Skip to content
Lubo Bali
BlueprintBook a call (opens Calendly in a new tab)

By Lubo Bali

Get your free AI blueprint

Tell me your problem the way you would on a call. In about a minute you get a first plan: how your work runs today, what it looks like with AI, who is involved, the risks and a timeline. Free, one per visitor.

Where does your data live today?
How many people on your team?

Please describe it in general terms: no customer data or passwords. What you type is saved without your name, so I can learn what people need. Emails and phone numbers are removed. How I handle your data

Free · about 1 minute · one blueprint per visitor

See what you'd get

6 example blueprints made by the tool. Each one starts from a typical story for that industry, never from a real client.

Choose an industry

AI blueprint · exampleData that's ready for AI

One fleet dashboard

Each Monday the office manager spends about 6 hours copying fuel and route data from 3 spreadsheets into one report. By the time the owner reads it, it is already old.

Goal Give the owner one fresh, trusted view of fuel and routes for all 40 trucks, without hours of copying.

Today6 h a week
  1. Export fuel card dataSpreadsheets
  2. Collect route dataSpreadsheets
  3. Copy 3 sheets into one reportSpreadsheets · 6 h a weekpain point
  4. Match card numbers to platesSpreadsheetspain point
  5. Owner reads the weekly reportSpreadsheetspain point
With your AI~35 min a week estimate

One fleet dashboard

  1. Fuel and route data flow in
  2. Each card is linked to its truck
  3. Report updates by itself
  4. Owner sees fresh numbers anytime

Who's involved

  • OwnerdecidesReads the report and decides on fuel and routes
  • Office managerdoes the workCopies the 3 sheets into one report every Monday
  • Dispatchersdoes the workPlan the routes for the trucks
Data readiness

40/100

Early, and that's normal: week 1 cleans up your data

  • To fix: Your data lives in spreadsheets
  • To fix: The apps don't share the same IDs
  • To fix: The same data is kept in several places
Timeline 4–5 weeks
  1. Week 1Clean up and connect your data
  2. Week 2Your data in one place
  3. Week 3Answers and reports on top
  4. Week 4Handover
Time each week

Today6 h

With your AI~35 min

From your own numbers. “With your AI” is an estimate.

Risks to watch

  • Fix firstFuel cards use card numbers, routes use plate numbers
  • WatchCards may move between trucks or drivers over time
  • WatchHand-typed sheets may hold gaps or typos

3 questions I'd ask you on our call

  1. Is there a list today that says which fuel card belongs to which truck plate, and who keeps it up to date?
  2. Which numbers does the owner look at first in the Monday report, and what decision does he make from them?
  3. Where do the fuel card and route sheets come from: are they downloaded from a provider or typed in by the dispatchers?

Your full technical plan

  • Which tools and AI models
  • How it is built (architecture)
  • How I solve each risk
  • Cost estimate

We go through it together on a free 30-min call.

Book a free 30-min call (opens Calendly in a new tab)

AI blueprint · exampleAnswers you can trust

One residual payout desk

Every month, agent splits are worked out by hand in Excel from 3 processor reports. It takes 2 full days and agents often dispute their numbers.

Goal Correct agent payouts every month, fast, with numbers agents can check and trust.

Today3.5 h a week
  1. Download residual reportsProcessor portals
  2. Match merchants, calculate splitsExcel · 16 h a monthpain point
  3. Check splits against contractsContract PDFspain point
  4. CEO approves the splitsExcel
  5. Finance pays, agents disputeExcelpain point
With your AI~45 min a week estimate

One residual payout desk

  1. All 3 residual reports in one place
  2. Merchants matched to agents
  3. Splits calculated from contracts
  4. CEO reviews and approves
  5. Agents see their own statement

Who's involved

  • CEOapprovesApproves every agent split before payout
  • Finance persondoes the workDownloads reports, matches merchants, pays agents
  • Sales agents (30)kept in the loopEarn a 40 to 60 percent split of their residuals
  • Processors (3)kept in the loopSend the monthly residual reports
Data readiness

16/100

Early, and that's normal: week 1 cleans up your data

  • To fix: Your data lives in spreadsheets
  • To fix: 3 different apps hold the data
  • To fix: The apps don't share the same IDs
Timeline 4–5 weeks
  1. Week 1Clean up and connect your data
  2. Week 2Numbers calculated by code
  3. Week 3The AI explains the numbers
  4. Week 4Handover
Time each week

Today3.5 h

With your AI~45 min

From your own numbers. “With your AI” is an estimate.

Risks to watch

  • Fix firstSplits in contract PDFs may not match the Excel tab
  • WatchEach processor may name or number merchants differently
  • WatchAgent pay and merchant data must stay private

3 questions I'd ask you on our call

  1. When an agent disputes a number, what is usually wrong: the merchant match, the split, or the processor report?
  2. Do any agents have special deals, like split changes over time, shared merchants, or sub-agents?
  3. How do you know which merchant belongs to which agent today, and who decides when two agents claim one?

