AI by task

Artificial intelligence (AI) for knowledge management: grounded answers, living SOPs and connected work

Knowledge management is the everyday machinery of how a business actually runs: the SOPs, policies, wikis, resolved tickets and contracts, plus the unwritten know-how that lives only in people's heads, scattered across email, shared drives, chat threads and business systems. Every organisation has it, because every organisation has questions that need a reliable answer and steps that must be repeated the same way every time. AI changes the economics here: it reads messy, unstructured content, answers in plain language with links back to the source, and increasingly carries out multi-step actions across the systems where the work happens.

That knowledge is usually trapped. A long-standing McKinsey study put the average knowledge worker at nearly a fifth of the working week just searching for internal information or tracking down the colleague who holds it. The figure predates today's tools, but the underlying problem, knowledge that is scattered, unsearchable and dependent on one person, is still the norm in most organisations, and it is exactly the single-point-of-failure risk that stalls a small business the moment a key person is off.

AI attacks it in three ways. It answers questions from your own curated content rather than the open internet, a pattern called retrieval-augmented generation, so answers stay company-specific, current and auditable. It turns raw material, a meeting transcript, an old checklist, a screen recording, into a clean structured procedure a person approves. And through standardised connectors it can read from one system and act in another, with a person supervising anything that carries consequences.

The useful side effect is that building the assistant forces scattered tribal knowledge into a documented, searchable form, because content that is missing or wrong becomes visible the first time someone asks about it. New starters ramp up faster, experienced staff field fewer repetitive questions, and the business becomes more resilient when someone is on leave or moves on.

The boundary is honesty about the limits. The commonest failure is retrieval, not invention: if the right document is missing or stale, the assistant answers confidently and wrongly. Even the best production deployments answer only a portion of questions well, so the realistic goal is meaningful deflection with a human fallback, and the system needs a named owner who keeps the source content current and the judgement calls with a person.

Internal Q&A over your own documents (a grounded assistant)

How it works: You connect an AI assistant to the knowledge sources you already have: your wiki, shared drives, policy PDFs, resolved support tickets. When someone asks a question, the system first retrieves the most relevant passages, then the model writes a plain-language answer grounded in them, ideally with links back to the source so the person can check it. The assistant answers from your own curated content rather than from whatever it memorised in training, which keeps answers specific, current and auditable.

Example: A dealership's service advisors ask an internal assistant how a particular warranty claim is handled and get a cited answer from the manufacturer's own documentation in seconds. As ecosystem proof at scale, Uber built an internal Slack copilot called Genie that answers engineers' questions from internal sources; per Uber's engineering blog it expanded to 154 Slack channels and answered over 70,000 questions, though its reported 48.9 per cent helpfulness rate is a useful reminder that even a successful assistant still needs a human fallback.

The benefit: Cuts the time your people lose hunting for answers and takes repetitive questions off senior or on-call staff, who are otherwise a bottleneck. Because answers cite their sources, the assistant doubles as a way to find the authoritative document, not just a paraphrase of it.

Turning raw material into clean, structured SOPs

How it works: Instead of starting from a blank page, you give the AI the raw inputs you already have: a meeting transcript where someone explained the process, an email thread, an old checklist, a screen recording. You ask for a structured standard operating procedure with numbered steps, roles, inputs and edge cases, and a human owner reviews and approves it. The AI keeps procedures consistent in tone and format, and rewrites them when the process changes, so documentation does not rot the moment the process shifts.

Example: A clinic's practice manager records a short call walking through how a new patient is registered and referred, then asks the model to turn the transcript into a step-by-step SOP with a checklist and a list of exceptions. The same approach drafts onboarding guides and policy summaries from scattered notes, with the owner correcting anything wrong before it is published.

The benefit: Documentation actually gets written, because the effort drops from hours to minutes and the activation energy that usually leaves processes undocumented disappears. That attacks the single point of failure directly: critical process knowledge stops living only in one person's head.

Cross-system workflow automation with connected tools

How it works: Through standardised connectors, most prominently the Model Context Protocol, the AI can be given controlled, permission-scoped access to your live business systems: CRM, ticketing, calendar, email, accounting. It then reads from one system and acts in another as a multi-step workflow: look up a record, draft a document, create a task, update a status. High-risk steps, anything that writes to a system of record or sends an external message, sit behind human approval, and every action is logged.

Example: A new enquiry arrives at a law firm. The assistant pulls the matter record, checks the fee-earner's calendar for availability, drafts a tailored follow-up email and creates a task, then pauses and waits for a person to approve before anything actually goes out.

The benefit: Removes the copy-paste-and-context-switch tax of moving information between tools by hand, and lets routine operational chains run with a person supervising rather than executing every keystroke.

Faster onboarding and capturing unwritten know-how

How it works: A grounded Q&A assistant lets new starters ask questions in natural language and get cited answers from the same internal sources a veteran would point them to, at any hour, without interrupting a colleague. There is a useful side effect: building the assistant forces scattered tribal knowledge into a documented, searchable form, because content that is missing or wrong becomes visible the first time someone asks about it.

