What will you find here?
This is a toolkit for scaling your GenAI strategy.
It has been developed based on learnings from various contexts and contains explanations, tasks, and examples that will help your team become powered by AI.
This material is a compilation of suggestions and tips you can follow to scale your GenAI strategy.
Come back to it every time you have doubts on what is the next step you should take.
Also, don’t feel you have to read the slides in a specific order. Evaluate what is your current challenge and go to the slide that will help you solve it.
How to Scale Up CI&T Flow?
From our experience, scaling GenAI is a 3 acts process:
But what does this means in pratice? Let’s start explaining the first act goal in a simple but more detailed way:
Act One Ignition
Before thinking about performance improvement and new ways of working, it is necessary to think about adoption. Adoption means that people are actively using GenAI in their daily work, therefore they are learning what are the possibilities of this new technology. During ignition we rollout GenAI in a structured way, ensuring that everyone is learning and applying proved use cases.
1. Strategy & Roadmap
jboubpjnpijpjklnpikn[omlknpubj0vbp
Kick-off with Leadership
Why is this the first step?
The strategy is not going to work if the leadership is not actively involved. They need to know what is being done and why, because they are the ones that will allow the teams to do it. The leadership will need to ensure that everyone in their teams have time to experiment and learn about GenAI and /Flow, and they are also the ones able to communicate the urgency of the matter.
⏮️ Before the Kickoff
Make sure you have a good understanding of the metrics that will be followed.
Also, take a look in the whole Ignition Framework to have an overview of the work that will need to be done.
▶️ During the Meeting
Use this meeting to align the speech with all leaders! Clarity prevents suspicion. Here are some important topics to cover:
GenAI is not about downsizing teams, it is about increase productivity and deliver more
GenAI not about replacing people, It’s about evolving towards a new way of working and developing new capabilities with the time that is saved using GenAI
GenAI not about increasing workload, it is about making repetitive tasks less burdensome
And remember! The plan is about GenAI. Don’t focus your speech in pushing people to use Flow, make people understand the benefits GenAI can bring and they will want to use Flow.
Defining the Enablement team
Adoption strategy responsible
Make it clear who will be the leadership responsible for pushing the adoption process.
Every leader will be responsible for making his/her part, pushing the team and taking actions to ensure the rollout of the GenAI strategy, but from our experience it is important to have one leader that will be pushing the follow up meetings, that will bring news from /Flow, that will organize PDCA and that will be up to date with the news from the GenAI universe.
It is extremely important that this person is engaged with the subject and has time to include this responsibility in his/her priorities.
A project without an owner will definitely be forgotten in the list of priorities we all have.
Flow expert
Flow Experts are the people that will most support you during the adoption process. Our suggestion is that they should be leaders on your contract, because they will need the skills and autonomy to help you rollout actions.
What are their responsibilities?
Have a deep understanding of how the Flow Engines work, what are the best scenarios to use each Engine and know how to configure them.
Create new agents that are tailored to the contract’s context.
Support the whole team in the use of Flow.
Ensure the rollout of adoption strategies in the team.
Participate in GenAI PDCAs to help define next steps
Attend meetings with the Flow Adoption Team and bring the information back to the team
Important reflection about Flow Experts
Besides knowing about Flow engines, prompt engineering and having the creative thinking to create new agents, the Flow Expert will need to help engage everyone in the contract and support the change process. So don’t look only for someone that is a great engineer, look for a proactive and charismatic person, that will be able to help with engagement actions as well.
Map Opportunities & Gap
Workflow analysis
Before thinking about using GenAI, you must understand how GenAI can help you. So start by doing a SDLC analysis (Software Development Life Cycle).
- Talk to people from every role necessary to deliver a piece of software in your context and ask them to tell you what are the steps they do to complete their work.
- During the conversation people will tell you what are their pain points and their bottlenecks, if they don’t you must ask. Take notes.
- Don’t make any decisions before talking to everyone and have a complete SDLC mapping.
- After having the complete analysis, bring your Flow Experts together and discuss what are the pain points GenAI can help solve.
Not every problem will be solved by GenAI.
When analyzing your opportunities remember two things:
- Not every problem will be solved by GenAI.
- To solve a problem with GenAI, you must create a new perspective.
Remember that even though GenAI is very versatile it is not the solution for every problem and we must find the correct solution for each problem.
But also, some problems might not look like they need a GenAI solution at first, but when we change perspective and understand how to leverage GenAI, we might find new ways to solve problems.
In summary, be rational and smart on deciding which problems to tackle.
Define Goals & Priorities
Setting clear goal
Establish Measurable Objectives:
Set specific, measurable, achievable, relevant, and time-bound (SMART) goals to capitalize on identified opportunities.
Examples: Increase adoption rate by 20% over the next quarter, install 100 new agents by the end of the month.
