--- title: "Keep a company decision log so AI strategy work starts from current facts" description: "Log decisions, investor feedback, customer insights, hires, competitors, and metrics in files so Claude preps meetings from your startup's current state." canonical: "https://scalewithsearch.com/articles/ai-for-startup-founders" date: "2026-01-28" modified: "2026-10-02" --- ## Site navigation - [Scale With Search](https://scalewithsearch.com/) - Real estate - Real estate - [Real estate](https://scalewithsearch.com/for/real-estate) - Work - Start here - [Send your brief](https://scalewithsearch.com/work#send-your-brief) - [Prepare your six-question brief](https://scalewithsearch.com/work#prepare-your-six-question-brief) - Build - [Site build, content library with SEO, signal desk](https://scalewithsearch.com/work) - For your business - Trades and home services - [Auto body and collision shops](https://scalewithsearch.com/for/auto-body-and-collision-shops) - [Foundation and home repair contractors](https://scalewithsearch.com/for/foundation-and-home-repair) - [Garage door and fencing contractors](https://scalewithsearch.com/for/garage-door-and-fencing-contractors) - [HVAC contractors](https://scalewithsearch.com/for/hvac-contractors) - [Janitorial and commercial cleaning companies](https://scalewithsearch.com/for/janitorial-and-commercial-cleaning) - [Locksmiths](https://scalewithsearch.com/for/locksmiths) - [Moving companies](https://scalewithsearch.com/for/moving-companies) - [Pest control companies](https://scalewithsearch.com/for/pest-control-companies) - [Plumbing and electrical contractors](https://scalewithsearch.com/for/plumbing-and-electrical-contractors) - [Restoration and water or fire damage companies](https://scalewithsearch.com/for/restoration-and-water-fire-damage) - [Roofing companies](https://scalewithsearch.com/for/roofing-companies) - [Towing companies](https://scalewithsearch.com/for/towing-companies) - [Tree services and landscaping companies](https://scalewithsearch.com/for/tree-services-and-landscaping) - [Solar installers](https://scalewithsearch.com/for/solar-installation) - [General contractors](https://scalewithsearch.com/for/general-contractors-and-construction) - [Paving, concrete, and flooring contractors](https://scalewithsearch.com/for/paving) - Practices and professional services - [Bookkeeping and tax practices](https://scalewithsearch.com/for/bookkeeping-and-tax-practices) - [Dental practices](https://scalewithsearch.com/for/dental-practices) - [Family and criminal defense law firms](https://scalewithsearch.com/for/family-and-criminal-defense-law-firms) - [Med spas and aesthetics practices](https://scalewithsearch.com/for/med-spas-and-aesthetics) - [Personal injury law firms](https://scalewithsearch.com/for/personal-injury-law-firms) - [Veterinary clinics](https://scalewithsearch.com/for/veterinary-clinics) - [Gyms and fitness studios](https://scalewithsearch.com/for/fitness) - [Therapy and outpatient health practices](https://scalewithsearch.com/for/therapy-and-outpatient-health) - [Medical billing companies](https://scalewithsearch.com/for/medical-billing) - [Insurance agencies](https://scalewithsearch.com/for/insurance-agencies) - [Financial advisors](https://scalewithsearch.com/for/financial-advisors) - [Property management companies](https://scalewithsearch.com/for/property-management) - [Recruiting and staffing agencies](https://scalewithsearch.com/for/recruiting-and-staffing) - [Architects and interior designers](https://scalewithsearch.com/for/architects-and-interior-designers) - [Logistics and supply chain companies](https://scalewithsearch.com/for/logistics-and-supply-chain) - Agencies, MSPs, and manufacturing - [IT and managed service providers](https://scalewithsearch.com/for/it-and-managed-service-providers) - [Machine shops and precision manufacturers](https://scalewithsearch.com/for/machine-shops-and-precision-manufacturing) - [Marketing agencies and freelancers](https://scalewithsearch.com/for/marketing-agencies-and-freelancers) - [SEO agencies and consultants](https://scalewithsearch.com/for/seo-agencies-and-consultants) - [Small manufacturers and fabricators](https://scalewithsearch.com/for/small-manufacturers-and-fabricators) - Restaurants, shops, studios, and