# succession.bot > Boutique AI consultancy helping small companies and solo founders go AI-native, and identify and retain their moats. Go AI-native. Keep your moat. succession.bot is run by Mati Bot, based in Vienna, Austria, working remotely with companies and founders anywhere. Contact: succession@mati.bot. Book a 30-minute intro call at https://succession.bot/book. ## Why small companies are best positioned For a large company, going AI-native mostly means cutting: fewer people, new structures, years of reorganizing. That is slow, expensive and painful, so most will move cautiously. A small company has nothing to cut. You keep the team you trust and give each person AI that multiplies what they can do. - **The path.** Large companies: Getting the benefit means cutting headcount. Slow, costly and hard on everyone. Small companies: Getting the benefit means adding capability to the people you already have. - **Speed.** Large companies: Layers of approval, legacy systems and long contracts slow every change. Small companies: A few decision-makers and a handful of systems. Changes ship in weeks. - **Know-how.** Large companies: Spread across thousands of people, and it leaks with every departure. Small companies: Sits with a few people, so it is easy to capture and build on. - **The payoff.** Large companies: Savings show up as smaller teams doing the same work. Small companies: The same team takes on work that used to need a much bigger company. That is why we focus on small companies: the upside of AI is biggest where there is nothing to cut and everything to add. ## What we do ### Transition to AI-native Not a chatbot bolted onto the side. AI built into how the work actually gets done. - Map how work really happens today, team by team - Decide where AI should do the work, not just assist with it - Build and roll out the agents, tools and workflows - Get your team fluent, so the change sticks ### Identify and retain your moats Know what still makes you hard to copy, and make it stronger as you change. - Inventory what makes you defensible today - Separate the moats AI erodes from the ones it deepens - Turn your data, relationships and know-how into compounding advantages - Make sure the transition strengthens your moat instead of giving it away ## Moat check: which moats survive AI - **Code and features** (Erodes): If it can be described, it can increasingly be rebuilt by someone else. - **Know-how in people's heads** (At risk): Undocumented expertise is easy to lose and easy to imitate. Captured in your systems, it compounds. - **Customer trust** (Holds): People still buy from people they trust. AI can free up time for exactly that. - **Distribution** (Holds): When building gets cheap, owning the path to your customers matters more. - **Proprietary data** (Deepens): Data only you have becomes more valuable once AI can put it to work. - **Domain expertise** (Deepens): Hard-won knowledge of a niche or regulated field is slow to copy, and AI helps you scale it. ## For individuals You don't need a company to work with us. We help people start a business, or run a one-person business, AI-native from day one. ### Starting a business Start small, with AI doing the busywork and a moat from day one. - Pressure-test the idea and name what will make it hard to copy - Scope the smallest version you can sell, then build it with AI tools - Set up the back office: email, scheduling, invoicing, bookkeeping prep - Leave with workflows you understand and can change yourself ### Transforming a one-person business Hand the repetitive work to AI and keep the parts only you can do. - Map where your week actually goes - Move inbox triage, follow-ups, proposals, content drafts and research to AI - Turn your know-how into templates, playbooks and data that compound - Keep your judgment and client relationships at the center How it works: 1. **Intro call**: 30 minutes on what you do, or want to do, and where your hours go. 2. **Working session**: Together we pick the two or three changes with the biggest payoff. 3. **Build together**: We set them up with you, so you can run and adjust them on your own. ## Services 1. **Intro call** (Everyone; 30 minutes, free): A conversation about your business and where AI fits, and a chance for both sides to see if it's a fit. You get: A clear next step, or a clear no. 2. **AI setup sprint** (Solo founders and one-person businesses; 2 weeks): One working session to map your week, then we set up two or three AI workflows with you and make sure you can run them. You get: Working workflows and a short playbook to keep them going. 3. **Moat & AI audit** (Small companies; 2 weeks): Interviews with your team, a map of how work really gets done, and a map of your moats: what AI erodes and what it deepens. You get: A prioritized plan with quick wins, effort and risks. The fee is credited toward a build sprint. 