Delivery Management and AI: Engineering Leadership Pulse

study+001Fieldwork open

Twenty-one questions for technology, product, and delivery leaders, on what changed in delivery once writing code stopped being the constraint. About eight minutes. Contributors get the paper first.

From Tom Hill, a fractional CPTO in the thick of AI adoption, and Andrei Gliga, a delivery expert working out what AI means for planning. Two ends of the same problem.

* marks the three required answers: your role and your area at the start, and your email at the end. Everything else is optional, and a partial answer is still useful.

What happens to your answers

Your answers are anonymous: no name, no identifier, no IP address, no cookies.

Your email is required. It is kept apart, in a database that holds only addresses, to stop anyone answering twice and to send you the paper. Nothing joins the two.

The full detail is at the end of the form.

01 of 06

You

+01 What level is your role?*
+02 Which area is your role in?*
+03 How many engineers work in the teams you lead or support?

02 of 06

What actually changed

These three questions test the assumption the rest of the survey rests on.

+04 How widely are AI coding tools used across the teams you are accountable for?
+05 Since adopting those tools, how has the volume of working software your teams produce changed?
+06 Since adopting those tools, how has the time from deciding to build something to it being live for users changed?The whole calendar, not just the part spent writing code.

03 of 06

Where the time goes

Think of one piece of work your teams delivered in the last quarter. Not the best or worst, a typical one.

+07 Roughly what share of its calendar time went to each stage?
Almost noneA small shareA significant shareMost of it
Deciding whether to build it at all
Getting alignment and sign-off
Design and specification
Writing and reviewing the code
Waiting on another team
Testing, release and deployment
+08 Which of these have become MORE of a constraint since adopting AI coding tools?Select any that apply.

04 of 06

Deciding

+09 How long does it typically take to get a decision on what to build next?
+10 Of the work delivered in the last quarter, roughly how much turned out to be the wrong thing (reworked, reverted, or abandoned)?
+11 Has that share changed since adopting AI coding tools?

05 of 06

Seeing it

+12 How much do you trust what your work tracking system says about the current state of work?
+13 How much of your tracking data would you have to clean up before you could forecast from it?
+14 How does a delivery date get arrived at in your organization?The way it most often happens.
+15 When a delivery date slips, when do you typically find out?
+16 Out of the last ten commitments your teams made, roughly how many landed on time?
+17 Can you see in a single view what every team is working on and when it is expected to land?
+18 Roughly how many hours a week go into assembling delivery status for people outside the teams?

06 of 06

The shape of it

+19 Since adopting AI coding tools, where has your teams' attention moved?
Much less attentionSomewhat lessNo changeSomewhat moreMuch more attention
Feature delivery
Architectural health
Developer experience
+20 What is the single greatest challenge you face right now in managing AI-empowered engineering teams?

0 / 5,000

+21 If writing code is no longer your constraint, what is?Optional, and the one we are most interested in.

0 / 5,000

Last

Where to send the paper

Your answers are stored with nothing attached to them. No name, no email address, no identifier of any kind. We do not record your IP address, and this page sets no cookies.

We do note which link brought you here, such as a LinkedIn post or one of our communities, so the paper can say where its answers came from. It names a route in, never a person.

Your email is stored separately, in a database that holds only addresses. It stops the same person answering twice and, if you leave the box ticked at the end, tells us where to send the paper. It is never added to our mailing list unless you also ask for future findings, and confirm.

Nothing joins the two. We cannot tell you which answers were yours, which also means they cannot be withdrawn once submitted.

To have your email deleted, write to [email protected]. Everything else is in the privacy notice.

Free-text answers are read before publication. Where one names a person or an employer, we paraphrase.

Kept apart from your answers, as set out above.