools-staff-are-already-using-without-approval</link><guid isPermaLink="true">https://meshin.co/faq/what-do-we-do-about-tools-staff-are-already-using-without-approval</guid><description>Add unapproved tools to the AI register, not a ban list.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>Add unapproved tools to the AI register, not a ban list. That beats banning them, which just pushes the same behaviour further underground. The Retainer&apos;s first step is exactly this: inventory the AI tools and data flows already in the building, then write a data classification and an acceptable-use policy your engineers can actually follow.</content:encoded><category>faq</category></item><item><title>What does a Fractional Chief AI Officer actually do, and how is that different from an advisor?</title><link>https://meshin.co/faq/what-does-a-fractional-chief-ai-officer-actually-do</link><guid isPermaLink="true">https://meshin.co/faq/what-does-a-fractional-chief-ai-officer-actually-do</guid><description>I carry the AI register as a named officer.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>I carry the AI register as a named officer. I review every initiative, sign the answers to security questionnaires, and sit in design reviews. An advisor gives an opinion and leaves; I own the outcome and answer for it. A week usually holds a security questionnaire, a design review, and a page for the board.</content:encoded><category>faq</category></item><item><title>What does the two-week Assessment produce, and is the fee credited if we continue?</title><link>https://meshin.co/faq/what-does-the-two-week-assessment-produce-and-is-the-fee-credited-if-we-continue</link><guid isPermaLink="true">https://meshin.co/faq/what-does-the-two-week-assessment-produce-and-is-the-fee-credited-if-we-continue</guid><description>The two-week Assessment produces a memo your board can act on.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>The two-week Assessment produces a memo your board can act on. The memo covers a readiness score across strategy, data, delivery, people and governance. It includes a gap list against NIST AI RMF, ISO/IEC 42001 and the EU AI Act. It also ranks three processes by what automating them is worth this quarter. Yes: if you move to the 90-Day Sprint within three months, the full Assessment fee is credited off the Sprint price.</content:encoded><category>faq</category></item><item><title>What happens if the evals say a workflow isn&apos;t good enough?</title><link>https://meshin.co/faq/what-happens-if-the-evals-say-a-workflow-isnt-good-enough</link><guid isPermaLink="true">https://meshin.co/faq/what-happens-if-the-evals-say-a-workflow-isnt-good-enough</guid><description>If the evals say no, the workflow does not ship.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>If the evals say no, the workflow does not ship. The Sprint ends with a written account of why the workflow fell short and what would need to change. That is a cheaper outcome than a launch that quietly under-performs until a customer notices. The eval numbers decide whether the workflow ships, whatever the deadline is.</content:encoded><category>faq</category></item><item><title>What if the AI we&apos;re building goes into a product we sell, not just an internal workflow?</title><link>https://meshin.co/faq/what-if-the-ai-we-are-building-goes-into-a-product-we-sell</link><guid isPermaLink="true">https://meshin.co/faq/what-if-the-ai-we-are-building-goes-into-a-product-we-sell</guid><description>I cover products you sell as well as internal workflows.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>I cover products you sell as well as internal workflows, and the obligations differ. Building AI into a product you sell can make your company a provider rather than a deployer under the EU AI Act. That carries a higher compliance bar than using AI only inside the business. I record that distinction in the AI register from the first review. That way you decide it early instead of discovering it after the feature has shipped.</content:encoded><category>faq</category></item><item><title>What if we already have a data science team or an internal AI lead?</title><link>https://meshin.co/faq/what-if-we-already-have-a-data-science-team-or-an-internal-ai-lead</link><guid isPermaLink="true">https://meshin.co/faq/what-if-we-already-have-a-data-science-team-or-an-internal-ai-lead</guid><description>Your data science team stays exactly where it is.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>Your data science team stays exactly where it is. I do not write models, tune pipelines, or manage engineers — what I bring is different. I own the AI register as the named officer, sign the procurement answers, and report to the board on AI decisions. A company of your size usually has the technical people already; what it lacks is someone accountable for AI outside the team itself.