ReadingBuilding a Copilot for PMs with Generative AI
blog

Building a Copilot for PMs with Generative AI

hero

Every Product Manager is driven by a desire to build innovative and powerful products to delight customers. However, more often than not, the reality of the job finds PMs caught up in multiple meetings, managing diverse stakeholders, and drowning in manual grunt work. In today’s PLG-driven environment, bringing the voice of the customer to the center of product development is more important than ever before. But in the midst of PMs having to keep up with operational work, the customer’s voice often gets drowned out.

How do we best set PMs up for success? The answer isn’t to simply throw more people at the problem and hope it solves the issue of bandwidth constraints. It’s to harness the power of AI to act as a partner to PMs and take on the manual grunt work, freeing PMs to create real value. A copilot for PMs.

From Project Management to Product Management

We envision a world where PMs aren’t stuck in meetings all day long, where they aren’t preoccupied with operational tasks like resource allocation, and where their main responsibility isn’t project management. We asked ourselves, “What if PMs could become redundant in customer meetings, sprint planning, and marketing communication, but still get more powerful insights and execute faster than ever before?”

With our native GPT-powered Turing AI, this is no longer a utopian idea. We’re bringing in powerful AI technology to revolutionize the role of a Product Manager. Here’s a glimpse into the exciting new AI features for PMs we’re adding to DevRev over the next quarter.

DevRev OneCRM
DevRev OneCRM

DevRev OneCRM

Bring the customer's voice to product development with DevRev. Purpose-built for modern software companies

Spend less time on sprint planning

A typical product roadmap consists of multiple moving parts and a delay in even one of these action items can result in the project not meeting its deadlines. Sprints remain the most popular way to organize developmental tasks. Managing and organizing sprints is crucial for PMs to execute product improvements and launches. However, the many ceremonies and protocols around traditional agile sprint planning can be daunting and may actually create blockers. At DevRev, we have used the power of AI to bring true agility back to sprints. As a PM, you can now focus less on project management and more on product strategy.

Here’s how to get started.

Summarize Sprints

With Turing AI, you can gain instant insights into the current state of a sprint. Use the ‘Summarize’ command to get real-time progress updates, view work distribution across parts, owners, stages, and tasks. You can also evaluate the performance of a sprint with metrics such as team velocity, blockers, and scope creep. This helps you understand at a glance what’s going well in a current sprint and what needs to be resolved. To dive deeper into sprint insights, you can type ‘smart queries’ into the conversational assistant. For example, when you type, “show me issues that are overdue by 2 weeks”, Turing immediately fetches and displays all issues that meet this criteria.

Predictive insights

With Turing AI, you can gain instant real-time insights into how a sprint and ongoing product enhancements are progressing and what the expected outcomes might be. Turing can predict whether work items in a sprint can be completed within the sprint timeframe by analyzing historical sprint data and existing bottlenecks. If it predicts an overflow, it can then suggest which work items can be moved out of the current sprint based on their priority. This saves PMs from spending valuable time on sprint analyses and increases performance outcomes from every sprint cycle.

Get a direct line to the customer

To create customer-centric products, PMs need to have a direct line to the voice of the customer. However, this isn’t usually the case. Today, PMs need to sift through mountains of Slack messages, support tickets and survey results to analyze them and infer customer insights. This information overload can make it difficult for PMs to make a distinction between requests for features that are already in the roadmap and high-priority items that need to be acted upon immediately. With Turing AI, PMs will soon be able to spend less time in meetings and classifying customer inputs. Here’s a peak into the AI and ML workflows we are building to automate the process of classification of inputs.

Ticket creation

  • Automatic processing: Detect the overall urgency of an action item using sentiment analysis and score it between 1-100 and send PMs a summary
  • Deduplication of action items: Turing AI ensures every action item from a customer is picked up only once to avoid duplicate tickets being created
  • Auto-creation of tickets: For every action item from a customer, Turing AI suggests a ticket to be created and auto-populates ticket attributes like title and description from the action item summary.

Classification and clustering of tickets

Turing AI can automatically classify tickets into buckets like:

  • Bugs
  • Feature requests
  • Adoption
  • Acquisition
  • Onboarding
  • Retention
  • Incident
  • Alert

It also goes one step further by crawling the web to get more context about incoming requests and then looks for synonyms or words that are interchangeably used to maximize accuracy of ticket classification. For new feature requests, the system automatically suggests enhancements to be created and for bugs, it suggests issues. Both new enhancements and issues are auto-populated with a title, description, owner and Part. As a PM, this reduces a huge amount of manual work that would usually go into creation and classification of tickets.

Prioritize work better

While most product management tools provide a purely subjective prioritization framework, DevRev uses generative AI to deliver a data-driven scoring method. We combine subjective scoring with a scientific score that encompasses product metrics (usage, adoption), customer metrics (retention, CSAT) and business metrics (revenue per user) to arrive at a more holistic prioritization matrix. PMs can simply chat with a conversational assistant using natural language queries to find high-priority items that need to be addressed. This enables them to identify critical areas for improvement and prioritize efforts accordingly. No more relying on ‘gut feeling’ during planning!

Never write a release note again

Every PM knows they should create release notes, but how many really enjoy writing them? Communicating product releases is essential to let internal and external stakeholders know about feature updates, bug fixes, and product improvements. With Turing AI, creating these release notes is one less thing PMs have to worry about. Turing uses natural language commands to auto-generate release notes in a matter of seconds.

With DevRev, creating a release note is as simple as opening Turing and entering a ‘create release notes’ command for an Enhancement. Turing then analyzes bullet points and issues under that specific issues and drafts a release note using this information. Want to add some of your own personality to the release note? Using natural language commands, you can ask Turing to change its tone, add some humor, and emojis until the draft has your unique voice. Finally, when it’s time to share your release note across channels, Turing can customize the formatting and tone to suit different channels like Slack, email, website, customer portals, PLuG, etc.

Allocate resources efficiently

With so many moving parts in each product, it can become difficult for management to track and allocate resources Turing AI provides a birds-eye view of resource allocation across different parts of your product. This allows CPOs and CTOs to get a top-down real-time view of individual contributors’ workloads and marshal resources effectively. Turing can analyze team members based on their skills, past performance, task dependencies, and current capacity to decide the most optimal distribution of work and answer questions like:

  • Do we need more team members on board or fewer?
  • Why are there dips or hikes in performance trends and what do we do about it?
  • Do we foresee any risks or roadblocks that may impact performance?
  • Which parts of the company or product do we cut and which ones do we nurture?
  • What is the quantum and value of work each team member is contributing?

Auto-create product architecture

Every product is made up of Parts. Parts are the fundamental glue that tie every piece of your product (tickets, work items, enhancements, etc) together. Trails is a unique feature in the DevRev app that lets you visualize every Part of your product and issues associated with it at one go. Turing AI can now discover your company’s product catalog, segregate it into Capabilities and Features, and then render a draft version of Trails based on this product architecture. To improve the accuracy of the auto-generated Trails, you can provide feedback and edits, after which the Trails will be regenerated. In just a matter of a few minutes, your DevRev app knows your product as well as you do!

Generative AI opens up massive potential for PMs to become more customer-focused, productive, and purpose-driven than ever before. Our own PMs are constantly exploring the possibilities of Gen AI to make DevRev the ultimate copilot for PMs and this is only the start. Explore DevRev and start building the next generation of customer-centric products. Get Started