feat: implement nlp based blocks prioritization strategy

feat: add weight to nlp entity schema and readapt feat: remove commented obsolete code feat: restore settings feat: apply feedback fix: re-adapt unit tests feat: priority scoring re-calculation & enabling weight modification in builtin nlp entities fix: remove obsolete code feat: refine unit tests, apply mr coderabbit suggestions fix: minor refactoring feat: add nlp cache map type feat: refine builtin nlp entities weight updates feat: add more test cases and refine edge case handling feat: add weight validation in UI fix: apply feedback feat: add a penalty factor & fix unit tests feat: add documentation fix: correct syntax fix: remove stale log statement fix: enforce nlp entity weight restrictions fix: correct typo in docs fix: typos in docs fix: fix formatting for function comment fix: restore matchNLP function previous code fix: remove blank line, make updateOne asynchronous fix: add AND operator in docs fix: handle dependency injection in chat module feat: refactor to use findAndPopulate in block score calculation feat: refine caching mechanisms feat: add typing and enforce safety checks fix: remove typo fix: remove async from block score calculation fix: remove typo fix: correct linting fix: refine nlp pattern type check fix: decompose code into helper utils, add nlp entity dto validation, remove type casting fix: minor refactoring feat: refactor current implementation
2025-06-26 18:27:28 +00:00 · 2025-03-26 13:11:07 +01:00 · 2025-03-26 13:11:07 +01:00 · bab2e3082f
commit bab2e3082f
parent 0db40680dc
31 changed files with 1061 additions and 49 deletions
--- a/api/docs/nlp/README.md
+++ b/api/docs/nlp/README.md
@ -0,0 +1,102 @@
+# NLP Block Scoring
+## Purpose
+
+**NLP Block Scoring** is a mechanism used to select the most relevant response block based on:
+
+- Matching patterns between user input and block definitions
+- Configurable weights assigned to each entity type
+- Confidence values provided by the NLU engine for detected entities
+
+It enables more intelligent and context-aware block selection in conversational flows.
+
+## Core Use Cases
+### Standard Matching
+
+A user input contains entities that directly match a block’s patterns.
+```ts
+Example: Input: intent = enquiry & subject = claim
+Block A: Patterns: intent: enquiry & subject: claim
+Block A will be selected.
+```
+
+### High Confidence, Partial Match
+
+A block may match only some patterns but have high-confidence input on those matched ones, making it a better candidate than others with full matches but low-confidence entities.
+**Note: Confidence is multiplied by a pre-defined weight for each entity type.**
+
+```ts
+Example:
+Input: intent = issue (confidence: 0.92) & subject = claim (confidence: 0.65)
+Block A: Pattern: intent: issue
+Block B: Pattern: subject: claim
+➤ Block A gets a high score based on confidence × weight (assuming both weights are equal to 1).
+```
+
+### Multiple Blocks with Similar Patterns
+
+```ts
+Input: intent = issue & subject = insurance
+Block A: intent = enquiry & subject = insurance
+Block B: subject = insurance
+➤ Block B is selected — Block A mismatches on intent.
+```
+
+### Exclusion Due to Extra Patterns
+
+If a block contains patterns that require entities not present in the user input, the block is excluded from scoring altogether. No penalties are applied — the block simply isn't considered a valid candidate.
+
+```ts
+Input: intent = issue & subject = insurance
+Block A: intent = enquiry & subject = insurance & location = office
+Block B: subject = insurance & time = morning
+➤ Neither block is selected due to unmatched required patterns (`location`, `time`)
+```
+
+### Tie-Breaking with Penalty Factors
+
+When multiple blocks receive similar scores, penalty factors can help break the tie — especially in cases where patterns are less specific (e.g., using `Any` as a value).
+
+```ts
+Input: intent = enquiry & subject = insurance
+
+Block A: intent = enquiry & subject = Any
+Block B: intent = enquiry & subject = insurance
+Block C: subject = insurance
+
+Scoring Summary:
+- Block A matches both patterns, but subject = Any is considered less specific.