</content:encoded><category>faq</category></item><item><title>What&apos;s the difference between the Assessment, the Sprint and the Retainer?</title><link>https://meshin.co/faq/whats-the-difference-between-the-assessment-the-sprint-and-the-retainer</link><guid isPermaLink="true">https://meshin.co/faq/whats-the-difference-between-the-assessment-the-sprint-and-the-retainer</guid><description>Three fixed offers, three stages.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>Three fixed offers, three stages. The Assessment is two weeks at a fixed fee, ending in a memo on where you stand and what to fix first. The Sprint is 90 days at a fee quoted from the agreed scope. It puts one agentic workflow live in production, with evals and a rollback path. The Retainer is a monthly fee, one or two days a week, as your named AI officer. If unsure, start with the Assessment.</content:encoded><category>faq</category></item><item><title>Who answers customer security questionnaires?</title><link>https://meshin.co/faq/who-answers-customer-security-questionnaires</link><guid isPermaLink="true">https://meshin.co/faq/who-answers-customer-security-questionnaires</guid><description>I do, as the named officer on record.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>I do, as the named officer on record. That is a core part of the Retainer: procurement answers and questionnaire responses as they arrive. My answers rest on the AI register and the acceptable-use policy, so they match what is actually running.</content:encoded><category>faq</category></item><item><title>Who owns the code and the data?</title><link>https://meshin.co/faq/who-owns-the-code-and-the-data</link><guid isPermaLink="true">https://meshin.co/faq/who-owns-the-code-and-the-data</guid><description>You own both the code and the data.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>You own both the code and the data. The code lives in your repositories from the first commit, under your licence and your review rules. The same principle applies to your data: nothing about the engagement moves ownership of the code or the data to me or to a vendor. What can change hands is access, temporarily and by agreement, for the work that needs it.</content:encoded><category>faq</category></item><item><title>Who owns the frameworks, templates and playbooks built during the engagement?</title><link>https://meshin.co/faq/who-owns-the-frameworks-templates-and-playbooks-built-during-the-engagement</link><guid isPermaLink="true">https://meshin.co/faq/who-owns-the-frameworks-templates-and-playbooks-built-during-the-engagement</guid><description>Everything written for your business belongs to you.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>Everything written for your business belongs to you. The AI register, the policies, the runbooks — all of it, the same as your code and data. What stays mine is the method: how I structure an assessment, build eval templates and run a risk review. I reuse that method with other clients, never your specific content, data or configuration.</content:encoded><category>faq</category></item><item><title>Why hire a fractional officer instead of a full-time head of AI?</title><link>https://meshin.co/faq/why-hire-a-fractional-officer-instead-of-a-full-time-head-of-ai</link><guid isPermaLink="true">https://meshin.co/faq/why-hire-a-fractional-officer-instead-of-a-full-time-head-of-ai</guid><description>You get a named officer&apos;s accountability, without a full-time salary.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>You get a named officer&apos;s accountability, without a full-time salary, a hiring process, or a seat that sits empty between AI decisions. That means a named officer on your customers&apos; security reviews and your board&apos;s roadmap. The comparison is between different kinds of cost, not price against price: a monthly retainer fee against a full-time hire&apos;s salary, equity, recruiting and ramp time. If you need someone hands-on across five systems at once, that is a full-time hire, not a Retainer.</content:encoded><category>faq</category></item><item><title>How does the Retainer end?</title><link>https://meshin.co/faq/how-does-the-retainer-end</link><guid isPermaLink="true">https://meshin.co/faq/how-does-the-retainer-end</guid><description>With a named internal owner in place, not a fading call.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>With a named internal owner in place, not a fading call. The decisions that got made are written down, not quietly dropped. The handover plan exists from the first month, not the last. A retainer running past two years would mean the design was wrong; the plan is written to end well before that.