Your team ships. The developers are good. And yet three questions keep landing back on your desk:
- What should we build next?
- Where does AI genuinely fit?
- Is the software ready to launch?
If those decisions keep returning to you, the gap is probably not development capacity. It is technology leadership: someone whose job is to decide what gets built and whether it should be, not only how.
Read the full script, scene by scene
-
1. The decisions still depend on you
On screen: Technology decisions
Your developers can build. But are you still deciding what to build, where AI fits, and whether it’s ready to launch?
-
2. Your fractional CTO
On screen: Fractional CTO
I’m Parag, founder of Your Tech Chief. I’ve spent ten years leading technology for businesses across India and the UK. As your fractional CTO, I bring that leadership without a full-time hire.
-
3. Start with the business goal
On screen: Business goal first
We start with your business goal. Then we decide which technology, AI, or automation genuinely helps.
-
4. The manual enquiry-to-quote journey
On screen: Too much manual work?
Say you run a service business. Each customer enquiry means reading the details, chasing missing information, and preparing a quote.
-
5. Extract the requirements
On screen: 1. Clear requirements
AI can help turn that enquiry into clear requirements and flag what’s missing. It shouldn’t guess the answer.
-
6. Assign the next action
On screen: 2. Clear ownership
Automation assigns the request to the right person, with a clear next action. So it doesn’t sit in an inbox.
-
7. Review and approve
On screen: 3. Human approval
Your team confirms the missing details, checks the scope and pricing, and approves the quote before it reaches the customer.
-
8. The business value
On screen: Less work. Faster responses.
The goal is less manual work and faster responses. That’s useful AI and automation, not technology for show.
-
9. The CTO judgement behind the solution
On screen: Judgement before tools
My role is to decide what’s worth automating, what needs human judgement, and whether the solution fits your business, budget, and existing systems.
-
10. From requirements to execution
On screen: Define. Prioritise. Deliver.
That’s one example. From defining requirements and setting the strategy to guiding execution, I keep technology aligned with your business goals.
-
11. Working alongside your team
On screen: Your technology partner
You set the business priorities. I work with your team to turn them into clear technology decisions and delivery.
-
12. Share your technology question
On screen: Comment or message Parag
What technology decision is holding your business back? Share your question in the comments, or message me to discuss it privately.
Developers answer “how”. Someone has to answer “whether”.
A good development team will build what is asked, well. What it can’t do on its own is weigh the business: which customer problem matters most this quarter, what the budget can carry, which existing system has to keep working, and how much risk is acceptable at launch. Those are leadership calls. When nobody owns them, they fall back to the founder — usually as a question in a chat thread, at the worst possible moment.
It helps to tell the two gaps apart, because they need different fixes:
| You see this | It points to | The fix |
|---|---|---|
| The plan is clear, the work is just slow | Development capacity | More or better builders |
| The team keeps asking you to choose between options you can’t judge | Technology leadership | Someone who owns the technical decisions |
| AI ideas pile up and none gets tested properly | Technology leadership | A decision on where AI fits, and a small, measured test |
| “Is it ready?” has no agreed answer, so launches slip | Technology leadership | A definition of ready, and one person accountable for it |
Adding developers to a leadership gap usually makes it louder: more people, more questions, the same desk.
A worked example: from enquiry to quote
Take a service business. An enquiry arrives: “We need IT support for two offices. Can you send us a quote?” Today, someone reads it, realises half the details are missing, chases them by email, waits, prepares a quote, and hopes nothing was lost on the way. Multiply that by every enquiry in a week. (This is an illustrative workflow, the same one shown in the film, not a specific client.)
Here is the same journey, with each step given to whatever does it best — and the reason why.
| Step | Who does it | Why there |
|---|---|---|
| 1. Turn the enquiry into a brief | AI | The input is free text and inconsistent. AI can pull out what is known (two offices, IT support) and flag what is missing (how many people, which locations, start date). It should not guess the answer. |
| 2. Give it an owner and a next action | Automation | A rule decides it: service type goes to the right team, with a task and a due time. No judgment needed, so no reason to spend a person or an AI on it. The enquiry stops sitting in an inbox. |
| 3. Confirm the missing details | A person (AI can draft the question) | This is the first real conversation with the customer. Tone, context and the relationship matter more than speed. |
| 4. Check scope and pricing, approve the quote | A person | A mistake here costs money or a customer. The decision stays human; the quote goes nowhere until it’s approved. |
| 5. Send the quote and set the follow-up | Automation | Once approved, sending it and scheduling the follow-up are rules again. |
The result is less manual work and faster responses, with judgment exactly where it is needed. That is useful AI and automation. Putting AI on every step, or automating the approval to save a few minutes, is technology for show — and it is how a fast system sends a wrong quote.
