Building a Target BRE List With AI, Start to Finish
Business retention and expansion work lives and dies on the quality of your call list. Visit the right employers in the right order and you catch expansions early and head off closures before they are announced. Visit alphabetically and you are doing motion instead of work.
AI has changed how I build these lists, not by replacing judgment but by collapsing the hours of assembly between having a territory and having a plan. Here is the workflow I used on a recent parish-level project.
Define the universe first. Before I opened any tool, I set the boundaries: geography, the sectors that matter to this economy, size thresholds worth a visit, and what “priority” means here. AI is excellent at filling a well-defined container and useless at deciding what the container should be. That decision is yours, and it is the whole game.
Assemble the raw list. With scope set, I used AI to pull together a first-pass roster of employers, with locations, sector, and estimated employment. This used to eat a day of tab-switching. The model gets you to a working draft in minutes. It also gets things wrong in minutes. Employment figures drift out of date, relocated companies show up at old addresses, subsidiaries get listed as independents. The raw list is a hypothesis, not a fact.
Verify, verify, verify. This is where the real work is and where AI helps least. For the targets that rose to the top, I confirmed the details against primary sources: company sites, recent local news, state filings, and a human who would know. Generative AI will state a plant’s headcount or an ownership change with total fluency and zero factual basis. Paste an unverified claim into a prospect memo and it is your credibility spent, not the model’s. The verification pass is not overhead on the workflow. It is the workflow.
Prioritize with criteria you can defend. Once verified, I ranked the list into primary targets and a secondary pipeline. AI helps you score and sort, but the ranking has to survive plain English to the person who signs your contract. “The model put it there” is not a reason.
Deliver something usable. I built the final product as a clean, sortable workbook, with the fields a caller needs at a glance and verification notes attached.
Add it up and the division of labor is clear. AI compressed the assembly. It did not decide what mattered, confirm what was true, or stand behind the result. Used as an accelerator bolted to your judgment, it hands you back the hours for where retention work is really done – in the room, with the employer, listening.