</content:encoded><category>faq</category></item><item><title>How quickly can you start, and how does availability work for a remote officer?</title><link>https://meshin.co/faq/how-quickly-can-you-start-and-how-does-availability-work-for-a-remote-officer</link><guid isPermaLink="true">https://meshin.co/faq/how-quickly-can-you-start-and-how-does-availability-work-for-a-remote-officer</guid><description>One client at a time.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>One client at a time. That is what keeps me reachable: one or two days a week, never spread across five accounts. I work remote worldwide, so time zones are a scheduling detail, not a blocker. Book a fit call to check current availability; if the timing is wrong, I will tell you rather than stall.</content:encoded><category>faq</category></item><item><title>What size of company do you actually work with?</title><link>https://meshin.co/faq/what-size-of-company-do-you-actually-work-with</link><guid isPermaLink="true">https://meshin.co/faq/what-size-of-company-do-you-actually-work-with</guid><description>Companies of 30 to 300 people, with at least one engineer.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>Companies of 30 to 300 people, with at least one engineer. Most companies here have run AI pilots, and they stalled before reaching production. Below that size, the process volume and engineering capacity are rarely there yet. Above it, you usually have an internal AI leader already and need something else.</content:encoded><category>faq</category></item><item><title>How do you charge for the work?</title><link>https://meshin.co/faq/how-do-you-charge-for-the-work</link><guid isPermaLink="true">https://meshin.co/faq/how-do-you-charge-for-the-work</guid><description>Three fixed structures, no hourly billing.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>Three fixed structures, no hourly billing. The Assessment is a fixed fee for the two weeks. The Sprint is quoted once, from the scope in the statement of work. The Retainer is a monthly fee for one or two days a week. You hear the number on the fit call, and it is written down before any work starts, so there is no surprise invoice partway through.</content:encoded><category>faq</category></item><item><title>How is cost per outcome measured?</title><link>https://meshin.co/faq/how-is-cost-per-outcome-measured</link><guid isPermaLink="true">https://meshin.co/faq/how-is-cost-per-outcome-measured</guid><description>On a dashboard built during the Sprint.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>On a dashboard built during the Sprint. Total spend on the workflow, model calls included, divided by how many times it produced the outcome. You define that outcome before the Sprint starts. The number sits next to tracing. So when it moves, you can see why: volume, a model change, or a real quality problem.</content:encoded><category>faq</category></item><item><title>What do NIST AI RMF, ISO/IEC 42001 and the EU AI Act each require of a company our size?</title><link>https://meshin.co/faq/what-do-nist-ai-rmf-iso-iec-42001-and-the-eu-ai-act-require</link><guid isPermaLink="true">https://meshin.co/faq/what-do-nist-ai-rmf-iso-iec-42001-and-the-eu-ai-act-require</guid><description>None of the three sets a headcount threshold.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>None of the three sets a headcount threshold. NIST&apos;s AI RMF is voluntary guidance built around four functions: govern, map, measure and manage. ISO/IEC 42001 is a certifiable standard for running an AI management system. The EU AI Act is binding law: what you owe depends on being a provider (building AI into what you sell) or a deployer (only using it inside the business), plus the system&apos;s risk tier, not your size.</content:encoded><category>faq</category></item><item><title>What happens on a fit call, and does it cost anything?</title><link>https://meshin.co/faq/what-happens-on-a-fit-call-and-does-it-cost-anything</link><guid isPermaLink="true">https://meshin.co/faq/what-happens-on-a-fit-call-and-does-it-cost-anything</guid><description>It costs nothing and runs thirty minutes, booked at cal.com/meshin.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><content:encoded>It costs nothing and runs thirty minutes, booked at cal.com/meshin. Tell me about the process you want in production, who owns it today and your timeline. By the end you will know whether an Assessment, a Sprint or neither is the right next step — no slide deck required.</content:encoded><category>faq</category></item></channel></rss>