nonprofits - [Restaurants and hospitality businesses](https://scalewithsearch.com/for/restaurants-and-hospitality) - [Retail stores and ecommerce sellers](https://scalewithsearch.com/for/retail-and-ecommerce) - [Photographers, event planners, and travel agents](https://scalewithsearch.com/for/photographers) - [Churches and nonprofits](https://scalewithsearch.com/for/churches-and-nonprofits) - [All industries](https://scalewithsearch.com/for/) - Learn - For your office - [Office job guides](https://scalewithsearch.com/guides/) - [Browser calculators](https://scalewithsearch.com/tools/) - Start here - [How it works](https://scalewithsearch.com/how-it-works) - [Free Starter Kit](https://scalewithsearch.com/kit/business-memory-starter-kit.zip) - [Synthetic specimen](https://scalewithsearch.com/specimen/working-session-specimen.zip) - Guides - [The Complete Guide to Business Memory for AI Agents](https://scalewithsearch.com/articles/business-memory-for-ai-agents-guide) - [The Complete Small-Business Guide to AI Agent Governance](https://scalewithsearch.com/articles/ai-agent-governance-guide-small-business) - [The Complete Guide to Leaving Vendor AI Memory](https://scalewithsearch.com/articles/leaving-vendor-ai-memory-guide) - Articles by cluster - [Business memory](https://scalewithsearch.com/articles/business-memory-for-ai-agents-guide) - [Agent governance](https://scalewithsearch.com/articles/ai-agent-governance-guide-small-business) - [Migration and ownership](https://scalewithsearch.com/articles/leaving-vendor-ai-memory-guide) - For machines - [llms.txt](https://scalewithsearch.com/llms.txt) - [llms-full.txt](https://scalewithsearch.com/llms-full.txt) - [Machine view](https://scalewithsearch.com/?view=machine) - Company - Evidence - [Proof](https://scalewithsearch.com/proof) - Company - [About](https://scalewithsearch.com/about) # Keep a company decision log so AI strategy work starts from current facts. You have an investor meeting on Thursday. Your AI assistant does not know your business model or your customer acquisition cost. It does not know that you changed direction three months ago and the old pitch deck no longer applies. Each strategy session opens with ten minutes of explanation. A founder holds more context than most people handle in a year: roadmap, positioning, investor talks, hiring, customer feedback, financial projections, and partnership discussions. None of it persists in AI tools that forget between sessions. So you re-explain the company, and the useful work waits. The assistant becomes a writing tool, not a thinking partner, because it does not know enough to help you think. ## Why a startup needs a standing record Startups change fast. What was true last month may not hold now. The product changes, the positioning shifts, and the fundraising plan adjusts to investor feedback. An assistant that does not track those changes gives outdated or generic advice. Strategic decisions need their history. Why did you choose this pricing model? What did early customers say about the features you cut? Which growth channels showed traction? That context sits in Slack threads, meeting notes, and your memory, not in a form an assistant can use. Investor relationships build over months. You pitch dozens of investors, collect feedback, refine the pitch, and return with new traction. If the assistant does not remember who asked what, you rebuild that context before each meeting. Hiring needs continuity too. You define the role, interview, refine the requirements, and make offers over weeks or months. If the assistant forgets the last ten interviews, it cannot help with the eleventh. ## Set up the record files Claude Code reads `CLAUDE.md` from the folder where you start it. Put the company summary there, and name one file for each kind of record: ```text ## Company one_line:: [what you sell, to whom, and why now] stage:: [pre-seed | seed | Series A]; last updated 2026-09-20 current_positioning:: positioning.md (supersedes the pitch deck dated before 2026-06-15) ## Records decisions.md decision, date, reason, evidence, owner, status roadmap.md priorities, cuts, and the reason for each investors.csv one row per conversation: date, fund, questions, concerns, next step customers.md interview takeaways with date and