case files
these are things i have actually built, researched or contributed to.
no hypotheticals.
AutoBizBot is a telegram-based business assistant built for small nigerian vendors, particularly in the perfume and lifestyle space. it handles order logging, customer follow-ups, invoice generation and payment collection through paystack — all inside telegram.
the problem it solves: most small vendors in nigeria run their entire business in whatsapp voice notes and scattered dms. AutoBizBot gives them a lightweight back-office without asking them to change how they communicate.
note: full automation services are handled separately at abdullahishaq.tech
see the automation side of my work → abdullahishaq.techrepresentation of the problem and the fix.
before state
[vendor forgets. no record. no receipt. no follow-up.]
after state
pay via paystack: [link]
your order is logged. you'll receive a confirmation once payment clears.
[automated receipt generated. order logged to google sheets. follow-up message scheduled for 48 hours.]
reviewed the data practices of a small nigerian e-commerce business against the requirements of the Nigeria Data Protection Regulation and the 2023 Nigeria Data Protection Act. identified seven gaps including absence of a lawful basis documentation framework, no privacy notice on the checkout flow and an unregistered data controller status with the NDPC.
produced a gap analysis report and a remediation roadmap. all seven gaps were addressed within sixty days.
brothers' keeper is an intensive overnight personal development programme for young men at obafemi awolowo university.
i chair the programme. i design its structure, facilitate its sessions and oversee its operations. the skills involved are the same ones that make a good lawyer: listening carefully, asking hard questions, holding a framework while people are uncomfortable.
i have been studying how decentralised autonomous organisations make decisions, resolve disputes and distribute authority without traditional legal structures. the governance questions here are serious legal questions: who is liable? what counts as a binding decision? how does a DAO respond to a data subject access request?
this is background work feeding into my broader research on ai and emerging tech governance.
a systematic review of how ai tools perform on nigerian law and how their limitations create legal risk for practitioners who rely on them uncritically.
key finding: most general-purpose AI tools have poor coverage of nigerian case law and subordinate legislation. a practitioner using them without verification is accepting a meaningful risk of error. the solution is not to stop using them. it is to understand exactly where they fail.
→ read the full article