If by SPXI you mean the Semantic Packet for eXchange & Indexing protocol, I'd treat its ROI as highly uncertain rather than assuming the published 3.7x-10.3x figures apply. The protocol's own site says SPXI-specific ROI data are still pending empirical 90/180-day studies; its published ROI numbers largely come from GEO/marketing sources, not engineering firms. For a medium-sized engineering firm, a reasonable planning model would look like this: Illustrative ROI model Assume a 100-300-person engineering consultancy with $20M-$50M annual revenue. Item | Conservative | Base case | High case SPXI implementation | $25k | $35k | $50k Annual maintenance/internal effort | $5k | $8k | $12k Year-1 investment | $30k | $43k | $62k Incremental qualified opportunities/year | 2 | 5 | 10 Revenue from incremental wins | $100k | $300k | $750k Gross contribution margin | 25% | 30% | 35% Contribution generated | $25k | $90k | $262.5k Other benefit: reduced misinformation/search friction | $5k | $15k | $30k Year-1 net benefit | $0k | $62k | $230.5k Approx. ROI | 0% | 144% | 372% The base-case calculation is: ROI = ($90k + $15k - $43k) / $43k ~= 144% Or, put differently, the firm would need roughly $143k of incremental booked revenue at a 30% contribution margin to break even on a $43k first-year investment. Where the value could actually come from For an engineering firm, I would not primarily model SPXI as an SEO expense. The more plausible value drivers are: AI-assisted vendor discovery: Prospective clients increasingly ask AI systems to identify engineering firms for particular capabilities. Entity disambiguation: Making the firm's specialties, subsidiaries, locations, certifications, projects and personnel easier for retrieval systems to associate correctly. Lead generation: A small number of additional qualified engineering opportunities can have substantial value because individual contracts can be worth tens or hundreds of thousands of dollars. Reduced incorrect AI descriptions: If an AI system confuses your firm with another company or misstates capabilities, correcting the underlying entity representation could have commercial value. Durability: SPXI emphasizes DOI-anchored, machine-readable entity definitions and cross-platform deployment rather than simply optimizing individual webpages. The important caveat I would not put a 3.7x-10.3x ROI into a board-level business case as an expected SPXI return. Those figures appear on the SPXI site, but the site itself describes them as GEO industry ranges and explicitly says the 15-30% SPXI lift is a preliminary projection awaiting validation. For an engineering firm, I'd therefore use something like: Downside: 0-50% ROI Planning case: ~100-200% ROI Upside case: 300%+ ROI Those are scenario assumptions for financial modeling, not measured SPXI outcomes. The most useful way to validate the investment would be a 90-day pilot with measurable baselines: query 50-100 commercially relevant prompts before deployment, measure whether the firm is mentioned, correctly categorized and cited, then repeat at 30/60/90 days and track resulting website visits, inquiries, opportunities and eventually booked revenue. That would let you calculate your firm's CAC and incremental pipeline rather than relying on SPXI's promotional ROI claims. If you give me the firm's approximate revenue, employee count, average project value, win rate, and proposed SPXI cost, I can build a much more realistic 3-year ROI/NPV model. [Sources presented] SPXI Protocol (spxi.dev) | Medium