AI in product management – perspectives from the legal industry, education, and entrepreneurship
Watch on YouTube
TLDR
In my recent conversation with product executive and former colleague Matt Coatney, we explored how artificial intelligence is transforming product management and innovation. The technology has evolved dramatically in the past decade, from fragile, expensive systems to powerful tools that integrate seamlessly into workflows. Product managers can leverage AI for everything from customer research and brainstorming to prototyping and workflow automation. While organizations must balance specialized versus general AI tools and address concerns like hallucinations and data privacy, the benefits for productivity and innovation are substantial. The most successful implementations focus on solving real customer problems and seamlessly integrate into existing workflows.
Key Topics
Using AI as a brainstorming partner to overcome creativity blocks
Accelerating product development with AI-powered prototyping tools
Integrating AI into product management platforms and workflows
Balancing specialized AI products versus general-purpose models
Managing AI hallucinations and verification challenges
Learning from AI adoption in risk-averse industries like legal
Impact on mentorship and professional development
Future trends: local AI models and data privacy
Introduction
In this episode, we had a free-form discussion. My guest doesn’t know what I’m going to ask him and I don’t know what he is going to ask me. Our goal is to make the discussion valuable for product managers, leaders, and innovators.
Joining me is a former colleague, Matt Coatney. We worked together on an important product for LexisNexis. I went on to teach graduate courses in innovation and coach product managers and leaders in organizations, while Matt got more involved in Information Technology, leading professional services and consulting operations for a few organizations as well as serving as CIO for one of the large law firms in the US. His career started in AI systems some 25 years ago and today he continues learning about and applying AI and is also is a product executive.
The Current State of AI in Product Management
Matt asked about my observations of the effects of AI, from the perspective of a product manager, entrepreneur, and educator.
Last year at the Product Development and Management Association (PDMA) conference, three separate sessions featured AI tools specifically designed for customer research. This wasn’t just theoretical discussion. These were practical applications already being implemented by forward-thinking product teams.
At the PDMA conference, I participated in a workshop led by Mike Hyzy, where we completed what would normally be a 3-5 day Design Sprint in just three hours. Our team consisted of four humans and one AI companion, which functioned as a fifth team member. The AI was operated by someone skilled in prompt writing who understood the product space.
What impressed me most was how the AI accelerated our work. When we brainstormed customer problems, the AI helped us explore details we hadn’t considered. It suggested unmet needs, offered additional perspectives, and helped us develop a comprehensive view in a fraction of the time it would have taken traditionally. By the end of those three hours, we had developed a solid marketing description for solving a customer problem and created a decent prototype—impressive results for such a compressed timeframe.
Practical AI Applications for Product Professionals
On a personal level, I’ve found AI tools like Claude to be valuable as brainstorming partners. Rather than trying to craft lengthy prompts with all my requirements upfront, I’ve shifted to a more conversational approach. Kicking ideas around with AI helps me overcome the inertia of starting a task.
Matt shared similar experiences, noting that this brainstorming use case is often underappreciated. While many focus on productivity enhancements like email responses and content generation, the creative thinking support is particularly valuable for entrepreneurs and product innovators. Many professionals today lack the “water cooler culture” opportunities to casually discuss ideas with colleagues, especially with remote work becoming more common. AI tools help fill this gap, providing an always-available thinking partner.
We also discussed the prototyping capabilities of AI. Matt mentioned tools like GitHub Copilot for assisting developers, and V0, which can build functional web applications directly from human prompts. These tools allow people with little or no coding knowledge to write code.
For entrepreneurs and product managers, these prototyping tools address a common challenge: developing clear user workflows before engaging software developers. Taking time to create detailed prototypes helps clarify thinking and identify assumptions that might confuse users. AI accelerates this process, allowing teams to get clarity on user experiences sooner and save significant time and money during development.
