[{"data":1,"prerenderedAt":239},["ShallowReactive",2],{"changelog-2026-11-17-questionnaire-pdf-guests-policy-gaps":3,"changelog-2026-11-17-questionnaire-pdf-guests-policy-gaps-surround":228},{"id":4,"title":5,"api":6,"authors":7,"body":13,"category":132,"date":133,"description":134,"extension":135,"faq":6,"features":136,"fixes":181,"highlight":194,"image":203,"improvements":205,"meta":222,"navigation":223,"path":224,"seo":225,"stem":226,"summary":6,"updated":6,"__hash__":227},"posts\u002F3.blog\u002F2026-11-17-questionnaire-pdf-guests-policy-gaps.md","Questionnaires Back in Their Own PDF, Guests on Assessments, and Policy Gaps",null,[8],{"name":9,"to":10,"avatar":11},"Justin Leapline","https:\u002F\u002Fwww.linkedin.com\u002Fin\u002Fjustinleapline\u002F",{"src":12},"\u002Fimages\u002Fjustinleapline.png",{"type":14,"value":15,"toc":128},"minimark",[16,20,27,33,39,61],[17,18,19],"p",{},"Most compliance work arrives from outside the company. A customer sends a questionnaire. An auditor needs access to one assessment. A framework asks whether a policy exists. This release is about the edges where episki meets those requests.",[17,21,22,26],{},[23,24,25],"strong",{},"Questionnaires go back the way they came."," A questionnaire arrives as a file, often a fillable PDF, and the requester wants that same file returned. Drop it on episki and the questions are read out of it, including any answers already filled in. Answer them from what you have already said, approve them, and episki writes the approved answers into the original form. You send back their PDF, not ours.",[17,28,29,32],{},[23,30,31],{},"Guests join one thing, not the whole workspace."," Customers and auditors are now invited onto a specific program or assessment, and that is all they can see. A contributor can upload evidence and comment on what is assigned to them. Everything else in the workspace stays out of view.",[17,34,35,38],{},[23,36,37],{},"Policies are checked against the framework."," Gap analysis shows which controls your policies cover, which they cover partly, and which they miss, and offers the language that would close each gap. When a control's requirement is the policy itself, the published policy can be its evidence, and that credit expires when the policy changes.",[17,40,41,42,45,46,45,49,52,53,56,57,60],{},"Around those three, ",[23,43,44],{},"goals now measure the program",", ",[23,47,48],{},"vendors are tiered by the data they touch",[23,50,51],{},"loops can be built from a sentence",", and the assistant ",[23,54,55],{},"draws real charts"," and ",[23,58,59],{},"shows what each AI evidence verdict was based on",".",[62,63,64,71,77,86,92,98,104,110,116,122],"ul",{},[65,66,67,70],"li",{},[23,68,69],{},"Questionnaires"," start from dropped files and are returned as the requester's own filled PDF",[65,72,73,76],{},[23,74,75],{},"Guests"," invited directly onto a program or assessment, and limited to it",[65,78,79,82,83],{},[23,80,81],{},"Policy gap analysis"," with suggested language, and ",[23,84,85],{},"published policies as evidence",[65,87,88,91],{},[23,89,90],{},"Goals"," with linked work, key results, and decisions",[65,93,94,97],{},[23,95,96],{},"Vendor data scopes"," that set a minimum tier, and requirements with alternatives",[65,99,100,103],{},[23,101,102],{},"Loops"," from a template gallery or a one-sentence description",[65,105,106,109],{},[23,107,108],{},"Charts in chat",", and overviews for controls, tasks, issues, and evidence",[65,111,112,115],{},[23,113,114],{},"A command palette"," with context, shortcuts, create-anywhere, and Tab to ask",[65,117,118,121],{},[23,119,120],{},"AI evidence review"," that reads screenshots and scans and records what it relied on",[65,123,124,127],{},[23,125,126],{},"Recently deleted",", ISO assessment statuses, implementation guidance, and desktop notifications",{"title":129,"searchDepth":130,"depth":130,"links":131},"",2,[],"changelog","2026-11-17","Drop a security questionnaire on episki and get the requester's own PDF form back with your approved answers filled in. Plus customers and auditors invited onto a single assessment, policy gap analysis with suggested language, published policies as evidence, goals with key results, vendor data scopes, loops from a template or a sentence, charts in chat, and a command palette that knows where you are.","md",[137,139,141,143,146,149,151,153,156,158,161,164,167,170,172,175,178],{"label":69,"text":138},"Dropping files on the Questionnaires list creates the questionnaire at once. Each file shows its progress — uploading, reading, finding the questions, \"48 questions · 12 already answered\" — and a failed file has its own Retry. The record's name, requester (the issuer, never the applicant), contact, due date, and summary fill