Your full technical plan

  • Which tools and AI models
  • How it is built (architecture)
  • How I solve each risk
  • Cost estimate

We go through it together on a free 30-min call.

Book a free 30-min call (opens Calendly in a new tab)

AI blueprint · exampleAI agents that do the work

One patient intake assistant

Patients fill paper and PDF intake forms that the front desk retypes by hand. Wrong or missing insurance details mean about 1 in 10 claims gets denied.

Goal Get patient details into your practice software right the first time, so fewer claims get denied.

Today15 h a week
  1. Patient fills intake formPaper or PDF
  2. Front desk retypes the formPractice software · 2 h a daypain point
  3. Claim sent to insurerPractice software
  4. Billing fixes denied claimsPractice software · 5 h a weekpain point
With your AI~2.5 h a week estimate

One patient intake assistant

  1. Forms are read automatically
  2. Details fill the practice software
  3. Missing insurance info is flagged
  4. Front desk checks only the flags
  5. Clean claims go out the first time

Who's involved

  • Doctors (4)decidesRun the clinic and see patients
  • Front deskdoes the workRetypes intake forms into the practice software
  • Billing persondoes the workFixes and resubmits denied claims by hand
  • Patientsuses itFill in paper and PDF intake forms
Data readiness

19/100

Early, and that's normal: week 1 cleans up your data

  • To fix: Key facts sit in documents and PDFs
  • To fix: Important facts are in free text
  • To fix: Nobody knows how clean the data is
Timeline 4–5 weeks
  1. Week 1Clean up and connect your data
  2. Week 2Connect the agent to your data
  3. Week 3Teach it your real tasks
  4. Week 4Handover
Time each week

Today15 h

With your AI~2.5 h

From your own numbers. “With your AI” is an estimate.

Risks to watch

  • Fix firstPatient health and insurance data must stay private
  • WatchHandwritten paper forms may be hard to read
  • WatchInsurance details are often missing at intake

3 questions I'd ask you on our call

  1. What are the most common reasons your claims get denied today?
  2. Roughly how many intake forms come in on paper versus as PDFs each day?
  3. Who in the clinic would sign off on a new tool that touches patient records?

Your full technical plan

  • Which tools and AI models
  • How it is built (architecture)
  • How I solve each risk
  • Cost estimate

We go through it together on a free 30-min call.

Book a free 30-min call (opens Calendly in a new tab)

AI blueprint · exampleAI agents that do the work

One client documents desk

Every month the team chases 40 clients by email for bank statements and receipts, often 3 or 4 reminders each, then types it all into QuickBooks by hand.

Goal Get client documents in on time and into the books without retyping, so month-end close is much shorter.

Today11 h a week
  1. Email clients for documentsEmail
  2. Send 3-4 reminders per clientEmailpain point
  3. Collect statements and receiptsDocuments and PDFs
  4. Type everything in by handQuickBookspain point
  5. Month-end closeQuickBooks · 48 h a monthpain point
With your AI~1.5 h a week estimate

One client documents desk

  1. Clients get automatic reminders
  2. Clients upload documents in one place
  3. Agent reads statements and receipts
  4. Entries prepared for QuickBooks
  5. Bookkeepers check and approve

Who's involved

  • Firm ownerdecidesRuns the firm and decides on new ways of working
  • Bookkeepers (2)uses itChase clients and type documents into QuickBooks
  • Rest of the teamdoes the workWork on client books and the month-end close
  • Small-business clientskept in the loopSend bank statements and receipts each month
Data readiness

13/100

Early, and that's normal: week 1 cleans up your data

  • To fix: Key facts sit in documents and PDFs
  • To fix: 3 different apps hold the data
  • To fix: Important facts are in free text
Timeline 4–5 weeks
  1. Week 1Clean up and connect your data
  2. Week 2Connect the agent to your data
  3. Week 3Teach it your real tasks
  4. Week 4Handover
Time each week

Today11 h

With your AI~1.5 h

From your own numbers. “With your AI” is an estimate.

Risks to watch

  • Fix firstBank statements arrive in many formats and bank layouts
  • Fix firstClient financial data is sensitive and must stay protected
  • WatchPhotos of receipts may be blurry or incomplete

3 questions I'd ask you on our call

  1. Of the 6 days of month-end close, how much is chasing and typing, and how much is review and checks?
  2. Do clients send documents as PDFs, photos or paper, and does each client have a different bank?
  3. Who in the firm signs off on entries today before they are final in QuickBooks?

Your full technical plan

  • Which tools and AI models
  • How it is built (architecture)
  • How I solve each risk
  • Cost estimate

We go through it together on a free 30-min call.

Book a free 30-min call (opens Calendly in a new tab)

AI blueprint · exampleAI agents that do the work

One job billing app

Approved change orders often never reach billing, so you lose money on work your crews already did.

