Re-Found: A Step-By-Step Guide to Becoming an AI-First Company

I recently read Ravi Gupta’s thought-provoking article “AI or Die” and found myself nodding along with his core thesis: companies that successfully integrate AI into their foundations will thrive, while those that treat it as a peripheral tool will struggle to survive.

He talks about “re-founding”, the idea that becoming AI-first requires rebuilding your company from the ground up with AI at its core. It’s not about adding an “AI strategy” slide to your quarterly deck or launching a token innovation lab. It’s about fundamentally reimagining how your business operates.

So how exactly do you become an AI-first company? That’s where this guide comes in. I’ve created a comprehensive framework to help you transform your organization into an AI-first company one methodical step at a time. This is based on work I’ve done with dozens of companies.

The Pyramid of AI Adoption

I previously wrote a post called the Pyramid of AI Adoption which illustrates how far along you are in becoming an AI-first company.

I suggest reading the full article but here are the Cliff’s Notes:

Stage 1: Augmentation – You’re using ChatGPT to write emails and summarize meetings. It’s like getting training wheels for your AI bicycle. Most companies are camping out here.

Stage 2: Automation – You’ve started changing how your company actually operates, automating away processes that eat up resources faster than I demolish a chocolate bar.

Stage 3: Innovation – You’re creating entirely new business models and products with AI that were previously impossible with your resources.

My aim in this guide is to show you how you can get to Stage 3. Of course, reading about it is the easy part. The hard part is implementing it! Let’s go…

Step 1: Immerse Leadership in AI

Transformation starts at the top. As a leader, you can’t just delegate “that AI stuff” to your tech team any more than Tony Stark could outsource being Iron Man.

Block out two hours every week to actually use these tools. I’m not talking about reading articles about AI (unless it’s my blog). I’m talking hands-on experience. Start with Claude or ChatGPT before venturing into more specialized tools.

Here’s your executive starter pack:

• Strategic Planning: Today’s AI models are extremely good at taking in large volumes of text and coming up valuable insights. Feed those massive reports into AI tools and watch them extract insights faster than a gossip columnist at a celebrity wedding.
• Communication Enhancement: Writing emails, drafting announcements, sending investor updates, these are all things that can be done faster and better with AI. You still provide the vision, but AI makes sure it doesn’t read like it was written at 2 AM after your fourth espresso.
• Meeting Follow-up: Tools that automatically generate meeting notes and action items? Yes, please! It’s like having a super-efficient assistant who never needs coffee breaks.
• Competitive Intelligence: New Deep Research capabilities (Google it… actually, no, ChatGPT it) let you gather information across hundreds of websites in minutes. Your intern can go back to getting you coffee now.

In a podcast with Patrick O’Shaughnessy, Ravi mentions how he fed ChatGPT some context about a dinner he was going to and asked it to give him some talking points. He read the talking points on the Uber ride over impressed his guests. This behavior should become second nature to you.

Step 2: Mandate AI Use Across Your Company

Once leadership is on board the AI train, it’s time to get everyone else tickets. Some employees are already secretly using these tools, but with leadership’s blessing, adoption spreads faster than rumours about office romance.

A fun and quick way to do this is to have leaders share their learnings in team meetings. You could perhaps call an All-Hands and have every leader walk through something they tried with AI and the results.

Another way is to have Department Heads conduct workshops for their departments to identify and experiment with AI tools. I’ve mentioned ChatGPT and Claude but there are hundreds of department specific AI tools that are more suited for certain tasks.

You also need to ensure employees have permission to try out different tools. At the very least, give everyone in your company a Team subscription to ChatGPT.

When I was running the Venture Studio at Forum VC, this is exactly what we did. The whole company got a subscription to ChatGPT, and we even mandated usage of it for my department.

Step 3: Conduct a Comprehensive AI Readiness Assessment

With your team engaged, the journey continues with a thorough understanding of your current state. I should warn you, this requires deep examination of how your company actually operates and you may find some surprises while doing this.

I’ll break down exactly how to conduct this assessment in actionable steps.

Map Your Workflows Systematically

First, create a standardized workflow documentation template that captures:

• Process name and department
• Inputs required (data sources, triggers, resources)
• Steps performed (sequential actions with decision points)
• Outputs produced (deliverables, decisions, impacts)
• Current time investment (hours per week/month)
• Estimated cost (labor, technology, opportunity cost)
• Business impact rating (critical, high, medium, low)

Next, organize department-by-department workflow collection sessions. I recommend instructing each department head to schedule dedicated 2-3 hour mapping workshops that include frontline staff.

These sessions should uncover both formal procedures and those informal “shadow processes” that exist in every organization but rarely appear in official documentation.

