- [Artificial intelligence](https://www.digica.com/index.php?option=com_tags&view=tag&id%5B0%5D=15:artificial-intelligence)
- [ai software for business](https://www.digica.com/index.php?option=com_tags&view=tag&id%5B0%5D=98:ai-software-for-business)
- [artificial intelligence applications](https://www.digica.com/index.php?option=com_tags&view=tag&id%5B0%5D=99:artificial-intelligence-applications)
- [AI](https://www.digica.com/index.php?option=com_tags&view=tag&id%5B0%5D=104:ai)
- [AI for business](https://www.digica.com/index.php?option=com_tags&view=tag&id%5B0%5D=107:ai-for-business)
- [software development](https://www.digica.com/index.php?option=com_tags&view=tag&id%5B0%5D=113:software-development)

#  How to choose a software engineering supplier: a CTO's checklist

Jim Carroll and Ben Stammers

 [ Back ](https://www.digica.com/index.php?option=com_content&view=category&id=27)

- [](https://www.facebook.com/sharer/sharer.php?u=https%3A%2F%2Fwww.digica.com%2Fblog%2Fhow-a-cto-picks-an-engineering-supplier.html.md)
- [](https://x.com/intent/post?url=https%3A%2F%2Fwww.digica.com%2Fblog%2Fhow-a-cto-picks-an-engineering-supplier.html.md&text=)
- [](https://www.linkedin.com/sharing/share-offsite/?url=https%3A%2F%2Fwww.digica.com%2Fblog%2Fhow-a-cto-picks-an-engineering-supplier.html.md)
- [](mailto:?subject=How%20to%20choose%20a%20software%20engineering%20supplier%3A%20a%20CTO%27s%20checklist&body=Hi%3Cbr%3EPlease%20check%20this%20article%3A%20index.php%3Foption%3Dcom_content%26view%3Darticle%26id%3D362%3Ahow-a-cto-picks-an-engineering-supplier%26catid%3D27)

*By Jim Carroll, CTO, and Ben Stammers, Head of Marketing, Digica*

![Jim Carroll, CTO of Digica](https://www.digica.com/images/Blog/2026/jim-cto-article/jim-carroll.png)

**In short:** Digica’s CTO, Jim Carroll, judges an engineering supplier on its process, not its technology. He checks who they’ve worked for, asks for evidence of how they work, picks a supplier big enough to scale and small enough to care, and judges the first call by the questions the supplier asks.

**If they’re not asking smart questions, then they’re not smart.**

That sums up most of what I need to know about a supplier. I’ve been an engineer for over 30 years, most of them in software services. Before Digica, I was CTO of a much larger engineering services company. Along the way, I’ve overseen hundreds of projects, some with more than 100 engineers working for a single client, for the world’s biggest semiconductor companies, household-name automotive brands and Silicon Valley giants. I’ve bought engineering as well as sold it, so I know what a buyer is really looking for.

My starting point is that the technology is rarely what separates one supplier from another.

Technology is often just table stakes. Lots of companies can do it. The key word in the phrase software engineering is not software; it’s engineering.

It’s the process that gives you reliability and predictability. If they say Thursday, you get it on Thursday.

Here’s how I narrow a long list down to the one I sign with.

## 1. Look at who they work with

First, I look at the sectors on a supplier’s website.

If they just do consumer software, such as websites or mobile apps, the engineering rigour doesn’t have to be there. If they’re building enterprise-grade applications, or building for highly regulated industries like finance, medical or defence, that’s the first indicator. Who do they work with?

Regulated work forces discipline. A supplier that has shipped into those sectors has had its process tested by someone stricter than the buyer.

Buyers in those sectors are wary. Most have heard a supplier say all the right things and then not deliver. So they want to meet you face to face and see that you’ve taken the time to understand their problem.

## 2. Look at the process, then ask for the evidence

What I want to see is how a supplier works, end to end. If I’m going to spend hundreds of thousands of dollars with somebody on building stuff, I want to know how they do it.

That doesn’t mean months of due diligence. It means seeing what their process looks like, what their tooling looks like, and some examples of their work.

When I get into a purchasing process, I want to see evidence of all of that. What’s on a website is a claim. The evidence is what a supplier shows when asked.

I’d like to see two things in particular. The first is some kind of technical report. It doesn’t matter what it’s about. I’m assessing whether the depth is appropriate: too deep is overthinking, too shallow is a failure of communication. I like to see something I hadn’t thought of: not necessarily technically deep, but a different angle.

