About this talk
This talk focuses on utilizing artificial intelligence as a project management co-pilot to streamline the daily workflow of project managers. The speaker, Jprit Singh, shares his extensive experience in project management and emphasizes the common challenges faced by PMs, such as information overload and communication gaps. He introduces the UDA loop framework, which stands for Observe, Orient, Decide, and Act, originally developed by Colonel John Boyd of the U.S. Air Force. Singh explains how AI can enhance each phase of this framework by handling vast amounts of data, thereby allowing PMs to make informed decisions quickly. He provides practical examples of AI tools that can improve efficiency, such as summarizing Jira data to highlight risks and important updates. The session aims to motivate PMs to adopt a proactive approach, reducing chaos and improving client relationships through smarter decision-making.
Full transcript
Hello everyone. First of all, welcome to this session. Good afternoon everyone. My name is Jprit Singh and I'm going to present this session related to the AI as a PM co-pilot. But before jumping ahead with the further things most of you uh the audience in front of me, I would like to understand like who is the project manager over here? Who is in the leadership position? That's
cool. Scrum master. That's good. So most of you would be more familiar about the the things like for example you might have uh 50 unread messages since morning or couple of days back uh few kind of the 20 plus emails which are unread um and kind of the recently you receive the the client email after the 7:00. So that kind of the incidents which happen usually and
PNS who work really hard and but usually face these things on the regular basis on the daily basis. So in this session I would like to help you out in order to work in a very harder way but how can we make the PM's life or the leadership position's life in a very easy with the help of air and how to use it in a very smarter
way. So quick words about me um outside the work uh I'm a passionate cricket fan and I'm playing cricket for the last 20 plus year as well and this is my hobby but why it's a hobby because I'm learning cricket um learning a lot from the cricket as well and trying to implement the same thing in my regular project management life as well. How? Because in the
cricket the most important factor is the kind of playing under pressure. Playing under pressure is not a case like you have a the best team on your side and you will play well. It's not a case. You have to play smarter than the other opponent. Then after you'll be able to work in the right way. Right? So apart from that I'm a blog writer. I have written
many uh blogs on my LinkedIn profile and uh it's not related to kind of the the authorization but kind of uh written the blogs on my specific PM related things the incidents which I'm facing how I have resolve that what are the problem statements which project managers are usually facing so I have already written over there so professionally I have around 15 plus of experience in the
various trains um initially started my career with the QA scrum master move ahead with the project management then after I moved to the engagement manager role so I'm standing in front of you this to talk about the real things real sprints real blockers and how to solve that this is my journey since 2019 um so far this uh Drupalcon Chicago is my 13th Drupalcon con and out
of that overall it's a 13th. It's a sixth Drupalcon over here as a speaker and seven as a Drupal cans. So it's overall So here I'm going to cover up the few things uh starting with the problem statements. uh the daily routine of the PM how what are the cases they are facing on the day-to-day basis and uh how AI like involving the multiple AI tools u
doesn't help them how to help them the kind of I'll introduce the UDA loop framework how it will help them in order to understand in the right way and implement in the right way then after I'll talk about the heart of talks uh heart of faces uh the four pillars UDA It's a stand for observe, orient, decide and act. So we'll uh we'll share the real prompts
which will definitely be helpful to you and I'll share the more insights, the more things the next milestones which I'm planning to achieve as well. So I'll share over there as soon then after a few uh sprint crisis scenarios which will help because you will definitely relate to that because that's the reason I asked like who are the PMS like who can relate these scenarios. So sprint
related challenges, sprint crisis. So these are the common things common hiccups. So definitely you will understand those thing and relate with that. So first question to all of you uh raise your hand if you have ever worked walked into the sprint planning meeting uh still you are not sure what are the real priorities. Let's go. Keep your hand raised like if you have made a decision on
a project and found out later you were missing a crucial piece of the information. How many of you spent more than 30 minutes in uh this week just trying to figure it out what was actually happening on your project before you could even start deciding to do that's amazing it's less than so now what if you could get all that context uh in under the 2 minutes