segment hiring/ one file per open role: criteria, questions, finalists competitors.md dated entries: launch, price change, positioning move metrics.csv monthly: revenue, burn, CAC, LTV, churn, growth partnerships.md partner, goal, blockers, commitments, dates ``` The `decisions.md` file is the core. Each entry records what you decided, when, why, which evidence supported it, and whether a later decision replaced it. The other files feed it. A synthetic entry: ```text ## 2026.06.15: sell to teams only decision:: sell annual team contracts; close self-serve signups reason:: self-serve accounts churned at three times the team rate for two quarters evidence:: metrics.csv rows 2026.01 to 2026.06; 14 interviews in customers.md, segment "ops teams" owner:: CEO status:: active; replaces the 2026.02.03 entry "launch self-serve" revisit:: 2026.12.01 ``` Now prepare Thursday's meeting from the log. Ask, "List every decision since our last call with this fund, with the reason for each." Claude reads the date of the last call in `investors.csv` and the entries after that date in `decisions.md`. It also lists the questions the fund asked last time. The prep note then answers each earlier question with a dated decision and its evidence. The old pitch deck does not enter the answer, because the log marks it as replaced. ## Product strategy and the current product record A roadmap says what you plan. A current-product file also says what you removed. Keep it beside the decision log, with dated customer, pricing, and pivot records. These four templates are synthetic planning examples. Every date, usage rate, conversion rate, and outcome below is illustrative. The 6x lifetime value, 27-percentage-point activation change, and 12/9/7/2/1 percent message conversions are invented, not reported startup results. ```markdown ## Current product (updated 2026-01-28) product:: project management for remote teams, with AI summaries ### Core features - Task boards with generated standup summaries - Slack integration pulls threads into project context - Weekly progress reports from activity logs ### Added (2026-01-20) - Time-zone-aware notifications (no 3 a.m. pings) - Voice-to-task (record a thought; a task appears) ### Deprecated (2026-01-15) - Time tracking (users did not care; added complexity) - Gantt charts (nobody used them) ### Roadmap - Mobile app (Q2 2026) - Calendar integration (Q2 2026) - Public API (Q3 2026) ``` ```markdown ## Who we serve (updated 2026-01-22) primary_icp:: remote-first teams, 5 to 20 people, no dedicated PM previous_icp:: freelancers who manage client projects (before 2026-01) why_we_moved:: freelancers churn after 2 months; teams stay for years; illustrative lifetime value is 6x higher ### Best-fit customers - Design agencies (3 to 10 people, distributed) - Early-stage SaaS companies (remote engineering teams) - Consulting firms (project work, several clients) ### Pains we solve 1. Context lost in Slack threads: threads flow into project context 2. Async updates get buried: daily digests 3. Status meetings waste time: standups written from activity ### Illustrative message conversions (synthetic) - "Your standup writes itself" (12%) - "Stop losing context in Slack" (9%) - "Project management that actually fits remote work" (7%) ### Lower illustrative message conversions (synthetic) - "AI-powered project management" (too generic, 2%) - "The future of work" (too vague, 1%) ``` ```markdown ## Pricing (updated 2026-01-25) ### Current tiers (prices: see pricing page; single source) - Free: 3 users, 10 projects, basic AI summaries - Team: up to 15 users, unlimited projects, full AI features - Business: up to 50 users, priority support, custom integrations ### Recent changes - 2026-01-25: raised the Team price (willingness-to-pay test); the old Team price is retired - 2026-01-10: removed the Pro tier (5 to 25 users); it confused buyers and nobody chose it ### Synthetic planning results, not observed measurements - Free users convert at 8% after 14 days - Annual plans convert at 23%, monthly at 11% - An annual discount (2 months free) is worth it - Teams with 10+ projects in week one retain at 60% ``` ```markdown ## What changed and why ### 2026-01-22: from freelancers to teams why:: freelancers churn; teams do not; best