Direct teams to focus particularly on processes that:

• Are performed frequently (daily or weekly)
• Follow clear patterns or rules
• Involve significant data processing or analysis
• Consume substantial employee time
• Create bottlenecks in delivering customer value

Do not include processes that are part of your core value as a business (just yet). Right now, we’re focussing only on processes that do not deliver core value and hence have low risk of automating them.

Score Each Process

Once that’s done, we score each process in terms of how “AI ready” it is. It doesn’t have to be complex. I usually create a 1-5 scale assessment across these dimensions:

• Data structure (1 = unstructured/analog information, 5 = highly structured digital data)
• Decision complexity (1 = requires nuanced human judgment, 5 = follows clear rules)
• Pattern recognition (1 = each situation is unique, 5 = clear patterns exist)
• Creativity required (1 = high creative input needed, 5 = minimal creativity needed)
• Emotional intelligence (1 = high empathy required, 5 = minimal EQ needed)

After scoring, create a quadrant analysis plotting each process on two axes:

• X-axis: AI Readiness (combined score of above dimensions)
• Y-axis: Potential business impact (importance, cost, time savings)

This visualization makes your prioritization decisions much clearer. Based on the scoring results, categorize processes into implementation timelines:

• Immediate candidates (Q1): High scores on both axes – your quick wins
• Mid-term candidates (Q2-Q3): High on AI readiness, medium on business impact
• Long-term vision (Year 2+): Medium scores or processes requiring significant redesign
• Human-centric processes: Low scores that should remain primarily human-driven

A sales agency I worked with had a very clear, structured, onboarding process that didn’t require much creativity. All they needed to do was gather requirements and turn that into a document for the delivery team.

Unfortunately, it took two weeks on average to complete onboarding, with multiple calls and emails between the onboarding team and the client. It’s not a core process but had high business impact and scored well on AI readiness. A prime candidate for automation.

Identify and Empower Your AI Champions

The final component of your assessment identifies the people who will drive transformation from within the organization.

Deploy a company-wide AI attitude survey with questions that reveal:

• Current use of AI tools (both personal and professional)
• Interest level in AI applications within their work
• Concerns or reservations about AI implementation
• Ideas for how AI could improve their specific functions
• Desire to participate in AI initiatives

A healthcare system I worked with was surprised to discover that some of their most enthusiastic AI advocates weren’t in IT or analytics, but in clinical departments where individuals had independently started exploring AI tools to solve daily challenges.

Analyze the assessment and survey results to identify potential champions, then conduct one-on-one interviews with promising candidates. Look for individuals who demonstrate:

• Practical AI knowledge or strong aptitude to learn
• Respect among peers (influence without authority)
• Ability to bridge technical and business perspectives
• Track record of successful change management
• Persistence through challenges

When your assessment is complete, you’ll have three critical assets:

1. A comprehensive map of your organization’s processes with clear AI potential scores
2. A prioritized transformation roadmap with timelines
3. A group of internal AI champions ready to drive change

This assessment provides the foundation for all your subsequent transformation efforts. It ensures you’re targeting the right opportunities, with the right sequence, and with the right people involved.

Step 4: Launch Strategic Pilot Projects

With your assessment complete, it’s time to move from theory to practice by launching some pilot projects.

Pick out 2-3 projects from the processes in the first quadrant in Step 2. If you have many contenders, prioritize projects with high visibility across your organization, or span multiple business functions. They should also have clear ROI potential.

I’m not going to go into how to run and manage projects here but it is extremely important. While everything I’ve mentioned so far sounds like a lot of fun, execution is usually where most companies stumble. This is really the make-or-break step, and to set you up for success, here are a few pointers:

Treat It As a Real Project

This is not a side project. Most side projects fail or don’t result in anything long-term because they aren’t taken seriously.

Have your AI Champions from Step 2 lead these projects, make it their primary KPI, and give them the team and resources they need.

Set Aggressive Implementation Timelines

To create momentum and prevent analysis paralysis, establish ambitious but achievable timelines for your initial projects.

I’ve helped companies launch and deliver AI automations within 30 days. Remember, we’re not looking for perfection here. We’re piloting a new way of doing things, and it just needs to be better than the old way.

Document Process Changes and Results Meticulously

Successful pilots will make the case for further transformation. Establish clear baseline metrics for the processes you want to automate, and then measure the results.

Document everything meticulously. These case studies become powerful tools for expanding your transformation. PS – you can use AI for this!

Create a Consistent Communication Cadence

Effective communication is often the difference between successful transformations and failed initiatives. Develop a systematic approach to sharing progress, learnings, and successes throughout your organization.