AI Application
26189_bb986f-32>
Value for Product Professionals
26189_317750-59>
Brainstorming partner
26189_7bf2d1-1c>
Overcomes creative blocks and inertia
26189_72a4ee-82>
Idea validation
26189_594cdb-10>
Tests concepts quickly without scheduling meetings
26189_a9b617-9d>
Prototyping assistance
26189_08e702-a9>
Accelerates creation of user interfaces and workflows
26189_4f4fef-22>
No-code development
26189_37a225-25>
Allows faster proof-of-concept creation
26189_3ec935-e1>
Design iteration
26189_089966-fa>
Enables rapid exploration of alternative approaches
26189_0f07b2-92>
AI-Enhanced Product Management Platforms
The integration of AI into existing product management tools represents a significant opportunity for enhancing team effectiveness. ProdPad, one of the more popular platforms for managing product management work, and many competitors, have recently added AI capabilities—currently called Co-pilot—to their toolkit.
What makes these integrated platforms valuable is their ability to serve as a central repository for product information. They help teams maintain alignment with overall strategy, track progress toward objectives, and understand user stories. With AI enhancement, these platforms can now help identify gaps in strategy alignment, surface unmet customer needs based on existing data, and answer questions from stakeholders outside the immediate product team.
Matt and I discussed an important consideration when choosing AI solutions: whether to invest in specialized AI products or use general-purpose AI with custom prompts. Many specialized tools are essentially using the same large language models as ChatGPT but with carefully engineered prompts and workflows tailored to specific use cases.
For organizations making these decisions, Matt shared insights from his experience co-leading AI initiatives at his law firm. They’ve taken a tiered approach, using one solid general-purpose language model for most applications while investing in legal-specific AI products for their revenue-generating lawyers. For highly specific tactical use cases, they evaluate additional specialized tools—but only when the return on investment justifies the significant cost.
AI Tool Approach
26189_df0bd7-91>
Pros
26189_1b6ce5-9f>
Cons
26189_35f7e3-a5>
General-purpose AI (e.g., ChatGPT, Claude)
26189_671b22-69>
Lower cost, versatility, continuous updates
26189_b78775-e9>
May require custom prompt engineering, less specialized
26189_534839-66>
Industry-specific AI solutions
26189_26f06d-66>
Optimized for domain-specific knowledge and workflows
26189_006306-e3>
Higher cost, potential duplication of capabilities
26189_739c7b-d2>
Integrated platform with AI features
26189_09f8da-3a>
Seamless workflow integration, centralized data
26189_e91090-10>
May have less advanced AI capabilities than specialized tools
26189_3a6b54-1d>
Custom-built internal AI tools
26189_c3110f-c4>
Precisely tailored to organization’s needs
26189_5b53ee-ab>
Resource-intensive to develop and maintain
26189_235583-a4>
Matt emphasized that the bar should be high for investing in specialized products when general tools can accomplish 90% of the required tasks. Organizations must consider whether the additional 10% improvement justifies spending five to six figures on multiple specialized tools, which could quickly add up to a million-dollar investment.
Workflow integration is important for successful AI implementation. Matt provided an example: if an organization has employees manually uploading invoices to ChatGPT, extracting data, and re-entering it into systems, they’re missing efficiency opportunities. The real value comes from automating these workflows to minimize manual steps.
Implementation Focus
26189_6ce0d0-41>
Key Considerations
26189_47c3fa-69>
Model Selection
26189_a0e861-84>
Balance between general and specialized capabilities
26189_99570f-79>
Integration Level
26189_95e9ea-54>
How seamlessly AI fits into existing workflows
26189_873300-59>
Data Strategy
26189_69e331-e2>
What information AI can access to maximize value
26189_931d13-eb>
ROI Analysis
26189_9bffc8-c3>
Justification for specialized AI investments
26189_8a07aa-6a>
User Adoption
26189_ce898e-29>
Support for different user groups based on needs
26189_30c89b-66>
Matt also reflected on the current state of AI capabilities. He noted that today’s models are becoming so robust that it’s increasingly difficult to find use cases they can’t handle, at least for English language applications. This growing general applicability raises the bar for specialized solutions to prove their value.
We briefly touched on the debate around artificial general intelligence (AGI), acknowledging that while the term itself may be somewhat ambiguous, the general applicability of today’s AI tools is already impressive. This evolution has significant implications for how organizations approach their AI strategy, suggesting that for many use cases, the focus should be on integration and workflow rather than pursuing incrementally more powerful specialized models.
