Features
Today ServiceNow has a large collection of AI and ML capabilities. AI in a Box is a product that gives you access to a self-hosted large language model server (llm) and a vector database (vectordb).
Let’s compare the two across, accuracy, speed, and cost.
A brief comparision of features so far;
- 👍 = Available
- 🚧 = Work in progress
- 🛣️ = On roadmap
- 🚫 = Blocked
| What | ServiceNow | AI in a Box |
|---|---|---|
| Summarization | 👍 (1 assist) | 👍 |
| Conversational agent | 👍 (10 assists) | 🚧 |
| Chat Summarization (and short description generation) | 👍 (1 assist) | 🛣️ |
| KB Article generation | 👍 (10 assists) | 🛣️ |
| Suggested Resolution Notes | 👍 (1 assist) | 🛣️ |
| Suggested Field Values (Predictive Intelligence) | 👍 ? | 🛣️ |
| Comment to Code | 👍 (1 assist) | 🛣️ |
| Flow recommendation | 👍 (1 assist) | 🚫 |
| Interface chat in now assist panel | 👍 (5 assists) | 🚫 |
| Q&A in AI search | 👍 (5 assists) | 🚫 |
| 👍 (10 assists) | 🚫 |
Summarization
ServiceNow has a model for each capability. That being said their summarization model may be more accurate than the one provided by AI in a Box. The AI in a Box model is designed to be fast and cost effective and customizable.
The speed of AI in a Box is trying to stay under 30 seconds. ServiceNow’s model is not available to compare.
| Category | ServiceNow | AI in a Box |
|---|---|---|
| Accuracy | 👍 | 👍 |
| Speed | 👍 | 👍 |
| Cost | 💲💲💲💲💲 | 💲 |
Suggested Resolution Notes
This feature doesn’t yet exist in AI in a Box but is in active development and is depenendent on the VectorDB.
ServiceNow’s does have this.
| Category | ServiceNow | AI in a Box |
|---|---|---|
| Accuracy | 👍 | 🚧 |
| Speed | 👍 | 🚧 |
| Cost | 💲💲💲💲💲 | 💲 |
Comment to Code
This feature doesn’t yet exist in AI in a Box but is in active development and is depenendent on the VectorDB.
ServiceNow’s does have this.
| Category | ServiceNow | AI in a Box |
|---|---|---|
| Accuracy | 👍 | 🚧 |
| Speed | 👍 | 🚧 |
| Cost | 💲💲💲💲💲 | 💲 |
Chat Summarization
This feature doesn’t yet exist in AI in a Box but is on the roadmap and is depenendent on the VectorDB.
ServiceNow’s does have this.
| Category | ServiceNow | AI in a Box |
|---|---|---|
| Accuracy | 👍 | 🛣️ |
| Speed | 👍 | 🛣️ |
| Cost | 💲💲💲💲💲 | 💲 |
Flow recommendation
This feature is on a page in ServiceNow I’m not sure can be created in AI in a Box.
| Category | ServiceNow | AI in a Box |
|---|---|---|
| Accuracy | 👍 | 🚫 |
| Speed | 👍 | 🚫 |
| Cost | 💲💲💲💲💲 | 🚫 |
Suggested Field Values (Predictive Intelligence)
This feature isn’t part of the GenAI products in ServiceNow but is part of the Predictive Intelligence product.
The requirement for this in ServiceNow was to have 30k records to train against and to set up some schedule to train the model in one environment and then push it to another.
This feature is not yet in AI in a Box but can by simply doing a similarity match against the VectorDB and then presenting the results in an intuitive way.
| Category | ServiceNow | AI in a Box |
|---|---|---|
| Accuracy | 👍 | 🚧 |
| Speed | 👍 | 👍 |
| Cost | 💲💲💲💲💲 | 💲 |
Interface chat in now assist panel
This feature is on a page in ServiceNow I’m not sure can be created in AI in a Box.
| Category | ServiceNow | AI in a Box |
|---|---|---|
| Accuracy | 👍 | 🚫 |
| Speed | 👍 | 🚫 |
| Cost | 💲💲💲💲💲 | 🚫 |
Q&A in AI search
This feature is on a page in ServiceNow I’m not sure can be created in AI in a Box.
Va conversational agent
With AI in a Box you can simply do REST calls to the llm and vectordb. This is not possible in ServiceNow. They allow this some other way.
| Category | ServiceNow | AI in a Box |
|---|---|---|
| Accuracy | 👍 | 👍 |
| Speed | 👍 | 👍 |
| Cost | 💲💲💲💲💲 | 💲 |
KB Article generation
Requires Text Generation.