// Industry pages: eight sectors, all rendered from INDUSTRY_PAGES. // Examples are written in the present tense as things we build. Do not add client names, results, statistics, or past-tense claims of delivered work. // `helps` entries link to a service page by slug. `questions` appears only on the longer pages. const INDUSTRY_PAGES = { 'manufacturing': { title: "Manufacturing", headline: "AI for the business side of the plant.", lede: "You know how to make the product. Quoting, scheduling, purchasing, and reporting often run on spreadsheets and a few people's memory. That side of the business is where we work. We stay off the production line.", situation: [ "Quoting depends on one estimator. Each job has its own spreadsheet. The schedule lives in a few heads, and the morning meeting is where it gets reconciled.", "Many shops have an ERP that handles orders and invoicing well. The schedule, the quotes, and supplier follow-up still happen outside it, in spreadsheets and email.", "Some of your data may be sensitive, such as customer drawings, specifications, or work covered by a contract or regulation. There can be strict rules about where it can be stored and processed, and most off-the-shelf AI accounts don't meet them. Your quoting, scheduling, and purchasing usually aren't sensitive. We draw that line with you and keep the sensitive data on the right side of it.", ], helps: [ { s: 'business-operations', p: "One schedule built from the spreadsheets you have now. Quotes drafted from similar past jobs. Fewer orders retyped from one system into another." }, { s: 'business-insight', p: "Years of quotes, job costs, and customer history connected, so you can ask which jobs make money and why." }, { s: 'organizational-intelligence', p: "Job files, past bids, and estimators' notes connected, so what the estimator knows is written down and your AI tools can use it." }, { s: 'product-supply-chain', p: "Bills of materials, supplier records, and lead times in one record, with flags for end-of-life parts and tariff exposure." }, { s: 'sales-fundraising', p: "Prospect research and outreach for a shop that has sold mostly through relationships. The same research works for finding new suppliers." }, ], examples: [ "We build quoting that learns from your past jobs, so a first-pass price is drafted for your estimator to review.", "We build a live view of every job, material, and supplier commitment across the plant and the office.", "We draw the line between sensitive and everyday data, so controlled work stays on your own systems while the rest can use AI tools.", ], engagementNote: "Most manufacturing work starts with one shared view, usually the schedule or the quote log. Supervisors and estimators learn it on their own jobs as it goes in.", questions: [ "How does a quote get made, and who can make one?", "Where does the schedule live, and who updates it?", "When a customer calls to ask where an order is, who answers, and where do they look?", "Is any of your work covered by a contract or regulation that limits where the data can go?", "Who would own these tools day to day?", ], profile: { typical: "Manufacturers, from single plants to multi-site companies, with an ERP and spreadsheets around it", first: "One shared schedule or quote log", leads: "Amanda Rinderle or Jonas Clark, with your CEO, estimators, and supervisors" }, firstProject: [ "A working schedule or quote log, built from the spreadsheets you use now.", "Training for the supervisors and estimators who use it, on their own jobs.", "A review step, so a person approves any quote or schedule change before it is used.", ], faq: [ { q: "Why not just use ChatGPT or Claude ourselves?", a: "An estimator can ask a chat tool to summarize a spec, and that works. Drafting a quote from your own past jobs takes more. The model has to read the ERP and the job records, be tested against quotes you have already priced, and hand anything unusual to a person. Customer drawings stay out of public tools." }, { q: "What data do you need to see first?", a: "A few dozen past quotes, the spreadsheet the schedule lives in, and a walk through how a job gets from order to shipment. We don't need drawings or specifications. If something in the sample is sensitive, tell us and we will work around it." }, { q: "What about our controlled or customer-owned data?", a: "Before we build, we work out with you which work is controlled. Your contracts and your compliance contact decide that. Controlled work stays on the systems you designate, and everyday quoting and scheduling can use more." }, { q: "Does this replace our estimator?", a: "No. The tool drafts a first-pass price from similar past jobs and the estimator corrects it. Those corrections are recorded, so some of what the estimator knows ends up written down." }, { q: "How do you handle a wrong quote or a bad schedule entry?", a: "Nothing goes to a customer until a person approves it. Before that, we run the tool on jobs you have already priced and compare its numbers with yours. Each draft is logged with the past jobs it drew on, so a bad number can be traced." }, ], closer: { heading: "Tell us how a job moves through your shop.", text: "We stay on the business side of the plant. Tell us where quotes and orders wait, and