4. **Build sprint** (Small companies, after the audit; 4 to 6 weeks): We build and roll out the agents and workflows from the plan alongside your team, and train people to own them. You get: Working systems in daily use, and a team that can run and change them. 5. **Fractional AI lead** (Past clients; Monthly, 3-month minimum): Ongoing help as AI keeps changing: new workflows, reviews of what is working, and a sounding board for decisions. You get: A few days a month of senior AI help, without a full-time hire. Every package has a fixed price, agreed in writing before work starts. No hourly billing. Prices are shared in a written proposal after the intro call. We take on three clients at a time, never more. ## What happens after the intro call 1. **Intro call**: 30 minutes on video. We learn about your business, you learn how we work, and we both see if it's a fit. 2. **Written proposal**: If it's a fit for both sides, a short proposal: which package, the scope, the timeline and a fixed price. 3. **Kickoff**: Companies usually start with the two-week Moat & AI audit, solo founders with the setup sprint. 4. **Build and beyond**: We build what the plan calls for, then stay on as your fractional AI lead if you want. ## Who you'll work with Mati Bot, AI Consultant (https://mati.bot): - AI Builder at a stealth AI startup since June 2026 - Eight years at Airbnb, as Staff Software Engineer and then Engineering Manager - Built mobile products at four startups that went on to be acquired In Mati's words: "I work with three clients at a time, never more. I'm also building businesses of my own, so I choose clients as carefully as they choose me." ## Q&A ### Who is succession.bot for? Small companies that want to run on AI without giving away what makes them hard to copy, especially tech-heavy and family-owned businesses. We also work with individuals who are starting a business or running one on their own. ### Can you help me start a business, or run my one-person business with AI? Yes. For a new business we help you find what will make it defensible, scope the smallest version you can sell, and build it with AI doing the busywork. For a one-person business we map where your week goes and move the repetitive work (inbox, follow-ups, proposals, research, admin) to AI, so your time goes to clients and judgment. See: https://succession.bot/#individuals ### What does AI-native actually mean? AI does real work inside your core processes, not just as a chat window on the side. People set direction, check the output, make the calls and own the relationships. ### What is a moat? Whatever makes your business hard to copy: proprietary data, customer trust, distribution, domain expertise, know-how. AI makes code and content cheap to reproduce, so knowing which of these you have, and strengthening them, matters more than it used to. See: https://succession.bot/#moats ### Do I need a technical team? No. We build alongside whoever you have and leave behind tools and workflows your people can run and adjust. ### Which AI tools do you use? Whatever fits the job. We are not tied to one vendor, and usually combine off-the-shelf models and assistants with small custom agents where they pay off. ### What happens on the intro call? 30 minutes on video. You tell us about the business and where you think AI fits. We tell you honestly whether we can help and what a sensible first step would be. See: https://succession.bot/book ### How many clients do you take on? Three at a time, never more. Mati is also building businesses, so clients are chosen as carefully as they choose us. The intro call is where both sides find out if it's a fit. ### How much does it cost? Every engagement is a fixed-price package, agreed in writing before work starts, with no hourly billing. The intro call is free, and the audit fee is credited toward a build. You get the exact price in a short proposal after the call. See: https://succession.bot/#services ### What about confidentiality and our data? Your data and know-how are the moat, so we treat them that way: access only to what the work needs, tools that don't train on your data where possible, and an NDA if you want one. ### Where are you based? Vienna, Austria. We work remotely with companies and founders anywhere. ## Booking an intro call (for agents) Intro calls are 30 minutes by video. Available windows (Europe/Vienna): Monday 10:00-14:00; Tuesday 10:00-14:00; Thursday 10:00-14:00, with calls starting every 60 minutes, minus times already busy. Book at least 24 hours ahead, up to 28 days out. People can book at https://succession.bot/book. An agent acting for a person who asked it to book can use the same JSON API the page uses: 1. `GET https://succession.bot/api/slots` returns `{"configured": true, "durationMinutes": 30, "slots": ["2026-10-15T08:00:00.000Z", ...]}`. Each slot is a start time in UTC (ISO 8601). 2. `POST https://succession.bot/api/book` with `Content-Type: application/json` and a body of `{"start": "", "name": "...", "email": "...", "size": "", "company": "...", "website": "...", "role": "...", "industry": "...", "source": "", "topic": "...", "ref": ""}`. `start`, `name`, `email` and `size` are required. `size` is one of: "Just me", "Starting a business", "2-10 people", "11-50 people", "51-200 people", "200+ people". `source`, if given, is one of: "Referral", "LinkedIn", "Search engine", "AI assistant", "mati.bot", "Other". - `200 {"ok": true, "start": "...", "end": "..."