in as files are read, and a field someone has edited is never overwritten.",{"label":69,"text":140},"Export → Filled PDF fills the requester's own form with your approved answers. Yes\u002Fno and select answers tick the matching box, multi-selects tick each choice and clear the rest, and text answers fill the text field. The filled file is kept with the questionnaire, and downloading the same answers again returns the same file.",{"label":75,"text":142},"Invite a customer or an auditor directly onto a program or an assessment. The invite email names what they are joining, and a guest sees only the programs and assessments they were added to — read-only pages, their own My Work, and, as a contributor, evidence upload, status, and comments on the items assigned to them.",{"label":144,"text":145},"Assessments","Programs and assessments have a Members tab with one Add people modal — search the workspace or type an email to invite someone new, then choose Can contribute or Read only. Notifications and evidence requests follow membership, and the assessment's Tasks tab now holds evidence requests alongside tasks.",{"label":147,"text":148},"Policies","Policies → Gap analysis checks your policies against a framework, control by control, and reports documentation coverage as No policy, Not addressed, Partly addressed, or Addressed. A control with no linked policy is judged against every policy in the workspace, and a gap comes with suggested language. Add the language, link the policy, raise a finding, or dismiss with a rationale — one row at a time or with Fix all, which saves each policy once.",{"label":147,"text":150},"A published policy can be the evidence for controls whose requirement is to have a documented policy. Each published version is its own evidence record, judged against its own text, and its credit is withdrawn when a new version is published, the policy is retired, you stop using it, or twelve months pass after approval.",{"label":90,"text":152},"Goals are now the measure of the security program. Link programs, frameworks, risks, work, policies, and vendors to a goal, add key results with measurable targets, and record achieved or abandoned as a decision with a rationale. The assistant suggests links when it sees work that serves a goal.",{"label":154,"text":155},"Vendors","Tag a vendor with the data it touches — PCI, PHI, PII, financial, customer data, production access, source code, AI data — and its tier rises to that scope's minimum, so a PCI vendor is Critical. Tier requirements accept alternatives, so a SOC 2, ISO 27001, HITRUST, FedRAMP, or CSA STAR report, or a questionnaire, can each satisfy \"security attestation\".",{"label":102,"text":157},"New loops start from a gallery of 13 templates across findings, vendors, risk and exceptions, reporting, and operations. Or describe the loop in a sentence — \"tell me when a vendor's SOC 2 report is about to expire\" — and the assistant picks the trigger, builds it, and opens the editor beside the conversation. Either way the loop starts unpublished and off until you publish it.",{"label":159,"text":160},"Chat","Ask the assistant for a chart and you get a real one, drawn with the same chart kit as the dashboards. The assistant picks the metric; episki computes every number from your workspace, and each segment links to the filtered list behind it.",{"label":162,"text":163},"Overviews","Controls, Tasks, Issues, and Evidence open on an overview — evidence coverage, owner load, status, due dates, ageing, and severity — and every tile and bar segment links to exactly the rows it counts.",{"label":165,"text":166},"Command palette","⌘K opens on what you can do where you are, with each action's shortcut beside it. Create any record from any page, jump back to recently viewed records, and press Tab to hand what you typed to the assistant.",{"label":168,"text":169},"Evidence","AI evidence review reads screenshots, scanned PDFs, and document text instead of judging a file by its name. Every verdict now records what it rested on — the evidence version, the model that answered, the prompt version, and a verbatim quote that episki checks against the source — so an auditor can sample AI verdicts like any other.",{"label":126,"text":171},"Deleted records go to Recently deleted for 30 days and can be restored from there. Admins can purge early, and the activity log stays readable through delete, restore, and purge.",{"label":173,"text":174},"ISO","ISO 27001, 42001, and 27701 assessments can be scored with ISO's own outcomes — Conforming, Opportunity for Improvement, Minor and Major Nonconformity, Not Tested, and N\u002FA.",{"label":176,"text":177},"Controls","The control overview shows the framework's implementation guidance under the