we will start there.", }, }, 'robotics-hardware': { title: "Robotics and hardware", headline: "Sales and sourcing for teams that would rather be engineering.", lede: "Most hardware companies are engineering teams first. Selling, sourcing, and supplier tracking get added later, by whoever has time. We help build out that side.", situation: [ "Sales and sourcing get staffed after the product works. Often a founder and one or two engineers carry both.", "Many companies start with one customer or one program, such as a government contract, and want a second market without hiring a full sales team.", "A robot or a device can carry hundreds of parts from dozens of suppliers. Parts go end-of-life, lead times move, and a tariff change can reprice a subassembly. Someone has to watch all of it, and usually it's an engineer with other work.", "Design files and some customer data may be export-controlled. Those stay off public AI services. When a tool needs them, we can run it on a model hosted on your own systems.", ], helps: [ { s: 'sales-fundraising', p: "Find markets and companies beyond your first customer, with a brief on each and outreach drafted for you to send." }, { s: 'product-supply-chain', p: "One record of every part, its supplier, lifecycle status, and lead time, with a flag when any of them changes." }, { s: 'organizational-intelligence', p: "Engineering notes, test results, support tickets, and customer feedback connected, so your team and its AI tools can search them by asking." }, { s: 'business-operations', p: "Order intake, quotes, RMAs, and certificates of conformance drafted by tools and approved by a person, instead of written by an engineer." }, ], examples: [ "We build a parts and supplier record that tracks lifecycle, lead times, and risk across every design.", "We build market research that maps new industries and buyers for your technology, with a brief on each.", ], engagementNote: "Hardware teams are short on time, so we start with one market to research or one bill of materials to map, and ask for as little engineering time as we can.", questions: [ "Who buys from you today, and who else should?", "Where is the bill of materials kept, and who maintains it?", "Which of your parts have only one source?", "Is any of your technical data export-controlled?", "Who does the sourcing, and what are they not doing while they do it?", ], profile: { typical: "Robotics and hardware companies, from engineering-led startups to established product lines", first: "One market to research, or one bill of materials to map", leads: "Amanda Rinderle or Jonas Clark, with your CEO, using as little of your engineers' time as possible" }, firstProject: [ "Either a research brief on one new market, listing the companies and buyers in it, or a mapped bill of materials with each part matched to its supplier and lifecycle status.", "Sources for every claim in the brief or the record, so an engineer can check them.", "A review step before any outreach goes out, and a walkthrough with whoever owns sales or sourcing.", ], faq: [ { q: "Why not just use ChatGPT or Claude ourselves?", a: "Your engineers probably already do, and for a quick question that's fine. Mapping a new market or tracking hundreds of parts needs your customer history and your bill of materials, sources someone can check, and a record that rechecks itself on a schedule. We build that, test it on markets and parts you already know, and keep design files out of public tools." }, { q: "What data do you need to see first?", a: "A bill of materials in whatever form it exists, or a list of current customers and what they buy. We don't need design files. Tell us what is off limits before you send anything." }, { q: "What about export-controlled files?", a: "They stay off public AI services, and most of this work can be done without them. When a project does need them, we can run the model on your own systems. We decide with you where each tool runs and what it can see. We are not export-control counsel, so your counsel decides what is controlled." }, { q: "How much engineering time do you need?", a: "As little as we can manage. Someone walks us through the bill of materials or the customer list, and later reviews what we produce." }, { q: "How do you handle mistakes in the research?", a: "Every brief and every flag cites its sources, and nothing goes to a prospect until a person has read it. We test on markets and parts you know well and tell you where the tool tends to be wrong. In market research, company size and who makes the buying decision are the usual weak spots." }, ], closer: { heading: "Tell us who buys from you, and who else should.", text: "The first conversation is twenty minutes and free. We ask about your customers and your parts, and not much else. Bring a bill of materials if you have one handy.", }, }, 'consumer-brands-retail': { title: "Consumer brands and retail", headline: "One view of the business across every channel.", lede: "A consumer brand runs on a storefront, a marketplace or two, an inventory system, an email platform, and several ad accounts. Most questions about the business need data from three of them at once.", situation: [ "Orders are in the store platform, stock is in the 3PL portal, ad spend is spread across