AI, automation or a person: four questions for any step
The example generalises. For each step in a process you want to improve, ask these in order and stop at the first “yes”:
- Is it rare, still changing, or cheap to do by hand? Then leave it alone for now. Automating a process that changes every month means rebuilding the automation every month.
- Does it commit the business to something? A price, a scope, a promise, anything a customer will see. Then a person decides. AI or automation can prepare the work, but a person approves it, because a mistake here costs money or trust.
- Can a written rule decide it? Then automate it. Rules are cheaper, faster and more predictable than AI, and easier to check.
- Is the input messy, but can a person check the output quickly? Then AI drafts and a person reviews. If nobody can check the output, it isn’t ready for AI.
Whatever the answer, every step still needs a named owner. Automation without an owner drifts; AI without an owner guesses; and a step nobody owns comes back to the founder.
When we rebuilt the enquiry journey for a premium real-estate business, the biggest change was not a tool: it was giving every lead an owner and a next step. Conversions rose from 0.6 to 1.55 a month. And when we built an AI that scores a charity on ten dimensions, every score came with its sources and a confidence level, and a person can override it. The AI did the reading; people kept the decision.
The ten-minute decision sheet
Pick one process that keeps landing on you. List its steps, then fill in one row per step. The last column is the one that matters: if you can’t write the reason, you haven’t made the decision yet.
| Step | How often | Cost of a mistake | Decision | Owner | Reason |
|---|---|---|---|---|---|
| Read the enquiry | Daily | Low (a person reviews) | AI drafts the brief | Sales lead | Messy input, output easy to check |
| Assign it | Daily | Medium (a lost enquiry) | Automate | Operations | A rule decides it |
| Approve the quote | Daily | High (money, trust) | A person | Account owner | Customer-facing and costly to get wrong |
| Annual price review | Yearly | High | Leave alone | Founder | Once a year; done by hand |
If the sheet raises more questions than it answers, that is useful too. Bring it to a free 30-minute review, and you get a written list of what to fix first within a few hours.
What to leave alone
The most valuable decision is often not to build or automate something. Leave it alone when:
- It happens rarely. A yearly task done by hand costs less than the automation that has to be maintained all year.
- The process is still changing. Settle how the work should run first; automate it once it stops moving.
- The relationship is the product. Some conversations should sound like a person because they are one.
- Nobody can check the AI’s output. If a wrong answer would go unnoticed, AI isn’t ready for that step.
- The tool costs more than the time it saves. Count the subscription, the set-up and the upkeep, not just the licence.
None of this is anti-technology. It is how you make sure the technology you do add earns its place. (For more on that, see three questions to ask before adding another tool.)
Where a fractional CTO fits
A fractional CTO gives a business senior technology leadership part-time, without a full-time hire. Alongside an existing team, that means three things: defining the requirements before anything is built, setting the priorities so the team works on what matters most, and guiding execution through to launch.
You set the business priorities. I work with your team to turn them into clear technology decisions and delivery — including what is worth automating, what needs human judgment, and whether a solution fits your business, budget and existing systems.
I’m Parag Bhadoria, and YTC is my fractional CTO practice, built on 10+ years of leading technology for businesses across the UK and India. You can see the situations founders usually bring and what changed for six of them.
Questions founders ask about deciding what to automate
How do I decide what to automate in my business?
Look at one process at a time and judge each step on its own. Automate steps a clear rule can decide, use AI where the input is messy but a person can check the output, keep a person on any step that commits the business to a price, a promise or anything a customer sees, and leave rare or still-changing steps alone. Write down the reason for each decision.
What is the difference between AI and automation?
Automation follows rules you write: when this happens, do that. It is predictable and cheap. AI handles work that needs some interpretation, such as reading a free-text enquiry and turning it into a brief. It is more flexible but can be wrong, so its output needs checking. Most good workflows use both, plus a person for the decisions that matter.
When should a person stay in the loop?
On any step that commits the business, where a mistake would cost money or trust: approving a quote, agreeing scope, committing to a price or a date. AI and automation can prepare that work, but a person should approve it before it reaches the customer.
Do I need a CTO if I already have developers?
Not always. If the plan is clear and the work is simply slow, you need more development capacity. If decisions about what to build, where AI fits or whether the software is ready keep coming back to you, the gap is technology leadership, and a fractional CTO can cover it part-time without a full-time hire.
What does a fractional CTO do alongside an existing team?
Defines requirements before work starts, sets priorities so the team builds what matters most, decides where AI and automation genuinely help, and guides execution through to launch. The founder sets the business priorities; the fractional CTO turns them into technology decisions and delivery.