Receive – Categorize – Conclude
PMO is all about monitoring and responding to every event in every project in-time, then presenting stats perfectly on portfolio; making smart decision on the basis of preset priorities as an event occurs.
You are limited on multi tasking, well you are human and you have a limited capacity. You can’t address every event manually when number grows. You will definitely be committing errors/delaying, and won’t be able to prioritize tasks and produce correct response after a certain level. What do you need to do?
Digitize maximum events that you encounter. Email (in/out), conversations (text/audio/video), documents, action points, from project in take till the delivery. Never rely on your capacity to remember everything about every project or program even if you’ve got brain of Einstein.
BUILD RAG ASAP !
There are tons of tools out there, if you have an AI guy go build or consult an expert to install local low-cost system; the more you spend on CPU & GPU the more you get out of it.
HOW ON EARTH?
1. Categorize information
2. Extract usable (ETL)
3. Feed information (minimalistic approach)
4. Extract actionable (set alerts in phase 2) – create receptors
5. Present visually
Bingo ! Question it and get tailored answers.
Now fetch custom crafted reports and dump to NoSQL !! And build whatever dashboard you like, PBI, Tableu, PowerBI or whatever.
Extend beyond traditional PMO? Yes you can! But such autonomous system requires hell of tuning. Will talk about it soon. Btw major objectives/components are:
Auto-feed (implement more receptors)
Auto question
Auto respond
Limitations
Level 2
Determine Action & Response categories
Setup robots, text, voice, visual
TOMMIRROW
A new model comes that supports embeddings better. Well no problem! We can transport whole knowledgebase to it, swap, and have it up and running in an hour or so! This also requires yet another writing, will do it soon!