ad accounts, and margin is in the accounting software.", "Catalogs outpace the teams that keep them. Titles, sizes, and colors drift apart between the website and the marketplaces, and nobody notices until a listing has a problem.", "Asking which products are slowing, and why, costs someone an afternoon of exports.", "Costs move faster than the team can track, from tariffs, freight rates, or a supplier's new terms.", ], helps: [ { s: 'business-operations', p: "Catalog cleanup, listings, returns handling, and customer service replies drafted for someone to approve." }, { s: 'business-insight', p: "Orders, products, customers, and ad data in one place, so a merchandiser can ask which products are slowing and see the likely reason." }, { s: 'product-supply-chain', p: "Supplier, landed cost, and country of origin tied to each product, with flags for tariff exposure." }, { s: 'sales-fundraising', p: "Research on wholesale accounts and retail buyers, with outreach drafted for the team." }, ], examples: [ "We build one view of the whole business, from orders, inventory, returns, and ad spend, that explains why a product is moving.", "We build monitoring for costs, tariffs, and supplier changes that tells you which products are affected.", ], engagementNote: "Most consumer work starts with the product data, because listings, feeds, and reporting all depend on it. Cleaning it up is a project. Keeping it clean is monthly support.", questions: [ "Where do your orders and product data live?", "Which question takes the longest to answer today?", "How many channels do you sell through, and who keeps the listings in sync?", "Which suppliers and costs matter most to your margin?", "Who on the team would use these tools?", ], profile: { typical: "Brands selling online, through marketplaces, in stores, and to wholesale accounts", first: "Cleaned product data and one view across channels", leads: "Amanda Rinderle or Jonas Clark, with your CEO and whoever runs merchandising and operations" }, firstProject: [ "A cleaned product catalog, with titles, sizes, and colors consistent between your site and your marketplaces.", "One view that joins orders, inventory, returns, and ad spend for the questions you ask most often.", "An approval step before any listing change goes live.", "Notes on where the data is thin, for example ad accounts that don't break out spend by SKU.", ], faq: [ { q: "Why not just use ChatGPT or Claude ourselves?", a: "A chat tool writes a decent product description, and it can make sense of one export. Asking why a product slowed down means joining the store, the 3PL, the ad accounts, and the books, and a chat window can't see any of them. We connect them, check the answers against numbers you already trust, and route listing changes through an approval step that logs who changed what." }, { q: "What data do you need to see first?", a: "A product export from your store, plus one or two reports someone rebuilds by hand each week. We start with product and order data and ask before going further." }, { q: "What happens to our customer data?", a: "Product and order-level data covers most catalog and reporting work, so that is where we start. If a project needs customer names or emails, we work out with you first what is included, where it runs, and how long it is kept." }, { q: "Does this replace our customer service team?", a: "The tool looks up the order and drafts a reply. Your team edits and sends it, and they decide when a customer needs more than a quick answer." }, { q: "How do you handle mistakes in listings or reports?", a: "A listing change waits for approval before it goes live. Each number in a report links back to the records behind it. We test against figures you already know. When they don't match, the cause is often in the data, such as a return logged in one system and not the other." }, ], closer: { heading: "Tell us which question takes the longest to answer.", text: "The first conversation is twenty minutes and free. Tell us where your orders and product data live, and we will talk through which pieces to join first.", }, }, 'wholesale-distribution': { title: "Wholesale and distribution", headline: "Faster quotes, cleaner item data, and every rebate claimed.", lede: "A good distributor knows which customer gets which price and which part can stand in for another. Most of that knowledge sits with a few people and in years of email, next to an ERP that was configured long ago.", situation: [ "The day is a stream of quotes, reorders, substitutions, and lead-time questions. They arrive by email, phone, and EDI, in no standard format.", "Every manufacturer you carry sends data in its own format with its own part numbers, so one item can sit under several identifiers.", "Customer contract pricing and vendor rebate programs are often tracked in spreadsheets.", "The ERP records transactions well. Questions that cross customers, margins, and suppliers still end in a spreadsheet.", ], helps: [ { s: 'business-operations', p: "Read incoming quote requests and purchase orders, look up pricing and availability, and draft the response for a rep to send." }, { s: 'business-insight', p: "Customer, order, and margin history in one place, so a sales manager can ask which accounts are slipping." }, { s: 