Give Sopact Sense a program's page and it rebuilds the logic model, grades every column by evidence, and flags the orphan activities and missing outcomes a reviewer circles first.
In short: To audit a logic model, give the program's page (or its text) to an AI and ask it to rebuild the columns — inputs → activities → outputs → outcomes → impact — and grade each by evidence: green where data backs it, amber where an activity is orphaned or a claim is untested, red where a column is missing. Sopact Sense does this in minutes straight from a URL, flagging anything it can't verify, so you see exactly what a reviewer would circle first.
An audit starts from what the program actually publishes — not what you assume it does. Point the Assistant at the program or proposal page and tell it to use only what's stated there:
You are the Sopact Sense Assistant. Audit the program/proposal page at [URL] (or I will paste its content). Use only what is stated on the page. Wait for my task.
Ask Sense to rebuild the logic model from the page and grade every column — naming the assumption under each link and flagging any activity that feeds no output:
Read [URL] and reconstruct its Logic Model (inputs → … → impact). Note the assumption under each link and flag any activity with no output. Do not infer what the page doesn't say. Grade every column green / amber / red.
Five elements make the audit rigorous: the input (the live [URL] or its text); reconstruct the columns (inputs to impact); orphan activities (an activity that leads nowhere on the model); no inference (it won't invent what the page doesn't say); and the grade (green / amber / red at a glance).
Sense returns the reconstructed columns with each element graded and the orphan activities and gaps called out. The demo audits Bright Futures Initiative, engineered to grade one green, one amber, one red:
Bright Futures Initiative helps ~150 first-generation students each year. We run weekly tutoring, college-application workshops, and a summer bridge camp. Our goal is for students to enroll in college and persist to year two. Last year 128 of 150 students (85%) submitted at least one college application, tracked through our application portal. We believe consistent mentoring builds the confidence students need to persist. We have not yet measured year-two persistence, and we do not state a baseline college-enrollment rate.
GRADE: green | 85% applied | 128 of 150, tracked in the portal; amber | summer bridge camp | orphan activity — feeds no output; red | year-2 persistence | no data, no baseline
The green column is carried by tracked data, the amber flag is an orphan activity — the bridge camp that doesn't connect to a tracked output — and the red box is the missing long-term outcome: no year-two persistence and no enrollment baseline.
The payoff of an audit is the fix. Take the lowest-graded element and make it measurable with one realistic change:
Take the lowest-graded element above and fix it using only what the program could realistically measure. Show the before → after grade and the single indicator/edit that moves it to green.
For Bright Futures, that's a year-two enrollment check against a stated baseline — turning the red persistence gap into something the program can actually show.
Generate a decision-first report in your own brand, then a shareable link:
Create a 'missing & incomplete' report from this analysis in Sopact branding [or paste your website URL / brand guideline to apply your own]. List every element graded amber or red, what is missing, and the one input that fixes each. Lead with the decision this report informs.
Create a shareable link for this report and open it in a new tab.
Orphan activities are the giveaway. A reviewer's eye goes straight to an activity that produces no output. Ask Sense to list orphan activities explicitly — it's the fastest read on whether a logic model is honest.
Audit from the live page. Pointing Sense at the program's URL keeps the audit grounded in what the program actually publishes — and surfaces the claims a funder would read first.
Re-audit after each fix. Add the one indicator Sense suggests, then re-run the audit next cycle and watch the grade climb.
Tighten the program page for accuracy. Once you've seen the grades, ask Sense to rewrite the page so its claims match the evidence:
Based on the grades above, suggest edits to the program page so its claims match the evidence. Flag every sentence that overstates what the program can show, and rewrite it to be accurate and specific.
A good logic model has no orphan activities (everything you do feeds an output), an evidenced link at each stage from inputs to outcomes, and a baseline for the outcomes it claims. Weak ones have activities that lead nowhere, outcomes with no indicator, or long-term impact with no follow-up. Grading each column green, amber or red makes those weak spots obvious at a glance.
A logic-model audit reconstructs a program's columns — inputs → activities → outputs → outcomes → impact — and grades each link by how much evidence supports it, flagging orphan activities and missing baselines. It tells you whether the logic is credible and exactly where it isn't, before a funder does.
Yes — give Sopact Sense the program's page URL and it rebuilds the logic model from what's published, grades every column, and flags anything it can't verify, without inventing outputs the page doesn't state. It takes minutes instead of a manual review.
Open Sopact Sense, paste your program description, and put it to work.
Try in Sopact