The second is some kind of project plan, in any form. It just has to demonstrate that the delivery commitments are based on something, not pulled out of the air.

> “They produce detailed weekly written status reports and excellent documentation.”
>
> *VP of Data Science, medical AI company, in a Clutch review*

## 3. Get the size right

I don’t want them to be big, and I don’t want them to be small. Not a hundred thousand people, and not five.

What matters is whether your project is a big deal to them, and whether they can carry it. A ten-million-dollar contract is a big deal for a mid-sized supplier and a line item for a giant, so the giant doesn’t really care. A five-person company can’t carry it. I’ve used the same line for fifteen years: big enough to scale, small enough to care.

## 4. A fast reply is a good sign, not the deciding one

On a large contract, I might send an RFI to ten or twenty suppliers, take proposals from no more than five, and shortlist two or three. That kind of cycle runs for months. Nobody spends ten million dollars that quickly. The difference between five minutes and two days is zero.

That doesn’t make speed worthless. A supplier that gets back quickly is usually well run, and buyers naturally end up talking to the ones who reply. But a quick answer only proves the sales team is on the ball. Whether the engineering is any good gets decided later.

## 5. In the first conversation, listen for the questions

The first call tells a buyer most of what they need to know, and the signal is the supplier’s questions, not just their answers.

I don’t want people just agreeing with me. If you disagree, you’ve got to be able to explain why. I don’t want somebody who just wants to take an order.

They need to speak the buyer’s language and push back when it’s warranted. Every pre-sales meeting I run is trying to answer three questions: what does the client want, what have they already got, and when do they want it. Nobody asks them that bluntly, but a good supplier gets to all three.

You measure suppliers by the questions they ask.

> “They knew how to ask the right questions, and their adaptability helped us achieve a structured scope.”
>
> *VP of Software Engineering, robotics company, in a Clutch review*

## 6. See who puts some skin in the game

The suppliers who stand out make an effort before they’re paid. They come to see the client in person. They build something small that works, a demo or even a UI mockup, to show how they think. If somebody’s made the effort to come and see you, that’s a good sign.

We did exactly that at Digica. We spent three weeks of our own time building a face detection and recognition demo for one of the world’s largest technology companies. It was convincing enough that they accepted our proposal to build a high-performance, quantised solution from scratch to run on their wearable device. Three weeks of effort before any contract won us the project.

> “Even if somebody came to me from one of the big consultancies, they may have done hundreds of these projects. They didn’t do them for me. I don’t necessarily trust them, because I know their model is billing.”
>
> *VP of Client Success, surgical technology company*

> “You know when a partner really is a partner rather than just a supplier when they did not give up, ask for more money or blame anything else. They simply doubled down.”
>
> *Technology Manager, farming company*

Before committing to a large project, I like to start with a small one, so both sides learn how the other works before the stakes go up.

## 7. Read the proposal for understanding, not length

A proposal used to be 10 to 15 pages as standard, and I don’t mean slides, I mean pages. That’s less clear-cut now, but a good one still has the same bones every time: an executive summary, a restatement of the problem, any interesting observations, the proposed solution, a timeline and the assumptions. It’s completely standard, and that’s the point. A supplier who restates the problem accurately has listened. One who doesn’t has sent a template.

## The checklist

![Checklist: eight questions to ask before you sign with an engineering supplier](https://www.digica.com/images/Blog/2026/jim-cto-article/jim-cto-checklist.png)

- **Sectors:** have they shipped into regulated industries?
- **Process:** can they show you how they work, end to end?
- **Evidence:** can they show you a technical report and a project plan?
- **Size:** is your project a big deal to them, and can they carry it?
- **Speed:** is their quick reply backed by engineering that’s just as good?
- **Questions:** did the first call challenge you, or just agree?
- **Effort:** did they show you something before they were paid?
- **Proposal:** does it restate your problem accurately?

The technology is table stakes. The engineering is the difference.

## Frequently asked questions

**How do I choose a software or AI engineering supplier?** Judge the process, not the technology. Check who they’ve worked for, ask to see how they work, pick a supplier your project matters to, and listen to the questions they ask in the first call. The checklist above covers the rest.