'product-supply-chain', p: "Match the same item across manufacturers' files, find alternates for allocated or end-of-life parts, and track supplier risk." }, { s: 'sales-fundraising', p: "Research on prospects and lapsed accounts, with outreach drafted for the team." }, ], examples: [ "We build quoting that reads any request, from a PDF to a one-line email, and drafts the price with the customer's terms.", "We build one clean product and supplier record across every manufacturer file, with alternates ready when stock runs out.", ], engagementNote: "Distribution work usually starts with the item and supplier record, because quoting and sourcing both depend on it.", questions: [ "How do quote requests reach you, and how long does a reply take?", "How many identifiers does one item carry across your suppliers?", "How do you track vendor rebates today?", "Which accounts do you worry about losing?", "What does your ERP do well, and what does it leave to spreadsheets?", ], profile: { typical: "Distributors carrying many manufacturers' lines, from regional houses to national networks", first: "A single item and supplier record", leads: "Amanda Rinderle or Jonas Clark, with your CEO, sales desk, and purchasing" }, firstProject: [ "One item and supplier record built from a few manufacturers' files, with the same item matched under each of its identifiers.", "Alternates listed for items that are allocated or near end of life.", "A list of the matches a person confirmed, and the files that didn't match cleanly.", "A review step before any quote or order built on the record goes out.", ], faq: [ { q: "Why not just use ChatGPT or Claude ourselves?", a: "A rep can clean up a reply in a chat tool. Quoting is harder. A request comes in as a PDF, a photo of a handwritten list, or one line in an email, and the price depends on that customer's contract, what is in stock, and whose part number they used. We connect the model to your ERP and pricing, test it against quotes you have already sent, route unclear requests to a person, and log each response." }, { q: "What data do you need to see first?", a: "One or two manufacturer files and a batch of recent quote requests, mess included. If there are customer details you would rather hold back, take them out first." }, { q: "Do we have to change our ERP?", a: "No. It handles transactions well, so we leave those to it. We connect to it and work on what has ended up in spreadsheets and email." }, { q: "Does this replace our reps or inside sales?", a: "No. The tool reads the request, looks up pricing and stock, and drafts a reply. A rep checks it and sends it, and customers keep dealing with the rep they know." }, { q: "How do you handle a wrong price or a wrong match?", a: "A rep approves each reply, so a person sees every price before the customer does. Item matches the tool isn't sure of go to someone to confirm. We test on past quotes and keep a list of the request types it handles badly, such as substitutions the customer didn't ask for." }, ], closer: { heading: "Tell us how quote requests reach you.", text: "Describe a typical day of quotes and reorders. We will want to know how many identifiers one item carries.", }, }, 'construction-trades': { title: "Construction and trades", headline: "Get more from the software you already pay for.", lede: "Contractors run on estimates, schedules, subs, and paperwork. The software you already pay for is adding AI features. We help you turn on the useful ones, connect them, and train the office and the field to use them.", situation: [ "The office handles bids, RFIs, submittals, change orders, pay applications, and lien waivers, much of it by hand and by email.", "A few experienced people know how to estimate a job, which subs show up, and what a change order will cost by the time it closes.", "Project management and accounting platforms are adding AI features. Few contractors have the time to turn them on, configure them, or connect the project platform to accounting.", "So most of our work here is setting up and connecting software you already own. We suggest new tools only when those run out.", ], helps: [ { s: 'business-operations', p: "Set up and connect the platforms you already own, and draft recurring paperwork such as change orders, daily logs, and pay applications for review." }, { s: 'business-insight', p: "Past bids, job costs, and subcontractor history in one place the estimators can ask." }, { s: 'product-supply-chain', p: "For firms that buy materials in volume, a record of suppliers, lead times, and price changes." }, { s: 'sales-fundraising', p: "Research on upcoming projects and the owners and GCs who award them, with outreach drafted for the team." }, ], examples: [ "We build estimating that draws on every past bid and job cost, so each new estimate starts from the company's whole history.", "We build the connection between project management and accounting, so job costs, field reports, and billing stay in step.", ], engagementNote: "We start with the platforms you already own and the paperwork that repeats on every job. Setup is usually a fixed-scope project, and training and changes afterward suit monthly support.", questions: [ "Which platforms do you already pay for, and which features are switched