If 2023 was the year of “personal AI” following the launch of ChatGPT, 2026 is the year of “workplace AI.” In 2026, AI models reached the point where they can handle a lot of workplace tasks, beyond software development. As a result, companies and organizations across the country have started to incorporate AI into their workflows. Per the Ramp AI Index, which tracks spending data of 70,000 firms, 56% of businesses pay for AI tools, as of September 2026. This may be modestly overstating things since Ramp’s customers probably skew toward the types of companies that are early AI adopters. On the other hand, it also doesn’t account for employees who have their own personal AI subscriptions that they use for work.
While it feels like AI has been around for a long time, it is still early days with this technology. The extent to which professionals use AI in their workflows, and what they use it for, seems mixed. We are probably still in the “early adopter” phase, or at most, the “early majority” phase for workplace AI. Thus, we thought it would be useful to provide some examples of how you can incorporate AI into your workflow. We have collectively found it to be very helpful in terms of generating better output and getting there more efficiently.
It’s also important to acknowledge the shortcomings of AI and highlight ways in which it can detract from the process. There are still some tasks better done by humans, and those are the areas to double down on as a professional. Overall, we think the ideal solution for most is “hybrid intelligence,” or combining human intelligence with artificial intelligence to create the optimal output.
Using AI at Work
So how do we use AI at work? Below are some examples, ordered from most basic to most complex.
Thought Partner & Research Assistant – AI works as a partner to check your ideas and push back on them. For example, you might have an idea about a real estate investment opportunity that you are considering exploring further. AI can help you think through it and make sure you are covering it from all angles. This will help you save time in the long run, since it may lead you to shift your approach or even abandon the idea altogether. You can set your AI to “low agreeableness,” so that it tends to push back on your ideas as opposed to immediately agreeing with them. Another example is researching a new market. AI can be particularly effective in this scenario when you connect it to paid tools (e.g. Costar, Green Street) that it would not otherwise have access to.
Reviewing/Underwriting Opportunities – AI can become your default way to look at new opportunities. If you get a broker package, assuming confidentially terms allow it, you can upload it up and get AI’s initial thoughts on the deal. It can quickly identify the opportunities and risks. Better yet, you can train AI on your specific investment style, so that it knows the types of opportunities that you look for. Or, you can set up a risk/return rubric to evaluate deals on a systematic risk-adjusted return basis. Finally, you can have AI develop an underwriting model, and even set up the formatting to your liking. AI is incredible at analysis, so it is very valuable to put it to work in these areas. Then you can shift your focus from being a “doer” to a “reviewer” and add more value that way.
Investment Diligence – If you have a typical investment diligence checklist, you can upload it to AI and have AI help you work through the list. While AI will not be able to handle all steps of the process, it can help you stay organized and review diligence documents more efficiently. It can help you generate lists of questions for the counterparty, review responses, and determine next steps. It can also be a thought partner when you need to make a call on whether to sign off on diligence or pass on the deal.
Investment Memos – Once the diligence process is complete, AI has very good context from which to draft an investment memo, which is the next step. You can upload a prior investment memo so that it has knowledge of your typical writing style, organization, and formatting. If AI built the model, it can incorporate excerpts or summary pages of the model into the memo. Once AI produces a draft memo, you can either make changes to it yourself, or work with AI to revise certain sections with some guidance. Then you can have AI review the final draft to suggest any final edits.
Checking Models – AI is useful for specific edits inside an existing model, especially when you can explain what needs to change and where it lives. AI is also good at helping audit the checks that you have built into your workbooks and investigating why something is not tying out. For example, checks can test whether sources & uses balance, debt balances roll forward, or distributions tie back to the cash available. Those are useful things to review after an edit. But a clean checks sheet only tells us that the tests we built are passing. If you never built a check for a particular issue, that issue can still be there.
Documenting Work – You can use AI to document your work via markdown files, which are plain text files with a .md extension. You can create global markdown files, documenting how you work. This might include standard operating procedures, model conventions, and instructions on how you approach things. You can also create project-specific markdown files, recording your progress on a project. As you work through projects, it’s helpful to have written records of what you worked on, any changes that you made, and what still needs attention. This is particularly relevant when working asynchronously in a team setting.
Meeting Notes and General Knowledge – AI is very useful for recording meetings and turning them into notes. You can review those notes and keep them with the project so that the discussion is available when you return to the work. AI is also very helpful for learning things outside of your experience. AI can get you up to speed on a variety of items, including legal issues, building websites, marketing, and design.
Shortcomings of AI
What are some shortcomings of AI?
99.9% vs. 90% done – In practice, we’ll typically be ready to make a decision when we are 80% or 90% finished with our work. In contrast, it feels like AI seeks to get to 99.9% before making a decision. This can bog down a process and kill deals. Sometimes it is impractical or impossible to get to 99.9%. In other cases, by the time we do all that work the deal is gone.
Confidently Wrong – We have seen numerous instances of AI being confidently wrong. Whether it be in research, analysis, or pressure testing our ideas, AI sometimes makes a statement that is clearly incorrect. Upon questioning, AI will back off and acknowledge the error, and then correct the response. While humans aren’t perfect either, typical professionals don’t confidently present something that they are not sure of. They will double-check first, before providing an answer.
Decision Making/Accountability – While AI is excellent for analysis, it is not as good at decision making. AI can provide a recommendation based on the facts and analysis, but it lacks judgment developed through years of experience and seeing different outcomes play out. Furthermore, AI lacks accountability. If professionals use AI in the decision-making process, the professional is still ultimately accountable for the decision.
Big Picture/Context Gaps – While AI can review all of our work and even observe calls or review transcripts, it typically doesn’t have 100% context. There are often gaps in our professional work that are unknown to AI. Or, AI has deep knowledge/context on one project, but lacks knowledge on another related project. Thus, AI often doesn’t see the big picture that is required to make a decision.
Confidentiality – Even with paid AI accounts with restricted data permissions, it is hard to be sure that documents are 100% confidential when uploaded to AI. While most work documents are not highly sensitive, there are some industries or types of documents that are very confidential and cannot be uploaded to AI.
Relationship Management – This is one area where humans have an edge for the foreseeable future. Some areas of work still need a human touch or a face-to-face conversation. It is hard to imagine a scenario where my AI forms a relationship with your AI, leading to a business deal.
Closing Thoughts
AI is an incredible technology that can supercharge your professional role in a variety of ways. It can help you complete projects more efficiently and also increase the quality of your output. However, it’s also important to acknowledge the limitations of AI. Aim for a hybrid approach: use AI for what it’s best at – including research, analysis, writing, and building. Focus your human efforts on areas where it isn’t as strong – including judgment, decision making, accountability, and relationship management.
Looking to learn more? Check out our various professional resources! Whether you have yet to break into the industry or are just starting out in a new role, we have everything you need. All you need to bring is your effort!
Thanks for reading!
Getting value from our free materials? Consider supporting us on Patreon