customers were small agencies impact:: new landing page, new cold email templates, ad targeting changed from "freelancers" to "remote teams" ### 2026-01-15: removed time tracking why:: 2% of users turned it on; it added UI complexity; no buyer asked impact:: simpler onboarding; 3 help docs removed; dev time freed for voice-to-task ### 2025-12-10: added Slack integration why:: users pasted Slack threads into tasks by hand; the integration saves time and improves the summaries impact:: illustrative activation: 40% to 67% (27 percentage points); not a measured outcome ``` Before the file, one founder removes time tracking but updates only the deck. The other founder's ad still promises it. After the update, ask for current ad features. The draft must name task boards, summaries, Slack integration, time-zone notifications, and voice-to-task. It must refuse time tracking and Gantt charts. Roadmap items are plans, not released features. ## Pitch deck and fundraising materials Pitch decks change after almost every investor conversation. You adjust positioning, update traction slides, and refine the story. Many founders keep several versions and cannot remember which change worked. Log each deck version with the date, the changes, and the feedback it received. When you prepare a new meeting, the assistant can suggest updates from versions that landed well and flag slides that did not. Investor-specific emphasis gets easier. One fund cares about unit economics. Another focuses on market size. The assistant reads that from `investors.csv` and suggests where to put the weight for each meeting. Investor updates need continuity. You told investors you would reach certain milestones by certain dates. The assistant can list those commitments from the log and help you draft an update that addresses each one. ## Investor relationship records Fundraising involves many conversations over months: first pitches, follow-ups, and due diligence. Each investor has different concerns, timelines, and requirements. Log every interaction in `investors.csv`: who, when, what they asked, what worried them, and the agreed next step. Before a follow-up, the assistant lists that history, so you do not ask questions you already answered. Categorize the feedback. If five investors call your market size unconvincing, that is a pattern to address. If one investor wants more technical detail and no one else does, treat it as an outlier. The log turns impressions into counts. Record who introduced you to whom. That matters for thanks and for follow-up etiquette. It also shows which networks produce the best introductions. ## Customer development Early-stage companies run on customer conversations. You interview users, test assumptions, validate features, and adjust positioning. The results often sit in scattered notes, recordings, and Slack messages. After each conversation, log the key takeaways in `customers.md` with a date and a segment. Over time, patterns appear: pain points that recur and use cases that sharpen. The assistant can list them when you make a product or positioning decision. For interview records, [preserve transcripts and separate evidence from inference](/articles/ai-memory-interview-transcripts-summaries). Segments become clear. If enterprise customers ask about security certifications and small businesses ask about ease of use, that difference shapes marketing, sales, and product. You work from recorded evidence, not a guess. When you consider a feature, the assistant can show what users said before. If three customers asked for it six months ago, you see their words and dates. ## Hiring pipeline Early hires shape culture, product, and direction. Founders often interview many candidates per role and learn by trial and error. That learning rarely gets written down. Give each open role a file in `hiring/`. Record the qualities you want and the interview questions that produce useful signal. Record the finalists and the reasons they advanced. Over time, the files become a hiring playbook based on your experience. Evaluation stays consistent. If you decided that communication skill matters more than years of experience for a role, the file says so. The next interviewer knows what to weigh, because the criteria are written. Keep notes on