on?", "How does an estimate get made, and who can make one?", "What paperwork repeats on every job?", "Where do field notes and photos end up?", "Who in the office would use these tools?", ], profile: { typical: "General and specialty contractors with an office team and a project platform", first: "Setting up and connecting the platforms you already own", leads: "Amanda Rinderle or Jonas Clark, with your CEO, the office, and the field" }, firstProject: [ "The project and accounting platforms you already pay for, set up and connected so job costs and billing stay in step.", "Drafts of one piece of recurring paperwork, such as change orders or daily logs, for a project manager to review.", "A record of which features are switched on and how each one is configured.", "Training for the office and the field, and a named owner on your side.", ], faq: [ { q: "Why not just use ChatGPT or Claude ourselves?", a: "For a one-off letter, use whatever chat tool you like. A change order or a pay application pulls from the job budget, the contract, and what happened in the field, and that data sits in your project platform. Most of our work here is switching on the AI features in the platforms you already own, connecting them to each other, and putting a review step in front of anything that gets sent." }, { q: "What do you need to see first?", a: "Your list of platforms and which features are switched on, plus a few examples of paperwork that repeats on every job. Past bids and job costs help if estimating is part of the work." }, { q: "Do we need to buy new software?", a: "Usually not. Project and accounting platforms are adding AI features faster than most offices can turn them on. We suggest something new only when what you have can't do the job." }, { q: "Does this replace our estimators or project managers?", a: "No. The tools draft and organize. Your estimators and project managers check the numbers and sign the documents, as they do now." }, { q: "What happens to our project and financial data?", a: "We start inside the platforms that already hold it and use their permissions. If anything has to leave them, we decide with you where it goes and who can reach it before we build." }, ], closer: { heading: "Tell us which platforms you already pay for.", text: "The first conversation is twenty minutes and free. Tell us what you own and which paperwork repeats on every job. That is usually where we start.", }, }, 'insurance-banking': { title: "Insurance and banking", headline: "Less re-keying, with client data kept inside your controls.", lede: "Agencies, brokerages, carriers, and banks keep years of client files, policies, and correspondence across several systems. Staff move information between them by hand.", situation: [ "At an agency, service staff re-key data between the management system, carrier portals, and email, and answer the same coverage questions many times a week. Policy checking, certificates, and renewals fill much of the day.", "At a bank, relationship managers and credit staff assemble loan files and client histories from the core system, spreadsheets, and email.", "Client data is regulated, and carriers, examiners, and auditors will all ask where it goes. We settle what a tool may see and where it runs before we build anything.", "Large carriers and banks often have AI teams of their own. We work alongside them on specific workflows, and where there is no such team, we fill that role.", ], helps: [ { s: 'business-operations', p: "Tools that read carrier documents, renewal notices, and applications, fill in the fields, and send exceptions to a person." }, { s: 'organizational-intelligence', p: "Client files and correspondence connected with access controls, so a service rep can find a policy history by asking." }, { s: 'sales-fundraising', p: "Prospect research and renewal outreach, drafted for producers and relationship managers to send." }, ], examples: [ "We build document review that checks policies and applications against each other and flags what a person needs to see.", "We build a client brief for every relationship, assembled from your own records and inside your controls.", ], engagementNote: "In regulated work we agree on the data boundary first, meaning what the tools may see, where they run, and who reviews what they produce. Then we build one workflow.", profile: { typical: "Agencies, brokerages, carriers, and banks", first: "One document workflow, inside limits you set on what the tools can see", leads: "Amanda Rinderle or Jonas Clark, with your CEO and your operations and compliance staff" }, firstProject: [ "A working tool on one document workflow, such as renewals or application intake, with exceptions sent to a person.", "Testing against files your staff have already completed, before the tool touches live work.", "A log built into the tool of what it produced and who checked it.", "Training for the staff who use it.", ], faq: [ { q: "Why not just use ChatGPT or Claude ourselves?", a: "Rewording a general email in a chat tool is fine. Client details are another matter, and they shouldn't go into a personal account. Renewals and loan files need the tool connected to your management or core system, tested against files you have already handled, and set up to send anything