Addressing AI Challenges and Concerns
Implementing AI in product development isn’t without challenges. Matt and I explored several concerns that organizations must address to effectively leverage these tools.
We discussed the potential impact on professional development, particularly in apprenticeship-model professions. If senior staff rely on AI instead of junior team members for certain tasks, how will those juniors develop expertise? Matt raised this concern for lawyers and software developers, and we discussed its relevance for product management as well.
However, we identified a potential upside: AI could handle routine tasks that would previously occupy junior employees’ time, freeing them to engage in higher-value learning experiences. By eliminating basic tasks like fixing simple coding errors or catching obvious document mistakes, AI might actually create more meaningful mentorship opportunities focused on strategic thinking and core professional skills.
Product managers should be aware of risks of AI hallucinations
AI hallucinations—where models generate plausible but incorrect information—remain a persistent challenge despite recent improvements. I shared a personal experience using AI to analyze a detailed lease agreement. While the AI successfully identified unfavorable clauses and accurately referenced their location in the document, I was initially concerned it might fabricate issues. In other contexts, I’ve frequently encountered hallucinations where AI adds information not present in the source material.
Both hallucinations and omissions pose serious risks, particularly in contexts like legal work. Matt’s firm strongly advises lawyers to verify all AI outputs, treating them as they would work from a junior associate. As Matt put it, the AI is like “a first-year associate that doesn’t sleep and is always there,” but still requires careful review.
Type of AI Error
26189_76bd13-79>
Risk
26189_403eaf-90>
Mitigation Strategy
26189_91b91e-07>
Hallucination
26189_f8260e-a8>
Introducing incorrect information
26189_36b002-f0>
Verify all outputs against source material
26189_1ee839-84>
Omission
26189_37da4d-85>
Missing critical information
26189_f5312b-05>
Verify outputs and use multiple prompts to ensure comprehensive analysis
26189_2e4ae0-58>
Misinterpretation
26189_9ad1be-a1>
Drawing incorrect conclusions
26189_957aea-db>
Apply domain expertise to evaluate outputs
26189_130d4a-b7>
Over-confidence
26189_fda92a-94>
Presenting speculation as fact
26189_780dc0-13>
Require citation of sources for key claims
26189_45a9b8-67>
For organizations implementing AI, establishing appropriate guardrails is essential. Matt described how his firm has developed policies that provide guidance without outright prohibiting most AI use cases. They focus on education about appropriate usage contexts, data confidentiality protections, and verification requirements, creating a balanced approach that manages risks while capturing benefits.
AI Adoption in Risk-Averse Industries
The legal profession offers insights into how AI transforms traditionally cautious industries. As a product management professional, I pay special attention to adoption patterns in risk-averse sectors—when they embrace new technology, it often signals well-established value and manageable risks.
Matt shared a progression of attitudes toward AI within law firms over the past two years. Initially, many leaders experienced fear about AI’s potential to disrupt their profession. This concern was both practical and financial: AI tools represented a significant investment while potentially reducing billable hours by increasing efficiency—a challenging value proposition in an hourly billing model.
Over time, with education and exposure, these perspectives evolved into more nuanced views. Today, many firms, including Matt’s, are bullish on AI’s capabilities within appropriate boundaries. Some see it as a competitive differentiator, while others pursue AI implementation to avoid falling behind competitors.
The adoption curve among individual lawyers follows patterns familiar to any technology implementation. Matt observed that his organization has moved beyond early adopters and is now entering the early majority phase. While some users try AI briefly before abandoning it, those who integrate it into their workflow show steadily increasing usage over time.
