{"slug": "csv-excel-and-google-sheets-in-famulor-knowledge-bases", "title": "CSV, Excel and Google Sheets in Famulor Knowledge Bases", "summary": "Famulor added support on September 5, 2026 for CSV, TSV, and XLSX files, plus Google Sheets connected through Drive, as structured knowledge-base sources, according to the company's changelog. The feature preserves original cell values and column positions with source file, sheet, and row references, supporting exact identifier lookup, description-based search, and defined filters and aggregations, but imported tables remain a snapshot from the last import rather than a live system. Famulor's documentation states that joins, arbitrary expressions, currency conversion, and formula evaluation are unsupported, so live inventory, personalized balances, order status, and binding current prices still require a tool or API.", "body_md": "### Summarize Content With:\n\nProduct codes, service descriptions, and reference prices often live\nin spreadsheets. Using them with an AI assistant has meant converting\nrows into prose or building an API for current values. Since September\n5, 2026, Famulor can process CSV, TSV, and XLSX files, as well as Google\nSheets connected through Drive, as structured knowledge-base sources.\nThe feature supports exact identifiers, description-based search, and\ndefined filters and aggregations ([Famulor changelog,\nSeptember 5, 2026](https://docs.famulor.io/changelog#2026-09-05)).\n\nThe critical qualifier is that an imported table remains a\n**snapshot from the last import**, not a live system.\nInventory, personalized balances, order status, and binding current\nprices still need a tool or API.\n\n**Key takeaways**\n\n- Famulor supports rectangular CSV, TSV, and XLSX tables, plus Google Sheets through a Drive connection.\n- Table knowledge is a good fit for stable catalogs, service matrices, and reference prices with a visible effective date.\n- Exact values, rows, counts, sums, minimums, and maximums are available within documented query boundaries.\n- Google Sheets are not queried as a live database; the last import controls freshness, and formulas are not recalculated.\n- Live, personalized, or writeback workflows still belong behind a mid-call tool or API.\n\n## What changed with structured table knowledge\n\nThe new capability refines an important architecture rule. General\nknowledge bases have traditionally served unstructured material such as\nmanuals, policies, and FAQs. Treating a price table like a document made\nreliable cell-level retrieval difficult. Famulor’s table model instead\npreserves original cell values and column positions, along with source\nfile, sheet, and row references ([knowledge-base\ndocumentation](https://docs.famulor.io/assistants/knowledge-base), checked September 10, 2026).\n\nAn assistant can search for a complete product code, locate a row from its description, or use defined table queries. Results include rows, count, sum, minimum, and maximum under constrained filters. Joins, arbitrary expressions, currency conversion, and formula evaluation are unsupported.\n\nThe previous guidance in [Knowledge\nBase vs System Prompt for AI Voice Agents](https://www.famulor.io/blog/knowledge-base-vs-system-prompt-for-ai-voice-agents-2026) therefore remains\ndirectionally sound but needs a post-update qualification:\n**structured imported tables now support defined retrieval and\naggregation tasks, while live data, personalized values, and writeback\nstill require a tool or API.**\n\n## When a spreadsheet knowledge base is the right fit\n\nA table is a strong candidate when its rows describe stable reference objects and snapshot freshness is acceptable. Examples include a spare-parts catalog, a service matrix, a list of branch capabilities, a tariff table with an effective date, or an approved reference price list.\n\nAsk four questions before importing:\n\n1. **Is the data stable enough for a snapshot?** If a\nvalue changes minute by minute, a knowledge base is the wrong\nsource.\n2. **Does each row represent a well-defined object?** A\nproduct, service, or region should be described through consistent\ncolumns.\n3. **Are filters and simple aggregations sufficient?** Multi-table logic, arbitrary calculations, and currency conversion need\nanother system.\n4. **Should every imported sheet be indexed?** Hidden XLSX\nsheets and every exported Google Sheets tab are included.\n\nThis separates knowledge retrieval from transactions. A catalog identifier may come from an approved import. “Is part X in stock right now?” requires a current system query, and a binding customer-specific quote should not be inferred from a general reference table.\n\n## Prepare CSV, TSV, XLSX, and Google Sheets before import\n\nSource quality starts with a clean structure. According to the\ncurrent [Famulor\ndocumentation](https://docs.famulor.io/assistants/knowledge-base.md), column headings must appear in the first populated\nrow. The table needs to be rectangular; merged cells and values beyond\nthe width defined by the header are rejected. Put separate tables on\nseparate sheets.\n\nThe documented format rules are specific:\n\n- CSV can use commas or semicolons as delimiters; multiline values must be quoted correctly.