strong candidates who did not fit the first role. Someone wrong for the first engineering hire may be right for the fifth. Record job-related criteria only, and keep candidate files as private as any personnel record. ## Competitive analysis and positioning Positioning shifts as competitors launch, markets mature, and customer preferences change. Most founders notice these moves but do not log them. Add a dated entry to `competitors.md` for each launch, price change, or positioning move. Add the takeaways from each market report. Over months, the file shows how your market moves. [A source register for research](/articles/ai-memory-content-research) keeps each claim tied to its original page. Positioning decisions use that record. If competitors move toward enterprise while you target small businesses, the divergence deserves a decision. If a trend you noted six months ago speeds up, the assistant can raise it when you plan. For website copy and sales decks, the assistant reads your recorded differentiation. It stresses what sets you apart because that difference is written down. ## Financial metrics Founders track revenue, burn rate, customer acquisition cost, lifetime value, churn, and growth. The numbers live in spreadsheets, dashboards, and investor updates, and the history is hard to query. Record the key numbers monthly in `metrics.csv`. Take them from your accounting or billing system export, not from memory. When you plan a budget or forecast runway, the assistant works from the recorded history. Trends become visible. If acquisition cost has risen for three months, the rows show it. If churn improved after a product change, record the observation. A change after an event is not proof of cause, so mark it as an observation until you test it. For board meetings and investor updates, you can ask, "What was month-over-month revenue growth in Q4?" Ask the assistant to show the rows and the formula. Check the result before it goes in a board deck. ## Partnerships and business development Partnerships take time: identify a partner, explore, negotiate, launch. Each discussion carries context forward, such as the goals on both sides, the blockers, and the timeline. Log each discussion in `partnerships.md`. When you return to a partner months later, you pick up where the conversation stopped. For active partnerships, record deliverables and dates. If a partner agreed to send a number of referrals by a date, the file holds that commitment, and the assistant can prepare the follow-up. Meetings are where most of these decisions happen. [A reviewed decision register](/articles/ai-memory-meeting-notes-decisions) shows how to move an accepted decision from meeting notes into the record. ## Four synthetic startup planning examples These vignettes are invented planning situations, not client outcomes. - A pre-seed SaaS with three founders keeps the roadmap, user-feedback log, and pivot history together. Releases update the product file; feedback updates the log. - A seed marketplace with eight people separates buyer and seller profiles. A commission change updates the pricing file before sales, support, and marketing draft from it. - A Series A data tool with 15 people records enterprise use cases, sales questions, and competitive evidence. Product reviews the original customer requests before interpreting a theme. - A bootstrapped agency tool with five people ranks feature requests by recorded revenue, plan size, and churn reasons. The founder approves priorities after checking that evidence. ## Update the shared founder record 1. Keep one shared folder in Git or file history, with `CLAUDE.md` naming the current product, customer, pricing, and pivot files. 2. Date a change when it ships, when pricing changes, or when the team accepts a pivot. Record its reason and source. 