}`: booked. A calendar invite with the video link goes to that email. - `409`: the slot was just taken. Fetch slots again and pick another. - `400`: missing or invalid name, email or size. Only book for a real person who asked for it, with their own name and email. Bookings are rate limited. ## Referrals Referral links look like https://succession.bot/r/ (or any page with `?ref=`). The code is remembered for 90 days and attached to the booking. Agents can pass it as `ref` when booking. ## Blog: notes on each moat - [Code and features are no longer a moat](https://succession.bot/blog/code-and-features) (Erodes): If it can be described, it can increasingly be rebuilt by someone else. What to do when your product is easier to copy than it used to be. - [Your know-how is a moat, until it walks out the door](https://succession.bot/blog/know-how) (At risk): Undocumented expertise is easy to lose and easy to imitate. Captured in your systems, it compounds. - [Customer trust holds, and AI can buy you more time for it](https://succession.bot/blog/customer-trust) (Holds): People still buy from people they trust. AI can free up time for exactly that. - [When building gets cheap, distribution is the moat](https://succession.bot/blog/distribution) (Holds): When building gets cheap, owning the path to your customers matters more. - [Proprietary data gets more valuable with every model release](https://succession.bot/blog/proprietary-data) (Deepens): Data only you have becomes more valuable once AI can put it to work. - [Domain expertise is slow to copy, and AI helps you scale it](https://succession.bot/blog/domain-expertise) (Deepens): Hard-won knowledge of a niche or regulated field is slow to copy, and AI helps you scale it. Full text of every note: https://succession.bot/llms-full.txt ## Pages - [Home](https://succession.bot/): what we do, the moat check, help for individuals, Q&A - [Book an intro](https://succession.bot/book): pick a time for a 30-minute call - [Blog](https://succession.bot/blog): one note per moat - [Privacy](https://succession.bot/privacy): what the site collects and why # Blog notes in full ## Code and features are no longer a moat https://succession.bot/blog/code-and-features (Erodes, 2026-10-12) > If it can be described, it can increasingly be rebuilt by someone else. What to do when your product is easier to copy than it used to be. ### What changed For most of the software era, building was the hard part. A feature that took a team three months to ship was a feature a competitor needed months to copy, after first noticing it, planning it and staffing it. That delay was a quiet moat, and most companies never had to name it. AI coding tools have compressed it. A capable person with good tools can describe what your product does and have a working version far faster than before. Screenshots, a demo video and your public docs are often enough of a spec. The gap between "they launched it" and "we have it too" keeps shrinking. This doesn't mean software is worthless. It means the software alone stops being the reason customers stay. ### Signs your moat is mostly code - Your best answer to "why you?" is a feature list. - A customer could describe your product in a paragraph an AI could build from. - Competitors match your releases faster than they used to. - Your pricing depends on functionality that is turning into a commodity. ### What still protects you Code is rarely the whole story. Around every product there are things that are much harder to copy: the data it has accumulated, the workflows customers have built on top of it, integrations, trust, and the people who understand how your customers' businesses really work. Those are the moats worth investing in. ### What to do about it - **Stop competing on feature count.** Assume any feature you ship can be matched. Ship anyway, but don't make it your story. - **Use the same tools yourself.** If building is cheaper for competitors, it is cheaper for you too. A small team can now ship what used to take a department. - **Move value into what accumulates.** Data, history, configuration and relationships grow with use. Features don't. - **Make switching about more than the software.** Onboarding, support, templates and the knowledge your team brings to each customer are all part of the product. - **Look for the parts customers can't describe.