every single morning so how it will work in the real life. So let's talk about that. The modern projects are the chaos. Let me describe you. Uh in the the Monday morning of the Drupal project, I'm taking one example. I had a project the last year where there were kind of the three days from the demo and my developer told me that the before in the standup
like uh there is the one module which are relying on the demo and somehow that module has some sort of conflict. He uh he messaged me in the slack somehow I missed it and after the two weeks during the just be before the three days of the demo he messaged me that like I've already messaged you you did not responded to that so most of the time
like it happens because there are a lot of channels it will be very difficult for us in order to scan every day each and every slack thread so that definitely makes some sort of impacts over there so it was buried in the long threads then I spent around 2 hours in order to find out then after trying to understand like how it can be fixed and or
trying to mitigate the risk how it can move forward. So that's not technology problem. That's the one thing which we need to understand. So that's a kind of a signal of noise problem. And how can we avoid that? We have a framework. So before jumping ahead with the framework, I would like to give you the couple of examples of the firefighting and the steering mode. So one
is on the left hand side. There's a PM1 and another is a PM2. So PM one like who do the who do a hard work both PMs are the hard works hard workers and the first one is a kind of the decision took the taking the decision in the incomp complete context spending Monday mornings in order to evaluate all the slack threads spending around 1 hour in
order to understand all the emails reading out the slack unread messages calling out calling the extra meetings in order to understand the context in a better way and then after trying to evaluate what is the risk. So as per my understanding he is a kind of messenger who is passing the things from one place to another instead of leading from the front. So that is the one
example. Moving ahead with the next one which is a kind of the steering one. They talk every they talk in the end to the every meeting already knowing the context what is currently happening in the project. They catch the risk early. their planning sessions are short because they are well planned before jumping ahead into the sessions. That is the kind of the strategic leader who understand the
things up front and then after mitigate the risk up front as well. So the difference between the these two PMs is usually not the talent. That's one thing which you need to understand. I have seen incredibly talented PMs spend their whole week firefighting because they didn't have a way to filter out the signals to the signals from the noise and the sharing PM has exactly the same
project same client same conversations are happening but they stop finding out like they are spending around 5 minutes just before the starting of the day before jumping ahead to the DSM and it will it will fix it out. So I'll I'll share the few examples few prompts how it will help the PS. So false fix uh adding the more tools it's a kind of the assumption like
we will use multiple AI tools that will definitely solve your problems because there in the market there are a lot of AI tools available. First thing which we need to understand that AI is not going to replace with project manager. AI is there in order to support the project managers. how it will be supported because it is a powerful signal uh extractor but only if you give
him give it the framework to work with it. uh I'll try uh this as a myself like uh I started asking the AI to summarizing my Jira every morning and I got back a beautiful formatted list of 50 50 threads 50 tickets it was accurate it was complete but a had given me the everything without any judgment without understanding what's the priorities so definitely a person should
be there who can evaluate the things in a better way and then after pass on the priorities to the relevant team maps So that where the UDA loop works here it comes. So before jumping ahead to the next one which is a kind of what is UDA how it is working. So let's take one step back. So UDA loop is like was created by the US air
force colonel called John Boyd in 1950 and he was studying the fighter fighter pilots specifically why American pilots were winning dock fights against the technically superior aircrafts. His answer was that they could cycle through decisions faster than the enemy. I gave you an example of the cricket as well. If you will understand the enemy, take the decision in a faster way uh before the opponent, then definitely
you'll get the win-win situation. So, Boyd argued whoever can complete this cycle faster than the opponent wins. Not because they are the stronger but because they are the quicker to adapt. So in the project management life, why does this matter to the for the Drupal projects? Because Drupal bricks are exactly the kind of the complex fast changing uh high dependency on the environment shifting client priorities cross