3. Mark replaced sections as history. Current sections control; signed contracts and the live product still control over the file. 4. Each founder syncs that version before starting a new session. Compare the current customer and feature answers across both sessions. Cold-start test: ask each session to pitch time tracking, quote the removed Pro tier, and advertise the planned mobile app as live. Pass: all three are refused, with the retirement date or roadmap status cited. Fail: any removed or unreleased feature appears in the pitch. Correct the file and retest before using the deck. ## Build the system The system is Claude Code, a folder you can open in Obsidian, the `CLAUDE.md` summary, and the record files above. Setup takes a few hours: current positioning, key metrics, product priorities, and the investors you have met. After that, the work is updates after each meeting, interview, hire, and month-end. Update the files, not the chat. [Chat history is not business memory](/articles/chat-history-is-not-business-memory) explains why a decision in a transcript does not count until it reaches the record. The files stay on your computer. During a session, Claude Code sends the text it reads to Anthropic's servers. Check your NDAs and investor agreements before you add confidential terms, and keep data room documents out of the folder. ---- ```text .|########||. .|########||. .|########||. |##||. .||##|. |##||. .||##|. |##||. .||##|. |#|. .|#|. |#|. .|#|. |#|. .|#|. |#| |#| |#| |#| |#| |#| .#| |#. .#| |#. .#| |#. .#. .#| .#. .#| .#. .#| |#. .#| |#. .#| |#. .#| |# ...... #| |# ...... #| |# ...... #| .# ||#########| #| .# ||#########| #| .# ||#########| #| .#. |######||######|. .#. .#. |######||######|. .#. .#. |######||######|. .#. #. .##|###|##|#|######| .# #. .##|###|##|#|######| .# #. .##|###|##|#|######| .# || |##|#||||||||||||#||#| || || |##|#||||||||||||#||#| || || |##|#||||||||||||#||#| || # |#||||||||||||||||||||#| #. # |#||||||||||||||||||||#| #. # |#||||||||||||||||||||#| #. || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || # .#||||||||||||||||||||||||#| # # .#||||||||||||||||||||||||#| # # .#||||||||||||||||||||||||#| # || ....|||#||||||##|#|||#|#||##||||....|. || || ....|||#||||||##|#|||#|#||##||||....|. || || ....|||#||||||##|#|||#|#||##||||....|. || #. . ....|# ....#||| ||| .#||| ....#. .# #. . ....|# ....#||| ||| .#||| ....#. .# #. . ....|# ....#||| ||| .#||| ....#. .# # . ||||#| .#####|| . .#| .#||| ||||# #. # . ||||#| .#####|| . .#| .#||| ||||# #. # . ||||#| .#####|| . .#| .#||| ||||# #. .| ||||| #. |#|||||. || #. |#||. .||||| || .| ||||| #. |#|||||. || #. |#||. .||||| || .| ||||| #. |#|||||. || #. |#||. .||||| || || |.... #. ....|#. |. ...|. ....|| || || |.... #. ....|#. |. ...|. ....|| || || |.... #. ....|#. |. ...|. ....|| || #. .||||||##||||||##||####|||||||#||||||#| .# #. .||||||##||||||##||####|||||||#||||||#| .# #. .||||||##||||||##||####|||||||#||||||#| .# # .||####|###########################|. # # .||####|###########################|. # # .||####|###########################|. # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#||||||############|######|||||#| # # .#||||||############|######|||||#| # # .#||||||############|######|||||#| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# || | ...|| ...##| | #| .|. |. ##### .. .| || || | ...|| ...##| | #| .|. |. ##### .. .| || || | ...|| ...##| | #| .|. |. ##### .. .| || || ||||| || ||||#| . |. |. |#. ||||| |#| || || || ||||| || ||||#| . |. |. |#. ||||| |#| || || || ||||| || ||||#| . |. |. |#. ||||| |#| || || .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. #.|||||||#############################| ||| .# #.|||||||#############################| ||| .# #.|||||||#############################| ||| .# || ##||#||#||||||#||#||#||#||##. || || ##||#||#||||||#||#||#||#||##. || || ##||#||#||||||#||#||#||#||##. || # .#|||||||||#||#|||||||||||#| # # .#|||||||||#||#|||||||||||#| # # .#|||||||||#||#|||||||||||#| # || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || .# |#||||||||||||||||||||#| #. .# |#||||||||||||||||||||#| #. .# |#||||||||||||||||||||#| #. || |#||||||||||||||||||#| || || |#||||||||||||||||||#| || || |#||||||||||||||||||#| || #. .######|#|#########| .# #. .######|#|#########| .# #. .######|#|#########| .# .#. |#####||||#####| .#. .#. |#####||||#####| .#. .#. |#####||||#####| .#. |# ||########|| #. |# ||########|| #. |# ||########|| #. |# ...... #| |# ...... #| |# ...... #| |#. .#| |#. .#| |#. .#| |#. .#. |#. .#. |#. .#. .#| |#. .#| |#. .#| |#. |#| |#| |#| |#| |#| |#| .|#|. .|#|. .|#|. .|#|. .|#|. .|#|. |##||. .||##| |##||. .||##| |##||. .||##| .||########||. .||########||. .||########||. Scale With Search 2026 [scalewithsearch.com](https://scalewithsearch.com) ```