Align with Business Objectives:
Ensure that goals align with broader business objectives and stakeholder expectations.
Prioritizing opportunities
Impact Assessment:
Evaluate the potential impact of each opportunity on business operations and customer satisfaction.
Use criteria such as cost-benefit analysis, resource availability, and risk assessment.
Stakeholder Engagement:
Involve key stakeholders in the prioritization process to ensure alignment and buy-in.
Kick-off With team
Why is this important
Adoption of GenAI is more about change management than technology. And change management is a lot about communication and leadership.
You must use this memento to get the buy-in of the people that are still skeptical about the subject and to align practical next steps with the ones that want to join the challenge.
Check below the steps to take before Team kick-off
Minimum topics to cover during the meeting
Why is GenAI important to us.
Why is GenAI important for each individual collaborator.
How the adoption metric will be calculated.
What are the current engines available on Flow.
What are the agents that the teams must start using.
How to access the correct tenant.
How teams will be trained in the usage of /Flow engines and agents.
What is already happening inside the contract concerning GenAI.
A final callout with next steps.
Set up a broadcast communication channel
The objective of this channel is to provide a centralized platform where team members can receive official updates, tips, ideas, examples, and announcements related to AI initiatives within the team context.
Channel Structure
1. Channel Name and Purpose
- Name: Choose a clear and descriptive name for the channel (e.g., “AI Insights”).
- Purpose Statement: Clearly define the purpose of the channel in the channel description. For example: “A space for sharing tips, ideas, resources, and announcements related to AI in our development team.”
2. Types of Content to Share
- Tips and Best Practices: Quick tips on how to effectively use AI tools and techniques.
- Example Use Cases: Share specific examples of how AI is being utilized within the team or industry.
- Event Announcements: Notify team members about upcoming meetings, webinars, and workshops related to AI.
- Resource Sharing: Provide links to articles, tutorials, and relevant tools that can aid in AI learning and application.
- Discussion Prompts: Encourage team members to share their own experiences and questions related to AI.
3. Guidelines for Communication
- Frequency: Establish a regular schedule for posting updates (e.g., weekly, bi-weekly) to maintain consistency.
- Engagement Encouragement: Encourage team members to ask questions and share their own insights or resources.
- Moderation: Assign a moderator or a small team responsible for managing the channel, ensuring that content is relevant and responses are timely.
Benefits
Improved Engagement: Regular updates can keep the team motivated and engaged with AI initiatives.
Centralized Information: A dedicated channel prevents information from getting lost in personal messages or unrelated discussions.
Enhanced Collaboration: Facilitates a culture of sharing and collaboration, allowing team members to learn from one another.
Defining and understanding the Adoption KPI
Headcount
Exploratoty User
Execution of at least 1 meaningful action* on Flow per month
Active User
Execution of at least 20 meaningful action on Flow per month (on average once a day).
Heavy User
Execution of more than 100 meaningful action on Flow per month.
To calculate the Percentage of Adoption in a contract we will consider Active User per the definition below:
2. Training Enablement Team
jboubpjnpijpjklnpikn[omlknpubj0vbp
How to train a Flow Expert
A Flow Expert is expected to have knowledge about: GenAI, Flow Engines, Flow Agents, Flow Platform and Change Management.
To start the training, the Flow Expert must watch all available material in CI&T University.
The training levels currently available are:
Interaction with other expert
It is highly recommended for the Flow Expert to have conversations with other Experts. One of the best ways to learn about GenAI is to see how other people are using GenAI, so look for opportunities to talk with other Experts.
3. Flow Onboarding
giyhbljoyfgighv
How to give access to the tools
Access to your tenant
As soon as your tenant is created all the people in your contract will get access to Flow.
After creation, as new people join your contract, users with admin role can add them to your tenant and give permission to specific Engines.
How to manage engines permission
Flow platform permissions are managed by Engine.
Every engine has an user and an admin permission level.
When your tenant was created some people were giving admin roles to all the engines. These are the people responsible for managing the access of other team members.
Remember that people with admin role can edit and delete prompts, dashboards and configurations on each engine, so ensure you are managing the people with this role.
Important
To give access to an Engine, the user must be an admin in that specific Engine.
Getting start with Flow agents
Engines & agents
To understand what we are going to do it is important to know the difference between Engines and Agents.
Engines are the applications used to run our AI prompts
and agents. We are going to focus our initial efforts on Flow/Chat and Steps.
Agents are components designed to perform specific tasks using a combination of AI techniques.
Before starting
It is of the utmost importance that before start defining and configuring agents to your context at least one person that will participate in the process have watched the Flow Expert videos available. Those videos will ensure that you have all the needed information to make guided decisions.
giyhbljoyfgighvnipnipsckadpvj-mepokmlapk mejODBJDNPIKD