uncertain to a person. We build in a log of what it read and what it produced." }, { q: "What data do you need to see first?", a: "A few sample documents, such as a renewal notice or a blank loan file template, with client details removed. We talk through what the tools may see before anything real changes hands." }, { q: "What happens to client data?", a: "Before we build, we decide with you what a tool may see, where it runs, and how long anything is kept. Your carriers, examiners, and auditors may have their own requirements, and your compliance contact decides what applies. We are not a compliance firm and do not advise on regulation." }, { q: "Does this replace our service staff or relationship managers?", a: "No. The tool reads documents and fills in fields. Your staff handle the exceptions and the clients." }, { q: "How do you handle mistakes?", a: "Anything the tool is unsure of goes to a person, and a reviewer checks the rest before it is used. We test on files your staff have already completed and tell you which cases it handles poorly, such as endorsements with unusual wording." }, ], closer: { heading: "Tell us what your tools may and may not see.", text: "We start with the limits on what the tools can see, then pick one workflow to build inside them. The first conversation is twenty minutes and free.", }, }, 'healthcare': { title: "Healthcare and medical devices", headline: "Fewer days lost to component changes and records searches.", lede: "Medical device and life sciences companies manage complex bills of materials, long supplier lists, and quality records that regulators can ask to see. Provider organizations carry a heavy administrative load. We work on the operations side of both, and leave clinical and regulatory judgment with your people.", situation: [ "A device can carry hundreds of components from dozens of suppliers. Parts go end-of-life, suppliers change their processes, and each change has to be assessed and documented before it reaches production.", "Quality, complaint, and design records sit in several systems. Answering a question that crosses them, such as which complaints involve a given component, takes someone days of searching.", "Regulated documents, from change assessments to supplier files, are assembled by hand from records that already exist elsewhere.", "In provider organizations, scheduling, billing, payer rules, and vendor paperwork take a large share of staff time.", "Your quality, regulatory, and compliance teams decide what the rules require. We aren't a regulatory, clinical, or compliance firm, and we don't advise on any of them.", ], helps: [ { s: 'product-supply-chain', p: "One record of every component, supplier, and lifecycle status across your products, with a flag when a part or supplier changes and a list of the products it touches." }, { s: 'organizational-intelligence', p: "Quality, complaint, and design records connected where they live, so staff can search them and see the source behind every answer." }, { s: 'business-operations', p: "First drafts of recurring documents, such as supplier change assessments, assembled from your own records for a qualified person to review and sign. For providers, invoices, supply orders, and scheduling." }, { s: 'sales-fundraising', p: "Research on hospital systems, group purchasing organizations, and the people who make buying decisions, with outreach drafted for your team." }, ], examples: [ "We build component and supplier monitoring that shows which products a change touches before anyone has to ask.", "We build searchable access to quality and complaint records, so an engineer finds every related record and its source in minutes.", "We build back-office workflows for provider organizations, from invoices to supply orders, without touching patient records.", ], engagementNote: "Device and life sciences work usually starts with the component and supplier record, because quality, sourcing, and change control all depend on it. With providers, we start with one piece of office work and a clear line around patient data.", profile: { typical: "Medical device and life sciences manufacturers, and provider organizations", first: "A component and supplier record, or one piece of office work", leads: "Amanda Rinderle or Jonas Clark, with your CEO and your operations, quality, or office leads" }, firstProject: [ "A working tool on one workflow, such as the component and supplier record or a recurring document.", "Testing against records your team has already reviewed, before the tool touches live work.", "A log of what the tool produced and who approved it, and a review step before anything is used.", "Training for the people who use it, and a named owner on your side.", ], faq: [ { q: "Why not just use ChatGPT or Claude ourselves?", a: "For writing with no product or patient data in it, a chat tool is fine. Tracing a component change across your products, or finding every complaint tied to one part, needs your ERP, quality system, and supplier records connected, tested against answers your team already knows, and set up so a person reviews the output. We build that, and keep controlled data where your teams say it belongs." }, { q: "What data do you need to see first?", a: "A bill of materials and the supplier