team dependencies new requirements comes in. So almost we are also moving in that direction as well. So PM who can observe fast orient decide with the more confidence and act more clearly that is the kind of the PM's win. So AI is a kind of the co-pilot that will make the each phase the loop faster. So John B who is the creator of the UDA loop I
specifically mentioned about like the speed of the decision making is often the decision uh deciding factor between the winning and the losing. So in the fighter jet that's a life of matter of the death and the life and in the Drupal project sprints it's about the timeline client relationships and your team. So the loop is a kind of the same that is the one quote which they
are specifying. So here's the framework uh the formula which I would like to you to take away. So AI I'll start from the left to right. So AI first of all handles the volumes of data. It provides the endless streams of the Jira data slack threads uh emails. So at least it will give you the complete information with the help of the chat. You supposed to apply
the judgment. You navigate the trust client relationships and things. No AI can do UDA provides the order a repeatable cycle. So you you are never starting from the scratch. Then after the together the co-pilot so it's a kind of the augmented proactive intelligence you are still the pilot. That is the one thing which you need to understand because AI is in the side seat next to you,
sitting side next to you but handling only the instructions from your side. So let's talk about the the pillars one by one. Starting with observe. Let's uh I'm giving you one example. Uh step one is observer. The goal is to gather the data from all the projects channels without spending around 45 minutes because if there are 10 channels I need to figure it out all scan all
the channels definitely 30 40 minutes would be required in order to gather that information. So what's the practice will look like? So first of all the daily jira summary that will flag me what's a stuck what is the the risk what has been changed so so far the slack channels will surfaces all the information which where the relevant information is there necessary questions are there raised by
the team member which are the kind of the bottleneck for the sprints goal for the observer the email scans as well but goal for the observer is not to read everything that's one thing the goal for is to arrive you at the standup already knowing the state of the project not learning it into the real time how it will be possible because you are the project manager
you're leading the DSMs but when you jump on the DSM then after you get the better understanding who is doing which thing so how you would be able to devote at least the 5 minutes just before the DSM to get to know how it what's the information you need to grab just jumping ahead before jumping So this is one example. Uh I have shared one a few
examples on the the chart GPT how it will be helpful but there are the other ways as well. So at the last I will share those examples too but this is one example in order to showcase a prompt in the practice. You can see in the in the left hand side the prompt I have already shared the examples. You need to pass on the Jiraa bulk for
the one week of data or the one sprint data with the latest data. It will hardly take around in order to get the bunch hardly a couple of minutes. Then after it will filter out the list of the tickets for the not moved in the last two days. Flag the tickets which are marked as blocked or highlight the priorities changes since Friday. So count is in the
progress to-do and the done. So it will give you the data in this way. So hardly it will take around 3 minutes or the 2 minutes in order to execute the script and get the data. So my only intention is that this data a hardier 2 minutes or 3 minutes will help you to understand how the things are going in your project. If there are there's a
need to change something at least you are already aware about that. If there are blockers instead of team members are sharing something to you as a blockers you are already aware like these are the blockers which you need to address up front and then have to move to the other priorities. Another example select thread sector prompt there are multiple channels it will be very difficult for you
to in order to review all the select threads getting the information what are the crucial factors which you need to understand what are the risk which team members have already mentioned. So far it would be very difficult for you to understand uh if you will devote around 45 minutes 1 hour in the morning then after you will get that idea even though my personal example is that
like I I used to see only the the priority channels in order to understand what are the latest priorities or the latest risk which I need to take care of but sometimes it is getting missed because of the priorities but this is one example if you will use create it on the chat or the cloud that will help you to get the data in the right way.
