It's that time of year again. Your firm wants a comprehensive self-assessment covering the past twelve months of your work. They'd like specific examples, quantifiable accomplishments, and thoughtful reflection on your professional development.
And you're sitting there trying to remember what you did in February.
I always find the annual self-assessment to be frustratingly challenging: you're asked to recall a year's worth of work product across dozens of matters while billing 2,000+ hours. You have all the information you need, but it's scattered in your time entries, your emails, your documents, your filings, and your mind. And best of all, you're likely not receiving credit towards your billable hours threshold for time spent running this down.
This is the exact kind of problem I like to solve with the help of AI; specifically, the tedious groundwork of gathering, organizing, quantifying, and drafting. Not the substantive thinking about what to highlight or how to frame your career trajectory (that's still on you). If you leverage AI properly, it will save you several hours.
Here's the workflow I've been using.
Pinpointing the Pain: Why Self-Assessments Are Uniquely Miserable
Before diving into the solution, it's important to understand the pain points associated with the problem. For self-assessments, it's not just that they are time-consuming; they're cognitively brutal in specific ways.
The recall problem. Depending on your practice area, you might have billed to upwards of forty or fifty matters or deals this year. And each matter consists of several, often countless, different tasks. You need to remember not just what you did for each matter, but which tasks were significant, and ideally include some concrete details that make your contributions tangible. That's a lot to reconstruct from memory.
The highlights problem. There's a difference between "stuff you did" and "accomplishments worth mentioning." You billed a lot of hours, but which of those hours represented your best work? Which tasks did the partners actually value?
The drafting problem. Even once you know what to write about, you still have to write it. And finding a few hours of uninterrupted time to draft when you're juggling active matters feels impossible.
Recency bias. You vividly remember the work you did last month but might have completely forgotten the important work you handled in February, even though that might have been more impressive work.
The blank page problem. Staring at an empty Word document is demoralizing. Most people find it far easier to react to a draft than to create one from scratch.
The narrative problem. When you're buried in day-to-day work, it's hard to see patterns in your own development. What skills did you build this year? What's the throughline? Stepping back to see the forest is difficult when you've spent twelve months focused on individual trees.
Invisible work. Some of your most valuable contributions (quick-turn research, coverage assignments, mentoring junior associates) didn't generate big hours but genuinely mattered. These are easy to forget and undersell.
AI can help with all of this.
My AI-Assisted Workflow
I break this into three phases: data gathering, context layering, and drafting. Each phase builds on the last, and the whole process turns what used to be a painful half-day exercise into something much more manageable.
Phase 1: Data Gathering (Let AI Do the Heavy Lifting)
Start with your time entries. Most firms use Intapp or similar billing software, and you can typically export your entries for the review period as a spreadsheet or PDF.
Upload that export to your AI tool of choice and ask it to analyze your year. Specifically, have it identify which clients and matters you logged the most hours on, what categories of work you performed, and a rough timeline of your year showing what you were working on and when.
You can try this prompt:
<context>
I am an associate at a law firm preparing my annual self-assessment. I have uploaded an export of my time entries from our billing system covering the review period. Each entry typically includes the date, client/matter name or number, hours billed, and a description of the work performed.
</context>
<task>
Analyze my time entries and produce a comprehensive summary of my year that I can use as the foundation for my self-assessment. I need to understand where my time went, what types of work I performed, and how my year unfolded chronologically.
</task>
<instructions>
Please analyze the uploaded time entries and provide the following:
1. Top Matters by Hours: Identify the clients and matters where I logged the most time. For each, provide the total hours and a brief characterization of the type of work based on the time entry descriptions.
2. Work Categories: Group my time entries into categories of legal work (e.g., document review, legal research, brief writing, deposition preparation, discovery, client communications, court appearances, etc.). Show the approximate percentage of my time spent in each category.
3. Quarterly Timeline: Break my year into quarters and summarize what I was primarily working on during each period. Flag any matters that were concentrated in the earlier part of the year, as these are easy to forget when writing a self-assessment.
4. Potential Highlights: Based on the time entry descriptions, identify entries or clusters of entries that suggest significant work product (e.g., drafting motions, preparing for hearings, taking or defending depositions, leading document reviews). These are candidates for accomplishments I should consider highlighting.