\n- TSV uses tab separators.\n- Supported text encodings are UTF-8 or UTF-16 with a BOM.\n- Legacy `.xls` files need to be saved as XLSX or CSV\nfirst.\n- XLSX imports include every sheet, including hidden sheets. Google Sheets imports include every exported sheet.\n- Formulas are not recalculated during import. Cached results can be marked unverified; missing results and cell errors remain unavailable.\n\nRemove helper sheets, confidential notes, and unnecessary columns\nbefore import. Add `valid_from` or `data_as_of`\nfor time-sensitive references so a conversation can state the source\ndate.\n\n## Set up the table in Famulor in five documented steps\n\nThis flow follows the current Famulor documentation linked above.\n\n### 1. Create the knowledge base\n\nOpen **Knowledge bases → New**, assign a clear name, and\ndescribe what the source contains. “DACH service catalog — September\n2026 snapshot” communicates purpose and freshness better than “new price\nlist.”\n\n### 2. Add the table source\n\nUpload a CSV, TSV, or XLSX file. For Google Sheets, first create a\nDrive OAuth connection under **Automations → Connections**,\nthen select that connection when adding the Drive source. The\ndocumentation describes cloud-drive sync as Beta and a separate plan\nfeature. It requires **Beta Features** under\n**Settings → Workspace** and follows the workspace’s\ncurrent sync rates (Famulor documentation, checked September 10, 2026).\nCheck the conditions in your own workspace rather than assuming\nuniversal availability.\n\n### 3. Check processing status\n\nWait until the item reports that it is ready. If processing fails, inspect the format, file size, header row, merged cells, and sheet layout. A visible filename alone does not prove that a new searchable index was created.\n\n### 4. Test with RAG Search\n\nUse **RAG Search** with realistic questions. As an\nimplementation practice, test an existing exact code, a nonexistent\nsimilar code, a description query, several matches, and a filter or\naggregation. This is a recommendation, not a product guarantee.\n\n### 5. Assign the knowledge base to the assistant\n\nSelect the knowledge base in the assistant editor. Famulor says that connection alone does not require a prompt change. Add workflow instructions for ambiguity, stale snapshots, or critical identifiers when needed.\n\n## What the assistant can ask of the table\n\nFor an exact identifier, search for the complete value in quotation marks. A similar-looking code is not evidence that the requested code exists. If several rows match, the assistant should clarify instead of choosing. Search results retain source references and positional column keys, including file, sheet, and row for traceability. That does not mean those references are automatically spoken to a caller.\n\nAdvanced implementations can use table queries through\n`POST /knowledge-bases/{id}/search` and the\n`search_knowledge_base` MCP tool. Under the current query\ndocumentation linked above, a query can return rows, `count`,\n`sum`, `min`, or `max`. It can combine\nup to five equality or range filters with AND; number and date ranges\nrequire an explicit type.\n\nThere is a firm trust boundary around numeric outcomes. Totals are rejected when matching rows contain missing, ambiguous, formula-based, or mixed-currency values. Do not calculate a full-table total by manually adding a handful of search excerpts. Use a table query—or a responsible backend when the operation is outside the supported set.\n\n## Practical example: a hypothetical B2B parts and service catalog\n\nSuppose an industrial service company maintains an approved reference catalog. The miniature table below is deliberately fictional and contains no real customer or product data:\n\nThe source file contains these three example records:\n\n- `FP-2048-A` : Standard filter package for MX-20, DACH\nregion, EUR 189.00 reference price, effective September 1, 2026.\n- `SV-3100-D` : Basic remote service for MX-30, DACH region,\nEUR 420.00 reference price, effective September 1, 2026.\n- `SV-3190-D` : Plus remote service for MX-30, DACH region,\nEUR 690.00 reference price, effective September 1, 2026.\n\nA caller might ask for “FP-2048-A.” Exact matching should identify\nthat record; “FP-2048-B” must not be treated as present merely because\nit looks similar. A second question might be, “How many DACH service\npackages for the MX-30 have a list price below €500?” Equality and range\nfilters plus `count` are designed for that request. The words\n**list price** and the effective date still matter in the\nresponse; this record must not silently become a binding current\noffer.\n\nFor voice channels, add an explicit confirmation routine as an implementation recommendation: have the assistant repeat critical codes in understandable groups and ask the caller to confirm before triggering a downstream step. This is not a claim that Famulor automatically spells or validates identifiers. Test likely pronunciations, background noise, and easily confused character sequences in real test calls.