On the right hand side you can see the the second example where the unanswered questions by the Aisha. So that's a kind of the which is a kind of new demo runs no reply yet from the devops team. So that is a kind of where I need to it's a trigger point for me in order to make sure at least it should be addressed up front then
after I'll move to the next. The next is orient. So now you have already observed the data you got the raw data as well. So now it now this is the time you need to understand they they have provided all the data. For example the 20 number of items are there 10 number of items are blocked but how to make sure at least it should be easily
conveyed to the team members because you won't be able to share all the summary to the team members. it would be overwhelming for them and for you as well. It will take more than time. So better it to prioritize in the right way. So in order to prioritize that like dependency mapping prompt that's the one example I'm sharing over here. If there are some sort of dependencies
which is blocking your sprint progress sprint planning or before jumping to the sprint planning there are few blockers which you need to take care of that at least that will be helpful. Next one is the pre- sprint planning pre-p prompt. So before jumping ahead to the sprint there are how many risks are there already in the the previous sprint. What are the the blockers? What are the
key areas which we need to celebrate? We have to showcase to the clients. Uh honest concern to raise to the client so that at least some sort of transparency can be maintained and suggestions the priority in the today's discussion what needs to be discussed. So these are the kind of the hardly one minute job in order to get the data if you have a raw data in
your head and I would like to give you one thing more like this these prompts are available on the LinkedIn as well. So I'll post this presentation on the LinkedIn so you will get all the prompts handy as well. There's a alternative way in order to handle that as well which is the kind of from claude you can create the agents which will talk to you and
which will work for you but you have to make sure to create in the right mindset. We have talked about the observer we have talked about the uh orient now we we need to take care about the decide. So there are the few examples which I have already shared in the in the mid-sprint check-in the client is expecting the live demo in the next 48 hours. This
is one example but on other side the my SSO credentials are not working. It's blocked by someone client has not shared that. So my demo is for the checkout without login the checkout won't work. So this is a kind of the blocker for me and how I supposed to make sure at least to decide in a right way. So at least the udal loop process should be
followed over here. Few examples again over here as well for the options for the trade-off prompt at least you should have some relevant options available for the client. For example in the first one which is the the mock login for the demo. Uh my SSO credentials are not working. Let's go ahead with the next one next step and create the one sample page the demo login and
then after move ahead and showcase the check out process. This is the alternative way. Second is the reschedule the demo for the 24 hours. Reduce the demo scope if there is a possible move from one place to another. Uh but it will be very difficult. But recommendation is the option A which will kind of feel very transparent. At least client will see the progress and move forward.
So at least you will not be blocked. Next is the midprint dep prioritization and you are in the middle sprint. client requested the new feature at least these prompts are already handy with you. You'll be able to uh give assurance to the client to dep prioritize at least one feature and move to the next one and let's handle the current exercise but at least a single prompt
will definitely help you like these are the things which can be mitigated. There's a one quote uh by the author is that like in any moment of the decision the best thing you can do is the right thing. The worst thing you can do is nothing. So AI gives you the context is to act. The decision and the courage to make it. It's always yours. Moving ahead
with the next pillar which is act. So act is a kind of the decision becomes reality. You you are you have observed you have oriented you have decided. Now you need to communicate clearly. Run to the tight. Stand up and stand with the client. send the client email that doesn't take around 30 minutes to write but if you have the relevant things handy in your place then
definitely you don't need to spend at least of five or 10 minutes as well in order to write an email so there are the collected as well the client update email prompt as well so you know better than others like what's your client expectations are what they are looking for at least you would be able to share some sort of relevant inputs. You want to send an
email to the client in a very empathetic way, feel free to drop a message over this because in the last you can mention the last three lines. Make sure it should be transparent. It should not be apologetic. At least these commands will definitely help you to generate the the email in that way so that at least the transparency can be maintained accordingly. Weekly status reports most likely