Example: A new front-desk hire at a hotel asks how a late checkout or a group booking is handled and gets a cited answer drawn from internal policy, instead of waiting for a busy supervisor to reply. Where the assistant cannot answer, that gap flags a missing or outdated document for the knowledge owner to fix.

The benefit: Shortens ramp-up time and eases the load on experienced staff, while making the organisation more resilient when a key person is on leave, unavailable or moves on.

How ready the AI technology is

Grounded internal Q&A, SOP drafting and document summarisation can be built and deployed reliably today, and cross-system agentic automation is maturing fast on the agent-drafts, human-approves pattern. Fully unattended multi-step operations remain early, and published productivity figures are a bearing rather than a promise.

Grounded internal Q&A is the most proven piece: the Uber Genie deployment has run in production since September 2023 at a scale no pilot matches. A system of this kind can be built now on your own wikis, drives and resolved tickets, with answers that cite their sources, and drafting and summarisation with a human editor, SOPs, minutes, FAQ answers, are dependable and low risk precisely because a person reviews before anything is used.[1]
Cross-system automation has matured unusually fast: the Model Context Protocol went from Anthropic's November 2024 launch to a cross-vendor de facto standard within about a year, adopted by OpenAI, Google, Microsoft and AWS. Connectors to CRM, ticketing, email and calendar are real and usable today; the dependable pattern is the agent drafts and a person approves, especially for anything that writes to a system of record.[1]
Where to stay sober: nobody gets a flawless, self-maintaining knowledge brain by pointing AI at a folder. Even Uber's successful deployment answers barely half of questions well by its own helpfulness rate, so the realistic goal is meaningful deflection with a human fallback, not every question answered. Treat published gains as somebody else's result, not a promise of yours, and measure deflection and answer quality on your own questions before scaling.

What to watch out for

The commonest failure is retrieval, not invention: if the right document is missing or stale, the assistant answers confidently and wrongly. In the UK the assistant must also respect existing access permissions and UK GDPR duties, and it needs a named owner who keeps the source content current.

The commonest real-world failure is retrieval, not invention: if the right document was never found, whether missing, out of date, badly tagged, or siloed in a system the assistant cannot see, the AI cannot answer correctly however capable the model is. Stale knowledge in means confident-sounding wrong answers out, and an answer can be faithful to its source yet still incomplete, so track which questions the assistant fails or declines rather than assuming it works.
UK data protection applies in full. There is no UK AI act, but the ICO already regulates AI through the UK GDPR and its dedicated guidance on AI and data protection, and an assistant that reads across systems must inherit and respect existing access permissions so it never surfaces salary data or confidential files to someone who should not see them. The fines behind that reach 17.5 million pounds or 4 per cent of worldwide annual turnover.[1]
Agentic actions need governance, and the whole system needs an owner. Anything that changes records or sends messages should require approval or run with tightly limited scopes and an audit log, and someone must be accountable for keeping the source content current and periodically evaluating answer quality. Without that, accuracy quietly degrades and the assistant slowly loses the team's trust. The judgement calls, and any action that commits the business, stay with a person.[1]

AI for knowledge management, in your industry

Pick your field to see how knowledge management works in context, and jump straight into that industry's page.

Manufacturers

→

Writes nonconformance reports from the job and surfaces patterns you can see.

Law firms

→

Accelerates legal research, verified against the primary source every time.

Law firms

→

Makes years of the firm's own best work something fee-earners can ask.

Conveyancers

→

Answers from your own approved precedents and notes, with a citation.

Mortgage brokers

→

Builds a fast lender shortlist, the adviser still owning the recommendation.

Mortgage brokers

→

Assembles the suitability file and gap-checks it before a review lands.

Pharmacies

→

Logs, routes and escalates complaints and GPhC-facing documentation.

Road haulage

→

Collates tachograph downloads and surfaces infringements for review.

Veterinary practices

→

Produces POM-V labels and records for a decision the vet has already made.

Home care providers

→

Keeps CQC evidence and inspection readiness current under the framework.

Block management

→

Keeps the golden thread of building-safety documentation current.

Engineering consultancies

→

Keeps the building-safety golden thread in order, competence left to the dutyholder.

Engineering consultancies

→

Drafts meeting minutes and actions, design decisions confirmed by the engineer.

Svennis Cloud Solutions

We build them, on Claude

These AI flows do not stay on paper. Svennis Cloud Solutions builds and integrates them into your systems, with a team of certified Claude architects, on Anthropic technology, from the first WhatsApp message to the finished invoice.

See live what we can doCertified Claude architects

See what this looks like in your business

Describe your business in one sentence and Claude will show you live where AI would make the biggest difference, using your own numbers.

Country

or start from an example:

Guide: AI for any company

The processes every business has, in detail.

→

Sources

Find out how ready your business is for AI