\n\n## Snapshot or live system? Make the boundary explicit\n\nKnowledge-base data reflects the last import. Google Sheets are therefore not queried live for every question. Famulor does not recalculate formulas and does not support joins, arbitrary expressions, or currency conversion. These are architecture-defining constraints, not footnotes (Famulor documentation, checked September 10, 2026).\n\nThe separate [Google Sheets\nintegration](https://www.famulor.io/integrations/google-sheets) and [Microsoft\nExcel 365 integration](https://www.famulor.io/integrations/microsoft-excel-365) pages cover workflow actions, triggers, and\nwriteback. This guide deliberately stays with retrieval and constrained\naggregation over imported table snapshots.\n\n| Requirement | Appropriate source | \n|---|---|\n| Stable product or service catalog with a snapshot date | Table knowledge base | \n| Exact reference identifier or description search | Table knowledge base | \n| Constrained filters, count, sum, minimum, or maximum | Table query within documented boundaries | \n| Current inventory or order status | Live tool/API | \n| Customer-specific balance or binding price | Authorized live tool/API | \n| Change a record, create an order, or update a CRM | Write-enabled tool/workflow | \n| Join across systems or arbitrary calculation | Responsible backend/API | \n\nIf a replacement or retry fails, the previous searchable index\nremains available and the error is reported, according to Famulor’s\ndocumentation. This preserves continuity, but it specifically does\n**not** prove that the new data was accepted. Check the\nstatus, snapshot date, and test set after every import.\n\n## Current limits at a glance\n\nThe values below come from the Famulor documentation linked above, checked on September 10, 2026, and may change over time:\n\n| Limit | Documented value | \n|---|---|\n| Records including headings across all sheets | 1,000 | \n| Columns | 50 | \n| Sheets | 20 | \n| Characters per rendered row | 6,000 | \n| Maximum XLSX expansion | 32 MB | \n| Google Sheets export | Explicit failure above 10 MB | \n| General knowledge-base capacity | 25 files, up to 20 MB each | \n\nFormal compliance with a size limit does not make a file semantically safe: ambiguous types, mixed currencies, or hidden tabs can still make a query unsuitable.\n\n## Privacy and telephone-operation checklist\n\nAccording to the documentation linked above, Famulor retains the\noriginal table file and XLSX imports include hidden sheets. A practical\ngovernance rule follows: **import only fields and sheets that are\nauthorized and necessary for this purpose.** This is\nimplementation advice, not a compliance guarantee.\n\nBefore rollout, verify the following:\n\n- Does the source contain personal customer or employee data that does not belong in a general catalog?\n- Are hidden sheets, comments, and helper columns genuinely intended for the assistant?\n- Is the Drive connection restricted to the required source and internally authorized?\n- Does the table include an effective or snapshot date for prices and tariffs?\n- Does the conversation avoid presenting reference values as current, binding promises?\n- Is there a clarification, tool call, or human handoff for unconfirmed codes, multiple matches, and missing live data?\n\nTelephone conversations add transmission risk: letters, digits, and similar-sounding identifiers can be misheard or misrecognized. Confirm critical identifiers and connect binding actions to a current system response. The assistant should state uncertainty rather than infer an answer from a similar match.\n\n## Conclusion: use table knowledge deliberately as a snapshot\n\nCSV, TSV, XLSX, and Google Sheets can now be more than flattened documents in Famulor. As structured knowledge sources, they support exact identifiers, description search, and tightly bounded table queries. That fills a practical gap for product catalogs, service matrices, and approved reference price lists.\n\nThe useful design question is not simply “spreadsheet or API?” but “snapshot or live process?” If the source is stable, rectangular, and clearly dated, a table knowledge base is a strong candidate. If the workflow is current, personalized, or write-enabled, an authorized tool or backend remains responsible. Prepare the source carefully, test exact and ambiguous cases with RAG Search, and verify that each new import is actually ready. That keeps the feature within its documented strengths without turning reference knowledge into an accidental real-time promise.\n\nWriter at Famulor", "url": "https://wpnews.pro/news/csv-excel-and-google-sheets-in-famulor-knowledge-bases", "canonical_source": "https://www.famulor.io/blog/spreadsheet-knowledge-base-famulor", "published_at": "2026-09-11 02:07:00+00:00", "updated_at": "2026-09-26 23:30:32.756781+00:00", "lang": "en", "topics": ["ai-products", "ai-tools", "structured-data", "artificial-intelligence"], "entities": ["Famulor", "Google Sheets", "Google Drive", "CSV", "TSV", "XLSX"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/csv-excel-and-google-sheets-in-famulor-knowledge-bases", "markdown": "https://wpnews.pro/news/csv-excel-and-google-sheets-in-famulor-knowledge-bases.md", "text": "https://wpnews.pro/news/csv-excel-and-google-sheets-in-famulor-knowledge-bases.txt", "jsonld": "https://wpnews.pro/news/csv-excel-and-google-sheets-in-famulor-knowledge-bases.jsonld"}}