New: Meet Smart Collections in LianaMonitor

What's new

Smart Collections is now available in LianaMonitor. Instead of building ever-longer keyword rules, you describe in your own words what belongs in a collection, and AI picks the matching hits from the monitors you selected as sources.

Your monitors stay exactly as they are. Smart Collections works on top of them, so you can keep a monitor broad and still see only what matters. The collection is then a set you can use like any other: HitsSummary, labels that can be inherited onto the hits, and dashboard widgets.

Use case 1: One monitor, several collections

A city monitor has to stay wide. It catches culture, the council, the tram, a zoning notice, and a lot that is none of those. Tightening the keywords usually means either more noise or missed coverage, and one search cannot serve every desk.

Smart Collections leaves that monitor alone. You create a collection for each job, use the same monitor as the source, and write what that collection is for.

For example, a municipality can keep one city monitor and give each team its own collection:

• Culture. Coverage where the city’s culture, venues, or festivals are the subject. Leave out sports, council process, and passing mentions.
• Local politics. Council, mayor, budget, and political decisions. Leave out event listings and stories that only use the city as a location.
• Transportation. Coverage where the city’s transport system is the subject: tram, buses, disruption, decisions. Leave out sports, culture, and venue listings.
• Zoning and land use. Planning, zoning, and land-use disputes where the plan or the site is the subject. Leave out apartment ads and generic housing pieces.

Same sources. Four collections. Four people who should not have to read the whole city feed.

Use case 2: Cut the noise from a generic keyword

If your brand name is also an everyday word, a monitor picks up a lot of unrelated mentions. Narrowing it down with keyword rules takes time, and it is hard to be sure the rules catch everything relevant without letting the noise back in.

With Smart Collections the monitor can stay broad. The instruction brings the focus:

Liana Technologies and its products: reviews, customer feedback and company news. Skip plants and nature, people named Liana and local businesses.

Mentions about rainforest vines, name days or a café with the same name stay in the monitor but out of your collection. You see only the mentions you actually need.

Use case 3: Sentiment from your point of view

Automatic sentiment scores the tone of the whole text. A story can be upbeat overall and still be critical of you, or gloomy overall and still show you in a good light. What usually matters is how the article treats your organisation.

With Smart Collections you write that into the brief. The collection keeps the hits where your organisation is presented negatively or positively, whatever the overall tone of the story. The sentiment score on the hit does not change. The brief decides what belongs.

For example, a city’s communications team can keep two collections on the same monitor:

• Negative about us. Coverage that presents the city as an organisation in a negative light: criticism of its decisions, services, staff or communication. Leave out stories where the city is only the location, and bad news that does not blame the city.
• Positive about us. Coverage that presents the city in a positive light: praise for its services, staff or decisions, and successful city projects. Leave out event listings and neutral announcements.

The negative collection is an early warning list for communications. The positive one gathers material for reports, recruitment and social media.

Use case 4: Your role in the story

A mention is not always about you. Your organisation can be the main subject of a story, one of several parties, or just a name in a list, a quote or an event location. When you report on visibility or react to coverage, the stories where you play the main role usually matter most.

With Smart Collections you can catch hits based on your role in them. Instruct the collection to include only the hits where you are in the main role, and leave the other mentions out:

Coverage where the city is the main subject or the main actor of the story. Leave out stories where the city is mentioned in passing, listed among others, or only the location.

Narrow it down to the bigger sources

If you want an even shorter list, add a condition on the size of the source. For example:

Coverage where the city is the main subject or the main actor. Only include hits where the source’s reach is more than 500,000, or where the author has more than 5,000 followers on social media.

The result is a short list of high-visibility coverage where you are at the centre, which works well for management reports and for deciding where to react first.

How to set up a smart collection

1. Create a new collection, or open an existing one, and go to its Collection settings.
2. Switch on Smart collections.
3. Under Source monitors, choose the monitors the collection should draw from. You can pick up to five.
4. In Smart collection instructions, describe in your own words what belongs in the collection. You can use up to 500 characters.
5. Click Save collection.

Your source monitors are not changed in any way. When you save, the collection judges up to 1,000 past documents from those monitors. New hits are judged against the instruction as they arrive.

Tips for writing good instructions

Write the instruction the way you would brief a colleague who reads the mentions for you.

• Say who or what the collection is about. Name the company, product, or topic clearly, for example “the city’s transport system” rather than just the city name.
• Say what kind of mentions you want. Reviews, customer feedback, news coverage, council decisions, planning disputes.
• Say what to leave out. Name the usual sources of noise: other meanings of the word, job ads, event listings, your own press releases.
• Keep it to one topic. If you follow two different things, create two smart collections.
• Adjust as you go. If unwanted mentions still appear, add them to the “leave out” part of the instruction. You do not need to change the monitor.

Good to know

Does Smart Collections change my monitors? No. Your monitors and their keyword rules stay as they are. The smart collection only decides which of their mentions belong in the collection.

Does it fill from the whole monitor history? No. It processes up to 1,000 past documents when you save, then judges new hits as they arrive.

How many monitors can a smart collection use? Up to five source monitors per collection.

How many Smart Collections can I have? Up to 10 active Smart Collections per customer at launch.

How long can the instruction be? Up to 500 characters. A short, specific instruction usually works best.

What can I do with the collection? The same things as any collection. HitsSummary, labels that can be inherited onto the hits, and dashboard widgets. The monitor stays the source. The collection is the set you summarise and share.

Does it work in Nordic languages? Yes.

How does it decide what is relevant? AI reads each mention from the source monitors and compares it with your instruction. Mentions that match are added to the collection.