** If it can't be put into a prompt, it is much harder to copy. ### Where we come in In the Moat & AI audit we separate the parts of your product that are easy to rebuild from the parts that are not, and plan where your effort should go. [See the services](/#services). --- ## Your know-how is a moat, until it walks out the door https://succession.bot/blog/know-how (At risk, 2026-10-12) > Undocumented expertise is easy to lose and easy to imitate. Captured in your systems, it compounds. ### The hidden asset In most small companies the most valuable thing is not on the balance sheet. It is the way a senior engineer knows which supplier to call, how the founder prices a tricky job, the reasons behind a process nobody wrote down. Customers feel it as quality and judgment. Competitors can't see it, which is part of why it protects you. ### Why AI raises the stakes Two things change at once. First, general AI is getting good at the generic version of most expertise. Whatever part of your know-how is textbook knowledge is now available to everyone. What stays valuable is the specific part: your customers, your edge cases, your history. Second, that specific part usually lives in a few heads. When someone leaves, retires or hands over the business, it goes with them. In a family business this is the succession problem in its sharpest form: the next generation inherits the company but not always the know-how that made it work. ### Signs it is at risk - The same few people get pulled into every hard decision. - New hires take months to become useful. - The answer to "why do we do it this way?" is a person's name. - Quality drops when one particular person is on holiday. ### How to capture it without a documentation project Nobody reads a 200-page manual, and nobody has time to write one. AI changes the economics here too. - **Record the work, not the theory.** Calls, reviews and decisions can be transcribed and summarized as they happen. - **Turn decisions into examples.** A library of real past cases (the request, the reasoning, the outcome) teaches people and AI systems better than a list of rules. - **Put it where the work happens.** An assistant that answers "how do we usually handle this?" from your own material gets used. A wiki mostly doesn't. - **Keep people in charge of judgment.** The goal is not to replace your experts. It is to make their knowledge outlive any single conversation. ### Why it compounds Captured know-how gets better with every case you add, and it is yours alone. A competitor can use the same AI models, but not your history. That is how an at-risk moat turns into one that deepens. ### Where we come in Capturing know-how is often the first thing we build after the audit, because it protects the business and makes everything else easier to automate. [See the services](/#services). --- ## Customer trust holds, and AI can buy you more time for it https://succession.bot/blog/customer-trust (Holds, 2026-10-12) > People still buy from people they trust. AI can free up time for exactly that. ### Why trust survives When the cost of producing things falls, the cost of choosing rises. Buyers face more options, more content and more polished pitches than ever, and much of it is generated. In that world, a supplier you already trust is worth more, not less. Trust is built from history: promises kept, problems handled well, honest advice when a sale was on the line. AI can't fake a track record. ### Where AI can damage it Trust holds, but it can be spent. The fastest way to lose it is to automate the wrong things: - Generic AI-written emails that sound like everyone else's. - Support bots that stand between customers and a person. - Proposals and content that are obviously produced at volume. - Mistakes made confidently by a system nobody checked. ### The opportunity Most small teams spend a large part of their week on work that doesn't build trust: admin, reporting, chasing information, formatting documents. Moving that work to AI frees time for the work that does: calls, visits, careful answers and following through on promises. - **Automate around the relationship, not inside it.** Let AI prepare the briefing. Let a person have the conversation. - **Use AI to remember.** Context about each customer, surfaced before every interaction, lets a small team feel attentive at scale. - **Be open about what is automated.** Customers accept AI help when it is honest and useful. - **Keep a person one step away.** A clear way to reach a human, always. ### Where we come in When we map your workflows we mark the ones that touch customer trust, and design the automation to protect them. [See the services](/#services). --- ## When building gets cheap, distribution is the moat https://succession.bot/blog/distribution (Holds, 2026-10-12) > When building gets cheap, owning the path to your customers matters more. ### The shift If products get easier to build, more of them get built. Every niche gets more competitors, and many of them will be good enough. The question stops being "can you make it?" and becomes "can you reach the people who need it, and do they already know you?" Distribution is any reliable path to customers that others can't easily copy: an audience that follows you, a partner network, a sales channel built over years, a reputation in a community, being the default name in your niche. ### Why it holds Distribution is built from relationships and attention, and both take time. A competitor can generate a thousand articles in a week, but not the newsletter readers who have trusted you for years, or the reseller who has recommended you for a decade. ### What changes - **Discovery is moving to AI assistants.