list behind it, or a sample of records with anything sensitive removed. For providers, nothing involving patients. We agree on what is in scope before anything is shared." }, { q: "Is this part of our validated quality system?", a: "That is your quality team's decision. Many first projects sit outside it, as tools that help people find and assemble information they then review. If a tool needs to be part of the quality system, we work with your team on what that requires before we build." }, { q: "Will you touch patient records?", a: "Not in a first project. Rules for patient data are your compliance officer's call, and we do not advise on them." }, { q: "How do you handle mistakes?", a: "A qualified person reviews anything the tool drafts before it is used. We test on records your team has already reviewed and tell you which cases the tool handles poorly, such as parts listed under more than one supplier number." }, ], closer: { heading: "Tell us where a component change or a records search costs the most time.", text: "The first conversation is twenty minutes and free. Bring a bill of materials or one recurring document, and we will talk through where to start.", }, }, 'professional-services': { title: "Professional services", headline: "More of your people's hours spent on billable work.", lede: "Law, accounting, engineering, and consulting firms sell the judgment of their people. An hour spent assembling a proposal or hunting through old files is usually an hour nobody bills.", situation: [ "Work moves through drafts, review, and sign-off. Much of what the firm knows sits in past documents that nobody can search well.", "In a licensed profession, the person who signs the work is responsible for it, however the draft was produced. Where AI helped with a draft, many firms want a record of who reviewed it and what they checked.", "Client confidentiality limits which tools can see which files, and in some firms ethical walls do too.", ], helps: [ { s: 'organizational-intelligence', p: "Past proposals, reports, and precedent documents connected, so a new hire or an AI tool can draw on them and cite the source." }, { s: 'business-operations', p: "Proposal assembly, status reports, and client intake. Where AI helps draft work product, we set up a simple log of which tool was used, what was checked, and who signed off." }, { s: 'sales-fundraising', p: "Prospect research and outreach drafts for firms that win work through business development." }, ], examples: [ "We build drafting and research tools with review steps and a record of who checked what.", "We build a searchable knowledge base of your past work, with client access rules built in.", ], engagementNote: "We start with one document-heavy workflow, such as proposals or reports, and set up the review step before anything else.", profile: { typical: "Law, accounting, engineering, and consulting firms", first: "One document-heavy workflow, with the review step set up first", leads: "Amanda Rinderle or Jonas Clark, with your managing partner and whoever owns the document process" }, firstProject: [ "A working tool on one document-heavy workflow, such as proposals or reports.", "A review step, with a log of which tool was used, what was checked, and who signed off.", "Access rules so the tool sees only the files each person is cleared to see.", "Training for the people who draft and the people who review.", ], faq: [ { q: "Why not just use ChatGPT or Claude ourselves?", a: "Many of your people already do, and for work with no client information in it that is often fine. Use across the firm raises other questions. The tool needs your past work to draw on, access rules that respect client matters, and a record of who reviewed each draft, since the person who signs is still responsible. We set that up and test it on documents you have already finished." }, { q: "What data do you need to see first?", a: "A handful of past proposals or reports, and a description of how a document moves from draft to sign-off. Client-identifying details can come out first." }, { q: "What about client confidentiality and ethical walls?", a: "We set access so the tool sees only the files a given person is allowed to see, and walls can be set up the same way. We work out with you first what is included, where it runs, and how long anything is kept. Your professional rules decide what is permitted, and we do not advise on them." }, { q: "Does this replace our associates or junior staff?", a: "No. It does the assembly and the first draft. Your associates still review, edit, and sign." }, { q: "How do you handle mistakes?", a: "Every draft has a named reviewer, and the log shows what they checked. We test on finished work and tell you which kinds of errors to watch for, such as a reference to the wrong version of a document." }, ], closer: { heading: "Tell us which document takes the longest to assemble.", text: "The first conversation is twenty minutes and free. We start with one workflow and set up the review step before we build anything.", }, }, 'private-equity': { title: "Venture capital and private equity", headline: "AI systems inside your portfolio companies, and a clear read on the next one.", lede: "We work inside the companies investors back, from venture-funded