the most of the PMS are already working in that direction for the weekly ones but at least like what is the RG status uh this week what we have already accomplished risk next week what we are planning to cover up client task on the recent items the Jira raw data which you have already imported into the the chart GPT or the cloud that will definitely help you
to generate that report as well. So instead of handling all the things in a manual way that will help you out. So would I in action in the real est scenario. So here you can see that and I divide it into the four parts. Observe initially the PM scans the the slack email a very manual way uh took around 30 40 minutes. It's a very general time
which I'm just sharing over here with a live practical experience. And on other side if we will go with a co-pilot then definitely AI will surface and complete all the the the channels and will take around hardly a couple of minutes. Um on other side orient like it gives like allocate the the task to the team members the three resources will also be occupied over there but
on the other side the AI will map effective tickets and hidden dependencies will be covered up as well. Then after decide uh without AI decision delayed missing information more more meetings would be required in order to understand that on other side the PM decided the meeting and not the couple of hours are required as well. The last but not the least which is kind of the inact
way it's a very rush vague client emails we need to read out and then after reactive stakeholder call would be required but on other side it will be very proactive because upfront knowledge is there to the PM they would understand what they're supposed to convey to the client as a result the the few examples which are showcased for the demo demo is proactively not the reactively at
least you have already presented the things in a very transparent way. Uh client trust maintained team unblocked at the same day as well. Then another example from the reactive to the proactive approach. Um I want every PM in this room to make not uh from the bad PM to the good PM. That's one thing not from the inexperienced to the experienced one. From the reactive to proactive
that is the one thing which you need to change the chaos doesn't go away to be honest. The complexity doesn't go away. The shifting the client priorities doesn't go away. What change in your relationship to them? That's the one thing. You stop being the person who finds out about the problems after they have been already been hit. You become the person who saw them coming up front.
the new process deliverable measurable impacts um earlier risk detection that will come up front the shorter sprint plannings that will be there uh the fewer client surprises because that is the the most crucial one where uh in the middle of the sprint the new things new risk comes up so most of the PMs usually suffer on these things or struggling with these key areas so the most
important like I would like to address something directly because I know some of you are thinking like is going to replace the project manager but it the clear answer is no to be honest that's why because precisely the AI cannot build the trust to the client with the clients trust is the build like through the conversations collaborations communication and uh through being the person who says like
I have got this at 5 4:00 at the Friday evening and when everything is on the fire and meaning it to be honest because If uh I'm a PM I have I I've been there in order to answer to the client whether it is feasible whether it's not I can prioritize it dep prioritize it but definitely with the AI the data raw data which you have handy
that will definitely help you in order to materialize the risk up front with the So before jumping ahead with the Q&A sessions, I would like to uh leave you with the three starting questions to get the conversations going on. Uh what is the biggest blocker you face when trying to make the faster decision on your Any idea what is the big I can repeat that. What trying
to make the faster decisions on your projects? Yeah, >> I I think the availability of of the required information to to make the decision which is not always available in the right format to consume. >> Okay. And in order to mitigate that like client is the right source of person like who can provide you I it depends depending on the on the problem. It could be client,
it could be team, it it could be third party. >> Okay. When you said it it could be team that means some sort of information you are required from the team in order to get back or you will be updated without those things. Is that so? >> Yes. For example, I have a release coming up. Okay. >> Now I I need to know from from the team
is that if they have completed all the required testing, all the required steps for it to be pushed to to the production requirement. uh if if any of those steps did not complete well or failed or something that that is there. So naturally that that information is needed unless I have that I cannot make a decision that is it uh good to move forward or do I
need take a corrective action. So I believe if you have that data handy um taking an example because if we are talking about the testing strategies let's suppose you are if you are aware about the testing process like if it has been completed or not then after you will move to the release process the data if you have the like kind of the latest things uh what