AI Use Case in Legal
26189_b45643-e3>
Description
26189_81bddb-46>
Value
26189_eb54e9-99>
Content summarization
26189_e0ee99-d9>
Condensing contracts, briefs, proceedings, and statutes
26189_3c154c-db>
Saves time on document review
26189_a64ae4-a4>
Synthesis
26189_fab9bf-d4>
Combining information from multiple sources
26189_a03f6e-0e>
Creates comprehensive understanding
26189_b6c5a2-7b>
Drafting assistance
26189_973e14-cc>
Generating initial document drafts
26189_668b21-42>
Accelerates document preparation
26189_45dc85-37>
Strategy brainstorming
26189_8f4f9b-08>
Exploring alternative approaches and counterarguments
26189_0453c3-40>
Enhances case preparation
26189_43aa64-3a>
Provision analysis
26189_ba4752-d2>
Identifying favorable/unfavorable contract terms
26189_d9f7a5-3a>
Improves negotiation position
26189_c95f7c-46>
What particularly impressed Matt after 25 years in the AI space was how dramatically the technology has evolved. The systems he worked with 15 years ago were fragile, expensive, rules-based, and easily broken when applied to adjacent use cases. Today’s models understand language nuance, adapt to specialized terminology, and apply reasoning to novel situations—capabilities previously thought to require years more development.
For product managers serving risk-averse industries, this evolution suggests several insights: emphasize verification and human oversight in your AI implementation, focus on specific high-value use cases with clear ROI, and recognize that resistance often transforms into enthusiasm as users experience benefits firsthand. The legal industry’s journey provides a roadmap for introducing AI into other conservative sectors, from healthcare and finance to government and education.
The Future of AI in Product Management
The most successful AI implementations in product management will be those that seamlessly integrate into existing workflows. Matt and I agreed that transparent integration represents the next frontier for AI tools, moving beyond standalone applications to become embedded features within the systems product teams already use.
This parallels our experience at LexisNexis, where we worked together on a product that integrated new capabilities without requiring users to change their behavior. I expect platforms like ProdPad to succeed by making AI assistance transparent and aligned with users’ natural work patterns.
By contrast, Microsoft’s Copilot approach in Office applications often feels disconnected from the actual workflow. As I mentioned to Matt, I frequently close the Copilot prompt when opening Word because it feels like an extra step that interrupts my process rather than enhancing it.
Our conversation also touched on recent developments in AI democratization. We discussed Deep Seek, which enables running sophisticated AI models on relatively inexpensive hardware—from Raspberry Pi devices to modest servers. This trend capability allows organizations concerned about data privacy and security to maintain complete control over their AI systems and data.
Matt predicted this will lead to a bifurcated market: cutting-edge models will continue to require substantial computing resources, while slightly older generations will become commoditized and available for edge computing applications. He envisioned an “AIOT” (AI + Internet of Things) future where smart devices incorporate local AI processing.
Future AI Trend
26189_c8b32d-ab>
Impact on Product Management
26189_762500-52>
Workflow integration
26189_33265e-24>
Reduced friction in adoption and usage
26189_d56176-6a>
Local AI models
26189_a4a5c9-43>
Enhanced data privacy and security control
26189_955655-d6>
Edge computing AI
26189_c1e643-b3>
New product possibilities with embedded intelligence
26189_45254d-af>
Democratized access
26189_882e3e-6b>
More accessible AI for smaller teams and organizations
26189_0a7ff9-af>
Specialized fine-tuning
26189_f9bccf-e5>
Tailored models for specific product domains
26189_e7e62e-68>
Beyond technical advancements, Matt expressed enthusiasm about AI’s potential social impact. He’s personally focused on applying AI to health and climate challenges. He noted that while large pharmaceutical companies are exploring AI applications, there’s tremendous untapped potential for nonprofits, NGOs, and small mission-driven organizations to leverage these tools.
This social dimension presents an opportunity for product managers to apply their skills beyond traditional business contexts. As AI becomes more accessible, product professionals can help mission-driven organizations integrate these capabilities into their workflows, potentially creating outsized impact through enhanced efficiency and effectiveness in addressing critical social challenges.
Key Insights for Product Managers
Throughout our conversation, several actionable insights emerged that product managers can apply immediately to their work with AI. We discussed the need to reframe executive demands for adding AI into customer-focused questions.
Matt and I both encountered situations where leadership teams push for AI integration without clarity about the specific value it will provide. Too often, senior executives make broad statements like “we need to add AI to what we’re doing” without understanding what that actually means for products or customers.