Goal Every approved change order and every hour on site gets billed, without the monthly scramble.

Today7 h a week
  1. Foremen log hours and materialsPaperpain point
  2. Change orders sent for signaturesEmailpain point
  3. PM and client approve change orderEmail
  4. Office builds monthly invoicesQuickBooks · 30 h a monthpain point
With your AI~1 h a week estimate

One job billing app

  1. Foremen log hours on site
  2. Change orders signed in the app
  3. Approved work flagged for billing
  4. Invoice drafts ready for review

Who's involved

  • Foremandoes the workWrites hours and materials on paper tickets
  • Project managerapprovesSigns off change orders on each job
  • ClientapprovesSigns off change orders and pays the invoices
  • Office teamdoes the workBuilds invoices each month from tickets and emails
  • OwnerdecidesWants every job billed in full
Data readiness

30/100

Early, and that's normal: week 1 cleans up your data

  • To fix: 3 different apps hold the data
  • To fix: Your data sits inside other apps
  • To fix: The apps don't share the same IDs
Timeline 4–5 weeks
  1. Week 1Clean up and connect your data
  2. Week 2Connect the agent to your data
  3. Week 3Teach it your real tasks
  4. Week 4Handover
Time each week

Today7 h

With your AI~1 h

From your own numbers. “With your AI” is an estimate.

Risks to watch

  • Fix firstPaper tickets and emails don't link to one job number
  • WatchSigned change orders are scattered across inboxes
  • WatchForemen may not switch from paper tickets easily

3 questions I'd ask you on our call

  1. How much do you think you lost last year on change orders that never got billed?
  2. How do you invoice today: fixed price, time and materials, or a mix per job?
  3. Do foremen have phones or tablets on site, and would they use them daily?

Your full technical plan

  • Which tools and AI models
  • How it is built (architecture)
  • How I solve each risk
  • Cost estimate

We go through it together on a free 30-min call.

Book a free 30-min call (opens Calendly in a new tab)

AI blueprint · exampleAI agents that do the work

One client desk for the firm

Client details are typed in by hand, conflict checks are manual, and lawyers search hundreds of PDFs to answer the same client questions, between cases.

Goal Free lawyers and the assistant from retyping and searching, so clients get fast, correct answers.

Today
  1. Client sends detailsEmail and phone
  2. Assistant types details inCase softwarepain point
  3. Conflict check by handCase softwarepain point
  4. Client asks about contractEmail
  5. Lawyer searches the PDFsPDFspain point
  6. Lawyer writes the replyEmail
With your AI

One client desk for the firm

  1. Client details read from email
  2. Case file filled in for you
  3. Conflict check runs, lawyer approves
  4. Answer drafted from client's PDFs
  5. Lawyer checks source, then sends

Who's involved

  • Lawyers (6)decidesAnswer client questions between cases
  • Assistantdoes the workTypes client details and runs conflict checks
  • Clientskept in the loopSend details and ask about contracts and deadlines
Data readiness

10/100

Early, and that's normal: week 1 cleans up your data

  • To fix: Key facts sit in documents and PDFs
  • To fix: 4 different apps hold the data
  • To fix: Important facts are in free text
Timeline 5–6 weeks
  1. Week 1Clean up and connect your data
  2. Week 2Connect the agent to your data
  3. Week 3Teach it your real tasks
  4. Week 4Connect the remaining apps
  5. Week 5Handover

Risks to watch

  • Fix firstClient files are confidential and privileged
  • WatchNames spelled differently can hide a conflict
  • WatchA wrong deadline in an answer could harm a client

3 questions I'd ask you on our call

  1. Which client questions come up most often, and which ones must always stay with a lawyer?
  2. How are your PDFs organised today: by client, by case, or all in one place?
  3. What exactly do you check in a conflict check, and where do those names come from?

Your full technical plan

  • Which tools and AI models
  • How it is built (architecture)
  • How I solve each risk
  • Cost estimate

We go through it together on a free 30-min call.

Book a free 30-min call (opens Calendly in a new tab)

What happens to your data

  • Your story goes to the AI only to write your blueprint, and only to AI providers that don't keep it or train on it. I also save it, without your name, emails or phone numbers, to learn what people need.
  • Your blueprint stays in your own browser, so you can see it again. It's saved on my server only if you click “Book a call”, for 90 days, so I can read it before we talk.
  • If you ask for the PDF, it goes to you and to my inbox. Your email isn't stored in the tool or added to any list.
  • To stop abuse, I keep a scrambled code of your browser for 12 months and of your IP for about a week. My server logs keep IP addresses for 14 days. No cookies, no ads.
  • Visits I count page visits, clicks and what you type, without cookies and without your name, to see what people look for. If your browser sends Do Not Track, nothing is counted.
  • Questions, or want something deleted? Email data@lubobali.com.