**What should I ask a supplier to show me before I sign?** A technical report and a project plan, in any form. Judge the report on its depth: too deep is overthinking, too shallow is a failure of communication. The best ones show you an angle you hadn’t thought of. The plan should show that the delivery dates are based on something, not pulled out of the air.

**How big should my engineering supplier be?** Big enough to scale, small enough to care. Your project should be a big deal to them, and they should be able to carry it. To a giant it’s a line item, and a five-person company can’t deliver it.

**Does a fast reply mean a supplier is good?** It’s a good sign, because well-run companies tend to reply quickly. But on a buying cycle that runs for months, five minutes or two days makes no difference. A quick answer only proves the sales team is on the ball. Whether the engineering is any good gets decided later.

**What should a good supplier proposal include?** An executive summary, a restatement of your problem, any observations, the proposed solution, a timeline and the assumptions. Pages, not slides. If they restate your problem accurately, they’ve listened.

**Should I start with a small project before a big one?** Yes, if you can. A small first project lets both sides learn how the other works before the stakes go up.

*Jim Carroll is CTO of Digica, a software and AI engineering company with 50+ engineers in the UK, Poland and the US. Ben Stammers is Digica’s Head of Marketing and co-wrote this piece with Jim.* [Discuss your project](https://www.digica.com/contact)

###  Tagged with

- [Artificial intelligence](https://www.digica.com/index.php?option=com_tags&view=tag&id%5B0%5D=15:artificial-intelligence)
- [ai software for business](https://www.digica.com/index.php?option=com_tags&view=tag&id%5B0%5D=98:ai-software-for-business)
- [artificial intelligence applications](https://www.digica.com/index.php?option=com_tags&view=tag&id%5B0%5D=99:artificial-intelligence-applications)
- [AI](https://www.digica.com/index.php?option=com_tags&view=tag&id%5B0%5D=104:ai)
- [AI for business](https://www.digica.com/index.php?option=com_tags&view=tag&id%5B0%5D=107:ai-for-business)
- [software development](https://www.digica.com/index.php?option=com_tags&view=tag&id%5B0%5D=113:software-development)
- [](https://www.facebook.com/sharer/sharer.php?u=https%3A%2F%2Fwww.digica.com%2Fblog%2Fhow-a-cto-picks-an-engineering-supplier.html.md)
- [](https://x.com/intent/post?url=https%3A%2F%2Fwww.digica.com%2Fblog%2Fhow-a-cto-picks-an-engineering-supplier.html.md&text=)
- [](https://www.linkedin.com/sharing/share-offsite/?url=https%3A%2F%2Fwww.digica.com%2Fblog%2Fhow-a-cto-picks-an-engineering-supplier.html.md)
- [](mailto:?subject=How%20to%20choose%20a%20software%20engineering%20supplier%3A%20a%20CTO%27s%20checklist&body=Hi%3Cbr%3EPlease%20check%20this%20article%3A%20index.php%3Foption%3Dcom_content%26view%3Darticle%26id%3D362%3Ahow-a-cto-picks-an-engineering-supplier%26catid%3D27)

How to choose a software engineering supplier: a CTO's checklist

###  Related posts

[ ![](https://www.digica.com/templates/yootheme/cache/f2/heart-f22c7316.png)

###  How to avoid denial in Data Science

01.05.2022

[Read more](https://www.digica.com/index.php?option=com_content&view=article&id=169:how-to-avoid-denial-in-data-science&catid=27)

 ](https://www.digica.com/index.php?option=com_content&view=article&id=169:how-to-avoid-denial-in-data-science&catid=27)

[ ![](https://www.digica.com/templates/yootheme/cache/83/AdobeStock_280791155-83a20d50.jpeg)

###  Do you know when you don’t know something?

04.24.2023

[Read more](https://www.digica.com/index.php?option=com_content&view=article&id=201:do-you-know-when-you-don-t-know-something&catid=27)

 ](https://www.digica.com/index.php?option=com_content&view=article&id=201:do-you-know-when-you-don-t-know-something&catid=27)

[ ![](https://www.digica.com/templates/yootheme/cache/1b/neural_networks-1b9ae2ab.jpeg)

###  How to find the best food for your AI model?

10.14.2022

[Read more](https://www.digica.com/index.php?option=com_content&view=article&id=182:how-to-find-the-best-food-for-your-ai-model&catid=27)

 ](https://www.digica.com/index.php?option=com_content&view=article&id=182:how-to-find-the-best-food-for-your-ai-model&catid=27)