Buy Vs Build

At least for the first few pilot projects, it makes sense to buy existing software or AI tools instead of build it out. You can roll your own AI once you’ve seen value.

One VC client wanted to automate their entire top of funnel deal flow. We could have developed an end-to-end AI automation but we decided instead to cobble together 3 different software. It’s not perfect but it improved investment throughput and we laid the foundation for more custom AI builds.

Launch the Projects

Don’t just build the pilot and leave it at that. Actually launch it and roll it out. See how it runs in the real world. Measure if it’s making a difference.

Getting a few successful pilots off the ground and communicating those successes sets you up for deeper transformation down the line.

If you’ve come this far, congratulations, you’re in the second level of the Pyramid of AI Adoption – Automation.

Step 5: Redesign Your Core Business Processes

Once your initial pilots demonstrate value, it’s time for deeper transformation. We’re getting to the third level on the pyramid.

This is where the “re-founding” concept becomes most apparent. You’re not just improving existing processes, you’re reimagining how work gets done.

Begin by identifying processes that form the backbone of your value creation. These are the processes I told you not to focus on in the previous step.

For a software company, this might be your development workflow; for a financial institution, your risk assessment process; for a healthcare provider, your patient care pathways.

Before redesigning this process, thoroughly document the current process to understand its complete flow, inefficiencies, and hidden dependencies. This mapping creates a baseline understanding that will inform your redesign.

For each selected process:

1. Conduct detailed observation sessions with the people who perform the work daily
2. Document every step, including unofficial workarounds and exceptions
3. Identify decision points and the information used to make those decisions
4. Measure time, cost, and quality metrics at each stage
5. Identify pain points, bottlenecks, and redundancies
6. Map data flows and information handoffs between systems and people
7. Document compliance and regulatory requirements

With this baseline, you can pick it apart and redesign it. The key to true transformation is starting with a clean slate rather than incrementally improving existing processes.

Conduct structured workshops where teams reimagine the process from first principles, considering AI capabilities as fundamental building blocks:

1. Begin with the core purpose of the process and desired outcomes
2. Challenge all assumptions about how work must be done
3. Ask: “If we were building this process from scratch today, with all of AI’s capabilities available, how would we design it?”
4. Identify which decisions could be automated, augmented, or should remain human-driven
5. Examine how to eliminate handoffs and information re-entry
6. Determine how humans and AI will collaborate within the redesigned process

Once you’ve redesigned the process, you can start the implementation. Again, I won’t go into project management here but keep in mind the advice I gave previously.

Since this is a redesign of your core process, you also want to start small. Pick one piece of the design to implement first, measure it, learn from it, and then move to the next piece. Like trying a new hairstyle, you don’t go from conservative cut to mohawk overnight.

Remember that process redesign is fundamentally about rethinking how work creates value, not just making existing processes more efficient. The organizations that achieve the greatest transformation benefits are those willing to challenge fundamental assumptions about how work must be done.

Step 6: Transform Your Product and Service Offerings

With internal transformation underway, turn your attention to market-facing opportunities. This is where AI fundamentally changes your value proposition in the marketplace. Rather than simply improving existing offerings, this step reimagines what’s possible when AI becomes central to your products and services.

Establish an AI Innovation Team

Create a dedicated team for AI-driven product innovation. Many companies are now hiring Chief AI Officers, and AI Product Managers to research and build AI-first products and features.

Create Rapid Prototyping Processes

With tools like Cursor, Windsurf, and Lovable, it’s extremely easy to rapidly prototype new products (especially in software). This doesn’t mean they’ll be instantly integrated into your core products and services but you can launch them as side tools and measure the response.

Build Customer Feedback Loops

Collect usage metrics not just on the product but also the AI features. A software company I advised built systems tracking not only when customers used their AI writing assistant but which suggestions were accepted, modified, or rejected, creating a rich dataset for improvement.

Update Your Pricing

As you develop these offerings, rethink your pricing strategy. AI-enhanced products often create exponentially more value than traditional alternatives.

AI Or Die

The pace of AI advancement isn’t slowing down.

The companies that thrive will be those that start transforming today. They’ll make mistakes and face challenges along the way, but they’ll develop the organizational capabilities needed to capitalize on each new AI breakthrough.

Is it easy? About as easy as teaching a cat to swim. It requires courage, commitment, and fundamentally rethinking how your business operates. But the alternative, watching AI-native competitors eat your lunch while you still decide what to order, is far more painful.

I’ve guided numerous organizations through this journey, and while each transformation is unique, the framework outlined here provides a proven path forward.

If you’re ready to begin your company’s AI transformation but need expert guidance, I’d welcome a conversation about how I can help you navigate this complex but essential transition.