An enterprise AI consulting firm pitched a local restaurant owner I know on $500k+ of development work ($100k per distinct project/tool build) for automating/improving various aspects of the business (staff scheduling, tracking beverage turnover, building profiles on customers, and some other things). What stuck out to me was the fact each project was quoted at the same cost split 50/50 between an MVP build, then productionizing it if shows success. I was pretty shocked and offered to take a review of their proposal... My first thought was that there had to be SaaS tools out there already tackling some of these areas for restaurant businesses, and likely with recently released or soon-to-be released AI features and functionality! (In the U.S., the restaurant industry is ~$1T industry!) Sure enough, running the firm's proposal through @claudeai, I generated a report within 30min aligning each distinct "feature" of the tools this firm was proposing to build to existing SaaS solutions in the market for restaurants and their features. While existing tools may not be able to tackle everything, I also know that some of the functions they were looking to automate and enhance could be done with no more than a couple days to a week's worth work of someone who's pretty competent with using agents. Question for you all: How do you bill or have you seen service providers bill for AI consulting/contracting work when building bespoke solutions for a business like a restaurant? (so containing integrations to their reservation system, POS, private events/dining booking, etc + spreadsheets, financials, and documents on procurement or training) Hourly rates? Project-based fixed fees? Outcome-based/fees tied to success of revenue lift / cost savings? Additionally, when you're working with a business that doesn't have an in-house tech/IT team to actually monitor, maintain, or continue to build on a system that's delivered from an outside firm, what does maintenance and updating look like over time?? This proposal I saw had no mention of hourly or monthly retainer rates for after these tools would've been turned over, and also nothing about estimated ongoing token costs for the various functions each tool would handle. In my mind, proceeding with the proposal would've resulted in quite a disaster... and for a firm to charge a hefty upfront fee for a tool delivered, but no incentives to ensure these tools are going to be reasonable to maintain and not burning through excess tokens seemed like a real flaw to this style of engagement. I couldn't help but think many such SMEs are being pitched similarly and am curious how many business owners/operated not that familiar with AI are potentially being led down paths without being fully informed... (For context, this restaurant does well north of $10 - 15M/yr so the proposal was clearly bid to be somewhat in line with knowledge of the relative size of revenue of the business, but I was still pretty shocked at the large figures that seemed much more fitting to Fortune 500.)