AI and the new era of startup documentation (PT 1): Information as an efficiency differentiator
When chatbots rule, your startup's documentation strategy becomes non-negotiable. Are you ready?
Bouncing across my Notion dashboard this week came a feature I had been patiently awaiting, but one that admittedly I didn’t expect to show up for at least some months.
Being able to query your knowledge and project management system is the dream of any always-remember-to-check-in-on-something founder. Instead of Slacking our teams for answers to the miscellaneous questions that tornado through our minds before, in between, and during each of the many meetings of the day, the promise of asking the robots to soothe our wandering thoughts sounds downright magical, and the effect on team culture, boundary setting, and a more relaxed digital working environment sounds even better.
As a founder of a startup built leveraging the minds of expert generalists to answer said questions and bring order to said chaos, the notion (pun intended) of Notion’s beta bot led me to an existential founder crisis.
“So, is this the end of system building for startups, and will the robots become our generalist glue, bringing all the answers to the team, and owning the space on how We Do?”
Not quite. I thought about how often in the thick of it in startups we deprioritize the unimportant work of pulling things out from our heads (here’s looking at your status updates, general process documentation, and thoughtfully written reflections), and consider them less important than our revenue-driving activity counterparts.
Stripe is famous for its documentation culture and is the cult fave of process nerds who like to talk shop, wishing their startup teams could just follow the dang rules they carefully wrote down in the first place. Why? Because their penchant for the written word allowed everyone to have a repository of rich insights into what was going on in the minds and meetings of their colleagues. And what Notion promises to a team like that is truly magical.
For most teams? Not so much. But we can change that. If the bots can read our internal data and provide updates and information, that information better be good if we want to make the best decisions. This means that what was formerly unimportant actually gives us the unbelievable ability to create efficiency gains we’d never be able to achieve before.
how DO you leverage the power of upcoming AI search tools to maximize your team’s chance of being more efficient with their time?
Design and implement a company-wide internal documentation strategy.
Why This and Why Now?
This is happening. Right. Now. Putting it off until things slow down “after the holidays”, or “once I raise”, or “once we’ve got that process buttoned up” will leave you scrambling last minute to throw some quick notes in a Notion page, wondering why the bot’s told you nothing.
Complexity compounds, and with every new face comes a new question, a new “Why did we do this like this before?” and “What’s this team working on and how does it relate to what I’m doing?”. Shorten your onboarding times. Reduce errors. Build a team that feels autonomy and agency. It all comes down to how you document the stuff that’s happening in a place where the data is searchable and usable.
how We Do: Designing and Implementing a Documentation Strategy for AI
Designing your Strategy
To effectively design your strategy, ask yourself and your team these questions to better understand what system will connect your existing processes and people into a knowledge-sharing machine.
Tools 🛠️
For this how We Do we’ll be focused on Notion given the prevalence of the tool in startupland, and the upcoming Q&A bot launch.
Rules (Process) 📝
Knowledge Management
What are the commonly asked questions from each function? Can they be answered with existing documentation?
What function on the team is most active in documenting their work and processes? What makes them successful in the current culture?
How much of what is written is outdated information? What can we reference with dates or categories to affirm that they represent a historical source of truth?
Where does knowledge live outside of the tool that would need to be queried? Are there integrations you can leverage to make Notion the place for searchable information? Can you leverage AI tools that take meeting notes and Slack debriefs and direct them where they need to go?
Project Management
Before a project starts, what information is being said in meetings or outside of Notion that needs to be captured to give the full scope of the project?
Do you capture roles and responsibilities before a project begins? Who captures them? Why?
What categories are being tracked that you ask most often about? Deadlines? Owners? Progress? Which of these are currently being captured and documented on the team?
When a project ends, how do you know it’s ended? How do you capture the result, outcome, and path forward?
People 🫶
Understanding your team:
Who on your team is a naturally organized documentarian? Can you leverage their work and natural proclivity towards documentation on the team?
Who is the ultimate team communicator? Whose prose wows you and gets the point across with enough information to be usable without going overboard?
Who sets the tone (literal voice and tone) for the team? Can you leverage their writing as an example of how to document in a branded way?
Who leads the team in upholding your values? Can you give them a role to gut-check team knowledge to make sure how things are getting done and what people know is consistent with team values?
Who struggles here? Who will need extra help to be successful?
Understanding how they spend their time:
Where are the natural points of work where documentation is referenced by the team? Can you use this as a point to gut-check what’s written for errors and omissions?
Where are the natural points of work where documentation is created by the team? Can you design a work habit that aligns with this so that creation is second nature and built into existing workflows?
Actually Actionable
Nice article. Now what?
Implementing Your Strategy
Team Buy-in:
Kickoff Meeting (60 min)
Discuss the purpose of the new documentation strategy, and allow for questions (you may have lots of them!) (15 min)
Present plan for implementation (45 min)
Gap-Filling:
Function by Function (2 weeks)
Timebox the buildout to ensure progress across the team. Some team members (the generalists) may be covering more than one function
Implement and test outside tool integrations to ensure the right information is plugging into your source of truth
Ongoing Grooming:
Establish a cadence of grooming aligned with the org
Keep a standing meeting where all those responsible for documentation get time to do so – or;
Set a cadence for when said grooming should occur. Only before new hires are onboarded? After projects are completed? Right before monthly planning syncs?
Q&A Gut-checks:
Ensuring that your source of truth is, indeed, true (Monthly, 30-minute team meeting)
Take a team pulse on how the tool is being used.
Does the information that’s coming back in the bot seem correct? Where have there been snafus? Should you design ways of mitigating against future issues? Can you pinpoint the root cause?
Before you go
Efficiency will be gained when the quality of information that AI is using is excellent, and data quality only comes from intentionality and commitment to the practice. The promise of self-serve knowledge management querying is one where efficiency gains in team performance and workload management will outweigh the time spent bringing in and updating key information. And by the time you implement this system, the bots will probably have a way to fix that problem, too.
The oAT team will be taking a break from how We Do next week to celebrate Thanksgiving, but will be back in your inbox on November 30th. Wishing you a wonderful holiday!
Writer: Britt