startups to private equity portfolio companies, building the systems that take hours out of quoting, ordering, and reporting without adding headcount. For deal teams, we also give an outside read on how a target runs, what its data can support, and how much of its AI is real.", situation: [ "Operating partners carry a long list of improvements and little capacity to build them, while management teams are busy running the business.", "Venture-backed companies add customers and staff before their operations catch up. Sales, support, and reporting often run on whatever the founders set up in the first year.", "After an acquisition or a roll-up, the data sits in several ERPs and CRMs that do not match, so combined reporting depends on someone rebuilding it every quarter.", "Diligence moves on a deadline. In a buyout, the questions are about the data, the systems, and how work gets done. In a venture deal, they are often about the technology itself: whether the AI in the pitch is the company's own and whether it holds up.", ], helps: [ { s: 'business-operations', p: "Order, quote, support, and reporting workflows inside a portfolio company, built to run on the systems it already has." }, { s: 'business-insight', p: "One view of customers, orders, and costs across a company, or across several acquired companies, with the source records behind every number." }, { s: 'ai-strategy-roadmap', p: "A short, ranked plan of where AI can improve a company and what its data can support today, for the operating partner, the founders, or, in diligence, the deal team." }, { s: 'sales-fundraising', p: "Prospect research and follow-up for a portfolio company with a thin sales function." }, ], examples: [ "We build a quoting or ordering workflow once and set it up across portfolio companies with similar processes.", "We connect the ERPs and CRMs of acquired companies into one reporting view, with every number traceable to its source.", "In diligence, we read how a target's operations, data, and AI hold up, and tell the deal team what is real and what its records can support.", ], engagementNote: "Work usually starts with one portfolio company and one workflow. We look at how the business runs, pick the workflow that takes the most hours, and build it.", profile: { typical: "Venture and private equity firms, operating partners, and the companies they back", first: "One workflow inside a portfolio company, or a read on a target for a deal team", leads: "Amanda Rinderle or Jonas Clark, with the operating partner or CEO and the company's managers" }, firstProject: [ "A working system on one workflow with a clear payoff, such as quoting, ordering, or reporting.", "A written read on the company: how work flows, where the data lives, and what to build next.", "A review step, so a person approves anything that leaves the building or changes a record.", "Training for the managers who will run it, and notes on what it can and cannot do.", ], faq: [ { q: "Why not just use ChatGPT or Claude ourselves?", a: "For a deal team's one-off questions, a chat tool is often enough. It does not connect to a company's ERP, hold one view across several acquired businesses, or keep confidential deal and company data inside agreed limits. We build those connections and test them on the company's own records." }, { q: "Do you work for the fund or for the company?", a: "Either, and often both. A fund, an operating partner, or a founder commissions the work, and the company's team uses what we build. We settle who the client is before we start." }, { q: "Can you assess a startup's AI?", a: "Yes. We look at what the product does, what data it depends on, and how much is the company's own work as opposed to a general model behind an interface, and we write it up for the deal team. We work only from what the company and the deal team choose to share." }, { q: "What about confidential deal information?", a: "Tell us what is off limits before anything is shared. We work inside those limits, and for diligence we use only the data the deal team chooses to provide." }, { q: "How does this fit a hold period or a funding runway?", a: "We scope the work around one workflow at a time, so value does not wait on a long program. We agree on the sequence with the operating partner or the founders." }, { q: "Does this replace management?", a: "No. The work takes on reporting, quoting, and data assembly. Management still makes the calls and approves the output." }, ], closer: { heading: "Tell us about the company, or the deal.", text: "The first conversation is twenty minutes and free. We usually start with one portfolio company and one workflow, and we can also start with a deal.", }, }, }; function IndustryPage({ setRoute, slug }) { const page = INDUSTRY_PAGES[slug]; if (!page) return null; const serviceLabel = (s) => (SERVICES.find(x => x.slug === s) || {}).label || s; return (
The situation

The situation in your sector.

{page.situation.map((t, i) =>

{t}

)}
Services

Where AI fits in {page.title.toLowerCase()}.

({ title: serviceLabel(h.s), body: h.p, link:
About this service
, }))} />
{page.questions && (
The first conversation

What we ask you about.

    {page.questions.map(q =>
  • {q}
    • )}
)}
); } Object.assign(window, { IndustryPage, INDUSTRY_PAGES });