** More buyers ask an assistant for a recommendation before they visit any website. Being clearly described, easy to cite and well reviewed matters more each year. (This site publishes an [llms.txt](/llms.txt) for exactly that reason.) - **Volume is no longer a signal.** Everyone can publish a lot. Being specific, credible and useful is what gets noticed. - **Small teams can run bigger channels.** AI can draft, personalize and follow up, so a team of five can keep up relationships that used to need a sales department. ### What to do about it - **Know where customers really come from.** Trace your last ten customers back to how they found you. - **Invest in channels you own.** Audience, partners and referrals compound. Ads and marketplace rankings are rented. - **Make your business easy for AI to understand.** Clear descriptions, structured information and genuine reviews help assistants recommend you accurately. - **Deepen before you widen.** Use AI to serve existing relationships better before chasing new ones. ### Where we come in Our audit looks at how customers actually find you and where AI can extend those channels without diluting them. [See the services](/#services). --- ## Proprietary data gets more valuable with every model release https://succession.bot/blog/proprietary-data (Deepens, 2026-10-12) > Data only you have becomes more valuable once AI can put it to work. ### Why it deepens Every company now has access to the same powerful models. That is a great equalizer, and it is also why data matters more: when everyone has the same engine, the difference is the fuel. Your records of jobs, quotes, customer questions, failures, measurements and outcomes are things no public model was trained on. Put to work, they make your AI better at your business than anyone else's. ### What counts More than most companies think: - Years of quotes, and which ones you won or lost. - Support requests and how they were resolved. - Project files, test results and inspection reports. - Customer conversations, notes and emails. - Operational data from machines, logistics or bookings. Much of it is messy and spread across inboxes, spreadsheets and old systems. That's normal. Messy data that only you have beats clean data that everyone has. ### How to put it to work - **Find it before you need it.** Inventory where your data lives and who can reach it. - **Start with one valuable question.** "Which quotes do we usually win?" or "What causes most of our rework?" is a better start than a data lake. - **Keep it yours.** Use tools and settings that don't train other people's models on your data, and keep copies in systems you control. - **Capture more as you go.** Every workflow you automate can also record what happened, so the data keeps growing. ### The catch Data only deepens your moat if you use it. Data sitting untouched in a drive is a liability, not an asset. The companies that pull ahead connect it to daily decisions. ### Where we come in The audit maps the data you already have. The build sprint puts the most valuable part of it to work. [See the services](/#services). --- ## Domain expertise is slow to copy, and AI helps you scale it https://succession.bot/blog/domain-expertise (Deepens, 2026-10-12) > Hard-won knowledge of a niche or regulated field is slow to copy, and AI helps you scale it. ### Why it's hard to copy Some fields take years to learn: reliability engineering, compliance, medical devices, specialized manufacturing, regulated finance, building codes. The knowledge sits in standards, in edge cases, and in the experience of having been wrong before. Newcomers can read the rules, but they don't know which rules bite, which customers are exceptions, or what an auditor will actually ask. ### Why AI deepens it General models know the textbook. They are least reliable exactly where expertise matters: the edge cases, the unwritten conventions, the judgment calls with real consequences. That makes experts more valuable, not less, because someone has to know when the AI is wrong. And the same AI that can't replace an expert can multiply one. An expert who delegates research, drafting, checking and formatting can serve many more clients without lowering the quality of the judgment they're paid for. ### How to scale expertise with AI - **Encode your checklists and standards** into the tools your team uses every day, so less experienced people work more like experienced ones. - **Let AI draft and experts decide.** Reports, assessments and proposals go faster when the expert reviews instead of writing from scratch. - **Build on your own cases.** Your past projects are the best material for your niche, especially once your [know-how is captured](/blog/know-how). - **Turn expertise into products.** Assessments, calculators and tools clients can use directly extend your reach beyond your own hours. ### The risk to avoid Expertise is a moat only while it stays ahead. Fields change, and AI speeds that up. The experts who win use AI to keep learning faster than their field moves. ### Where we come in For expert firms we usually start by finding the work experts do that doesn't need an expert, and hand that to AI. [See the services](/#services).