5. Patterns and Observations: Note any patterns you observe, such as increasing responsibility over time, variety of work across practice areas, or sustained involvement in major matters.
</instructions>
<output_format>
Please organize your analysis with clear headings for each section. Use tables where helpful for the quantitative data (top matters, work categories). For the timeline and highlights sections, use narrative descriptions that will help jog my memory about specific matters and tasks.
</output_format>
<note>
Time entry descriptions are often written in shorthand for billing purposes, so interpret them generously. If a description is ambiguous, make reasonable inferences about what the work likely involved and flag any assumptions.
</note>
Please let me know if you have any clarifying questions before beginning. What this gives you is a high-level map. Instead of trying to reconstruct twelve months from memory, you now have a structured summary that shows you where your time actually went. This alone defeats recency bias. Now you can proceed to analyzing and contextualizing.
Phase 2: Context Layering (Adding the "What Actually Happened")
Time entries don't tell you enough details, including, importantly, the importance of the work product to your team (i.e. was the team in a tough spot working on a tight deadline vs. having weeks to complete the assignment), the supervising attorney's feedback regarding your work, or what the client or court said about the work product.
For that, you need context. And the best source of context is probably your email account.
I take the summary of my billing entries from Phase 1 and feed it into an AI platform that has access to your email. I use ChatGPT Enterprise synced to Outlook. Ask the AI to cross-reference your billing data with your email threads for those matters.
Have it identify the specific deliverables you produced, any feedback you received from partners or supervising attorneys, and how matters progressed or resolved.
What this gives you is the connective tissue. You're no longer just listing matters. You're identifying actual accomplishments with concrete details. This is the difference between a forgettable self-assessment and one that effectively reminds the partnership committee of your key contributions.
Phase 3: Drafting (From Raw Material to Polished Accomplishments)
Now you have materials to work with. Review the AI's output and let it refresh your memory. You'll probably remember things the AI missed, so add those in. You'll also notice things the AI included that aren't actually worth highlighting, so cut those.
Once you've refined your notes, upload them back to the AI along with your firm's specific format or any requirements for how accomplishments should be framed. Have the AI draft accomplishment statements that you can then edit and personalize.
What this gives you is a working draft to react to. You're no longer staring at a blank page. You're editing, refining, and adding your voice to something that already exists. That's a fundamentally easier cognitive task.
If You Don't Have Email-Synced AI
Not every firm has an enterprise AI platform with email integration. If that's your situation, here are some workarounds.
Manual email export. Search your sent folder for the key matter names from your billing summary. Export those threads as PDFs and upload them to your AI tool. It's more manual, but it gets you the same context.
Calendar mining. Export your calendar for the review period and have AI identify significant meetings, deadlines, and events. These can trigger recall about what you were working on and when.
Document management pull. If you have access to your recent documents in iManage or NetDocs, export a list of file names or pull key documents. The document titles alone can remind you what you produced.
The brain dump method. Use your billing summary as a memory trigger, then just start talking. Record a voice memo of everything you remember about the year. Stream of consciousness is fine. Transcribe it and upload to AI for organization. Sometimes the fastest way to get information out of your head is to speak it rather than type it.
What AI Can't Do
A quick caveat: AI handles the gathering and drafting, but it can't tell you what actually mattered.
Only you know which assignments stretched you, which partner relationships you're trying to cultivate, or which skills you're emphasizing for your next review cycle. The AI doesn't know your career goals or your firm's internal structure. It can surface the raw material, but you have to decide what story to tell with it. Don't be lazy.
Think of this as editing and curating, not outsourcing. You're still the one making the substantive judgments. You're just not wasting hours on the mechanical work of gathering data and generating first drafts.
What used to take four to six hours of painful recall and drafting can become one to two hours of reviewing and refining. That's not nothing when you're trying to bill and still have a life in December (or whenever is the end of your firm's fiscal year).
More broadly, this is what AI is definitely good for in legal practice. It's the low hanging fruit. Self-assessments are a small example, but the pattern applies across a lot of administrative work that eats into your day.
Your time is valuable. Spend it on the parts only you can do (and bill for).