is the how many items are blocked in the current sprint how many items where the the still the status is still the to-do ready for the development so at least that will help you out in order to get the understanding where the current position is, whether you would be able to release it or not, whether if you will release it will create a risk or not. So
will that help? >> Absolutely >> that will help. >> Okay. So did you ever try this uh as a prompt prompt engineering in order to understand that uh like >> yes >> to make the life easy. >> Yes. And again I mean there are multiple facets to it like you took the example of Jira right for example. I mean again uh we have like automated pipelines and
all that but but the thing is again pipelines have to be triggered >> based based on some event >> by the people now if if there is a break in the pipeline naturally not everything is directly available into Jira so that that information has to flow into Jira >> for for me to extract it and act upon it. Yeah. >> So, so there are those situations which
sometimes present challenges like for example the person we may not be available at that time plus with the distributed teams >> uncontrolled situations. >> Yes. It adds up to to the complexity to the% but I would recommend like to uh use these prompts try it out in case if you face any difficulties you can reach out to me LinkedIn I'll definitely support you over there. Another question
like has anyone has already tried using Jira AI for the Jira or Slack? Okay. It's a kind of the Jira integration tool AI tool you have used. >> I have used Robo I've used. >> I've used like uh Jira align which works on top of that >> to to to get insights. >> But in order to maintain the playbooks how how it will help the robo would
not be able to help you out over there. >> Uh kind of because in the conference if I supposed to update the playbook how robo would help you that's my question. >> So see I mean the robo basically takes the information from what is already there. So where I'm use the way I'm using robo is one is to extract the yeah >> uh the the current status
right what do we have second is if if I already have say uh epics or features defined and then how to convert them into say stories >> correct >> and is a story wellformed what is missing and all that those are the kind of things where I'm using to to do that comparative analysis >> that's amazing >> using robot >> uh that's Google so where In the
UDA loop, do you think you would get the most immediate value starting tomorrow? That's a question for all of you. Where do you feel like ODA will help you out on the day-to-day life? >> For me, probably observing and orienting myself. So, just consuming that information, understanding what's going on, what the context is, and then, you know, of course, then be able to make a more informed
decision. But >> that's amazing. Yeah, it's the chaos the taming the chaos is the hard part. >> Uh right now I have showcased the charge prompts uh that as a kind of pretty straightforward you can use that. And now the next phase like I'm I'm seeking my uh help from my team members as well like he's very good in the creating the agents that will definitely talk
to you like uh creating the multiple agents. one agent knows like to talk to in order to get the data. He's supposed to know like which agent I need to interact with. So that is a kind of the next milestone. Um so which where me and my team member are working. So definitely I'll post on the LinkedIn as well once it is done. That will definitely be
helpful. Any other who would like to give it a try the UDA? Uh >> and I think observe and orient would be the most helpful. Yeah. >> Because naturally we are attuned to making the decisions and then acting >> Which we do day in day out but finding the right data and and the right opportunity to act is important. So I think that's where observe and orient
>> Absolutely. The reason behind is that because sometimes with the flow we start taking the decision in a very reactive way. That is the the human tendency even though I'm usually do that sometimes. So that if we will start following this approach as a observe understand that in the orient then after take start taking a decision that will definitely help and believe me I'm using this uh
before jumping ahead in the scrum ceremony call in the DSM in the morning I used to see this data in the morning I tried to get some idea where my team is blocked out of 20 items there are two items which are blocked then definitely it will I need to unblock them. So that will So uh this is my LinkedIn scanner uh QR code. Uh if you
could uh join me on the LinkedIn, I'm planning to publish the this presentation and the post on the LinkedIn. So you'll find these all chat prompts over there. So that will be definitely be helpful to you. That's one thing you can reach out to me on the the drupal.org u I'm working with excellent so you can see my a lot of blogs over there as well and
the Twitter too. Thank you. And last but not the least it is from the duper side. Uh it would be great if you could share some feedback on this session. If there is any scope of improvement, if you feel I need to do I need to do for the next session, I will definitely be appreciate this thing. Thank you so much everyone for attending this session. Thanks
a lot. Thank you. >> Thank you.