As product professionals, our responsibility is to translate these directives into customer-oriented questions: How can we enhance our product’s value using AI to solve problems customers actually care about? Will AI help solve these problems faster, better, or more comprehensively? This reframing helps ensure AI serves genuine customer needs rather than becoming a superficial feature.
Another opportunity Matt highlighted was AI’s ability to process unstructured data. He noted how frequently we encounter friction in our daily lives—retyping information into forms or manually extracting data from documents only to re-enter it elsewhere. These pain points represent prime opportunities for AI-enhanced products.
Customer Pain Point
26189_cb85d5-5e>
AI Application Opportunity
26189_41a141-4e>
Form completion
26189_83a80d-d4>
Auto-extraction of information from existing documents
26189_d2eae2-25>
Data transcription
26189_5b4336-47>
Converting formats without manual retyping
26189_17177c-9e>
Information synthesis
26189_236f21-9c>
Combining data from multiple sources automatically
26189_3c6c6e-cb>
Content transformation
26189_a5c3d4-34>
Converting between visual and text formats
26189_dec9e1-3f>
Pattern recognition
26189_f75355-73>
Identifying trends in unstructured information
26189_982c93-72>
Matt shared personal examples of AI’s potential, including using it to solve scientific puzzles from Scientific American and helping his daughter understand idioms in her schoolwork. These seemingly simple applications demonstrate how AI can remove friction points that we’ve previously accepted as unavoidable.
Conclusion
The integration of AI into product management isn’t just a passing trend—it’s fundamentally transforming how we research customer needs, prototype solutions, and create value. AI has evolved from a specialized, experimental technology to an essential tool in the product manager’s toolkit. The question is no longer whether to incorporate AI into product development processes, but how to do so most effectively.
As you incorporate AI into your product management practice, remember that the technology itself is just a tool. The true value comes from how you apply it to understand customer needs, solve meaningful problems, and create products that improve people’s lives. By focusing on these fundamentals while embracing AI’s capabilities, you’ll be well-positioned to thrive in this new era of product innovation.
Useful Links
Connect with Matt on LinkedIn
Learn more about Matt’s mission-based organization, 10 Billion Ventures
Innovation Quotes
“The best way to predict the future is to create it.” – attributed to Alan Kay and Peter Drucker
“The way to get started is to quit talking and begin doing.” – Walt Disney
Application Questions
How could you integrate AI as a brainstorming partner in your current product development process? What specific activities (customer research, feature ideation, user story creation) might benefit most from this partnership approach?
In what ways could your team use AI to accelerate prototyping without sacrificing quality? How might this change your current approach to validating product concepts before full development?
How could you transform executive requests to “add AI” into customer-focused initiatives? What specific customer problems in your product area could AI help solve more effectively?
How could your organization balance investment in specialized AI tools versus leveraging general-purpose AI models with custom prompts? What ROI metrics would you use to make these decisions?
How could you ensure AI integration enhances rather than disrupts your existing product workflows? What would seamless integration look like for your specific product and team?
Bio
Matt Coatney is a seasoned C-level AI and product executive with 25 years of diverse experience. His expertise includes: artificial intelligence, business growth, and product development. Matt has also supported a wide range of industries such as manufacturing, media, law, life sciences, government, and finance. His client list includes some of the largest, most well-known organizations in the world, including Microsoft, IBM, the Bill and Melinda Gates Foundation, Pfizer, Deloitte, HP, and the US government.
Matt writes and speaks frequently on technology and product topics. In addition to a TED talk and keynotes, his work has been published by MIT, HarperCollins, and O’Reilly and has appeared in books, journals, and international conferences. Matt’s latest book is The Human Cloud: How Today’s Changemakers Use Artificial Intelligence and the Freelance Economy to Transform Work, with Matthew Mottola.
Thanks!
Thank you for taking the journey to product mastery and learning with me from the successes and failures of product innovators, managers, and developers. If you enjoyed the discussion, help out a fellow product manager by sharing it using the social media buttons you see below.
Source