description, starting with ISO\u002FIEC 42001 Annex B, linked to its source.",{"label":179,"text":180},"Notifications","Desktop is now a notification channel next to Email, Chat, and Slack. In the desktop app, events arrive as native notifications and open the record when clicked.",[182,185,188,191],{"label":183,"text":184},"Security","A hardening pass tightened row-level security across tenants, stopped trusting client-written integration, domain, and identity columns, pinned each session to its own workspace, rate-limited the AI routes, and hardened markdown rendering, OAuth callbacks, storage paths, and inbound mail matching. Chat attachments and agent conversations are private to the person who created them.",{"label":186,"text":187},"Findings","Findings that resolve automatically move to Done, and two findings opened for the same problem at the same moment converge into one.",{"label":189,"text":190},"Integrations","AWS reads S3 buckets with dots in their names, and a Microsoft Graph walk that times out returns what it found instead of failing.",{"label":192,"text":193},"Frameworks","Frameworks adopted from the Frameworks page match their catalog entry, so mappings and guidance apply.",{"title":195,"description":196,"icon":197,"items":198},"The questionnaire goes back the way it came","Customers, insurers, and auditors send questionnaires as files, often more than one per request: a renewal application plus a ransomware supplemental. Drop them on the Questionnaires list and the request exists immediately, with each file read in the background and the requester, contact, due date, and summary filled in as it goes. Answers already ticked or typed into a fillable PDF import as answered. Once the answers are approved, Export → Filled PDF returns the requester's own form with every checkbox, radio button, and text field set, and nothing else touched.","i-lucide-file-input",[199,200,201,202],"Drop several files at once; they become one questionnaire with live progress per file","Fillable PDF forms keep the answers already in them","Approved answers are written back into the original PDF form, not a new document","Anything that could not be placed is listed on the Answers tab, and everything else is still filled",{"src":204},"\u002Fimages\u002Fblog\u002Fsecurity-questionnaires.webp",[206,209,212,215,217,220],{"label":207,"text":208},"Onboarding","Guided setup starts from your company website. It reads the site and its DNS to suggest your industry, the frameworks it names, your identity provider and tools, and your hosting — each labelled as a suggestion for you to confirm. It asks which standards are on your radar and whether anyone has asked for them yet, and recommends only the frameworks your answers call for.",{"label":210,"text":211},"Dates","Each person has a time zone in Settings → Profile, automatic by default, and emails use it. A due date picked as the 20th is the 20th for everyone, in every zone.",{"label":213,"text":214},"AI","The assistant runs on Claude Sonnet 5.5 and Haiku 5.5, and drafting assessment responses and extracting questionnaires run concurrently.",{"label":159,"text":216},"A reply cut off mid-stream is recovered from the saved turn, attachments are limited per message rather than per chat, spreadsheets can be attached, and the agent can file an attached document as evidence.",{"label":218,"text":219},"Lists","Tables run to the right edge with consistent row hover and bulk select, every bulk action lives in one actions palette, and bulk deletes can be undone. The created toast has an Open action on every create path.",{"label":147,"text":221},"A new document's owner defaults to the person who created it, and a published policy keeps its rationale.",{},true,"\u002Fblog\u002F2026-11-17-questionnaire-pdf-guests-policy-gaps",{"title":5,"description":134},"3.blog\u002F2026-11-17-questionnaire-pdf-guests-policy-gaps","gdcdm-3f_vKyzouYi0RBfJUNUg7hh2yKFatI7LsJxXI",[229,234],{"title":230,"path":231,"stem":232,"description":233,"children":-1},"Security Certifications: Are They Worth It?","\u002Fblog\u002Fsecurity-certifications-worth-it","3.blog\u002F2026-11-13-security-certifications-worth-it","CISSP, CISA, Security+, ISO Lead Auditor. Which personal certifications actually move a security or GRC career forward, which ones are just letters, and how to decide before you spend the money.",{"title":235,"path":236,"stem":237,"description":238,"children":-1},"Agent-first GRC: what changes when AI runs the program","\u002Fblog\u002Fagent-first-grc","3.blog\u002Fagent-first-grc","Most GRC tools added AI as a feature. Agent-first GRC treats agents as the operator — drafting policies, answering questionnaires, and running the program with humans approving the work that matters.",1795055240264]