How to Link Modern AI Into Your Dated Intranet
Learn how to add modern AI capabilities to your existing SharePoint, Confluence, or legacy intranet. No full rebuild required.
The Intranet Everyone Hates (But No One Can Replace)
You know the one.
It was built in 2015. Maybe 2010. It runs on SharePoint, or Confluence, or some custom thing that the developer who built it has long since left.
Everyone complains about it. Finding anything takes forever. The search is useless. Documents are buried 5 folders deep.
But replacing it? Thatâs a ÂŁ500k project and 2 years of pain.
Hereâs the good news: you donât need to replace it.
The Third Option: AI as a Better Front Door
You have three choices:
Option 1: Live with the pain (what youâre doing now)
Option 2: Full replacement project (expensive, risky, 18+ months)
Option 3: Add an AI layer that makes your existing intranet actually usable
Option 3 is what I want to talk about.
What âAI Layerâ Actually Means
Imagine this:
Your employee asks: âWhatâs our parental leave policy?â
Old way:
- Open intranet
- Try search (fails)
- Navigate: HR â Policies â Benefits â Family â Parental Leave
- Find 3 documents, not sure which is current
- Read through to find the answer
- Total time: 15 minutes
With AI:
- âWhatâs our parental leave policy?â
- âEmployees get 26 weeks paid leave. Hereâs the full policy: [link]â
- Total time: 30 seconds
The AI reads your existing documents. It knows whatâs in them. It answers questions in natural language.
Your intranet stays exactly the same. AI just makes it accessible.
What Modern AI Can Actually Do with Legacy Systems
1. Natural Language Search
Not keyword matching. Actual understanding.
âWho do I contact about expenses?â â Answer + contact details + relevant forms
âWhatâs the process for requesting time off?â â Step-by-step answer
âWhereâs the latest brand guidelines?â â Direct link to current document
2. Answer Questions from Documents
AI reads your PDFs, Word docs, wikis, and pages.
âWhatâs our discount policy for enterprise clients?â â Pulls answer from sales playbook
âWhat software do we use for X?â â Answers from IT documentation
âHow do I set up VPN?â â Step-by-step from IT guides
3. Surface Related Information
AI understands context.
Looking at a project brief? AI suggests related templates, previous similar projects, relevant contacts.
4. Multi-Source Answers
Your information is scattered across SharePoint, Confluence, Google Drive, and that one wiki nobody remembers.
AI searches all of them. One question, comprehensive answer.
How It Works Technically (Simple Version)
Step 1: Connect to your systems
We connect to:
- SharePoint / Microsoft 365
- Confluence
- Google Workspace
- File servers
- Custom databases
- Whatever you have
Step 2: Index your content
AI reads and understands your documents. Not just keywords â actual meaning.
Step 3: Deploy the interface
Options:
- Chat widget on your intranet
- Standalone page
- Microsoft Teams integration
- Slack integration
- All of the above
Step 4: Employees ask questions
They type naturally. AI answers. With sources.
What doesnât change:
- Your existing intranet stays exactly as is
- Your file structure stays the same
- Your permissions stay the same
- Your content management process stays the same
Timeline: 2-4 weeks for most organisations
Security and Permissions (The Important Bit)
Your data stays secure:
- AI respects existing permissions
- If you canât access a document, AI wonât show you info from it
- No data sent to third parties for training
- Can be deployed on your infrastructure if required
- Audit logs for all queries
Compliance covered:
- GDPR compliant
- Can meet financial services requirements
- Healthcare/NHS compatible
- Government security standards available
Real Results from Real Companies
Manufacturing Company (2,000 employees)
- Time spent searching for information: Down 65%
- IT helpdesk tickets: Down 40%
- Employee satisfaction with intranet: Up 50 NPS points
Professional Services Firm (500 employees)
- New employee onboarding time: Reduced by 2 weeks
- âWhere do I find X?â questions to colleagues: Down 70%
- Policy compliance: Improved (people actually find and read policies)
Retail Head Office (300 employees)
- Store manager queries to HQ: Down 55%
- Time to answer routine questions: From hours to seconds
- HR team admin time: Saved 15 hours/week
What It Wonât Fix (Honesty Section)
Outdated information
If your documents are from 2018, AI will give 2018 answers. You still need to keep content updated.
Missing documentation
If something isnât documented anywhere, AI canât magically know it. But it WILL highlight gaps.
Complex approval workflows
AI can explain the process. It canât replace your approval systems.
Political/organisational issues
If nobody uses the intranet because of culture, AI helps but doesnât solve the root cause.
The Cost Reality
What youâre paying now:
- Hours of employee time searching for things
- Duplicated work because people canât find existing resources
- Wrong decisions from outdated information
- IT helpdesk overload with âwhere is X?â questions
What AI costs:
- Setup: Depends on complexity
- Ongoing: Based on usage
- Typical range: ÂŁ500-2,000/month for mid-size companies
ROI calculation:
If you save each employee 30 minutes/week searchingâŠ
- 200 employees Ă 0.5 hours Ă ÂŁ30/hour = ÂŁ3,000/week saved
- Thatâs ÂŁ156k/year in recovered productivity
Is This Right for You?
Good fit if:
- You have lots of documentation nobody can find
- Employees constantly ask âwhere is X?â
- Your search function is basically useless
- Replacing your intranet isnât practical
- You want quick wins without major IT projects
Maybe not if:
- You have very little documented content
- Youâre a tiny team (under 20 people)
- Youâre already planning a full replacement
Want to See What This Looks Like?
We can demo with sample content similar to yours.
Or, if youâre comfortable, we can do a proof-of-concept with your actual intranet.
30-minute call to see if this makes sense for your organisation.