+- Block B has a redundant but fully specific match.
+- Block C matches only one pattern.
+
+➤ Block A and Block B have similar raw scores.
+➤ A penalty factor is applied to Block A due to its use of Any, reducing its final score.
+➤ Block B is selected.
+```
+
+## How Scoring Works
+### Matching and Confidence
+
+For each entity in the block's pattern:
+- If the entity `matches` an entity in the user input:
+    - the score is increased by: `confidence × weight`
+        - `Confidence` is a value between 0 and 1, returned by the NLU engine.
+        - `Weight` (default value is `1`) is a configured importance factor for that specific entity type.
+- If the match is a wildcard (i.e., the block accepts any value):
+    - A **penalty factor** is applied to slightly reduce its contribution:
+    ``confidence × weight × penaltyFactor``. This encourages more specific matches when available.
+
+### Scoring Formula Summary
+
+For each matched entity: 
+
+```ts
+score += confidence × weight × [optional penalty factor if wildcard]
+```
+
+The total block score is the sum of all matched patterns in that block.
+
+### Penalty Factor
+
+The **penalty factor** is a global multiplier (typically less than `1`, e.g., `0.8`) applied when the match type is less specific — such as wildcard or loose entity type matches. It allows the system to:
+- Break ties in favor of more precise blocks
+- Discourage overly generic blocks from being selected when better matches are available
--- a/api/src/chat/chat.module.ts
+++ b/api/src/chat/chat.module.ts
@ -16,6 +16,7 @@ import { AttachmentModel } from '@/attachment/schemas/attachment.schema';
 import { AttachmentService } from '@/attachment/services/attachment.service';
 import { ChannelModule } from '@/channel/channel.module';
 import { CmsModule } from '@/cms/cms.module';
+import { NlpModule } from '@/nlp/nlp.module';
 import { UserModule } from '@/user/user.module';

 import { BlockController } from './controllers/block.controller';
@ -68,6 +69,7 @@ import { SubscriberService } from './services/subscriber.service';
    AttachmentModule,
    EventEmitter2,
    UserModule,
+    NlpModule,
  ],
  controllers: [
    CategoryController,
--- a/api/src/chat/controllers/block.controller.spec.ts
+++ b/api/src/chat/controllers/block.controller.spec.ts
@ -20,6 +20,15 @@ import { LanguageRepository } from '@/i18n/repositories/language.repository';
 import { LanguageModel } from '@/i18n/schemas/language.schema';
 import { I18nService } from '@/i18n/services/i18n.service';
 import { LanguageService } from '@/i18n/services/language.service';
+import { LoggerService } from '@/logger/logger.service';
+import { NlpEntityRepository } from '@/nlp/repositories/nlp-entity.repository';
+import { NlpSampleEntityRepository } from '@/nlp/repositories/nlp-sample-entity.repository';
+import { NlpValueRepository } from '@/nlp/repositories/nlp-value.repository';
+import { NlpEntityModel } from '@/nlp/schemas/nlp-entity.schema';
+import { NlpSampleEntityModel } from '@/nlp/schemas/nlp-sample-entity.schema';
+import { NlpValueModel } from '@/nlp/schemas/nlp-value.schema';
+import { NlpEntityService } from '@/nlp/services/nlp-entity.service';
+import { NlpValueService } from '@/nlp/services/nlp-value.service';
 import { PluginService } from '@/plugins/plugins.service';
 import { SettingService } from '@/setting/services/setting.service';
 import { InvitationRepository } from '@/user/repositories/invitation.repository';
@ -93,6 +102,9 @@ describe('BlockController', () => {
          RoleModel,
          PermissionModel,
          LanguageModel,
+          NlpEntityModel,
+          NlpSampleEntityModel,
+          NlpValueModel,
        ]),
      ],
      providers: [
@ -116,6 +128,12 @@ describe('BlockController', () => {
        PermissionService,
        LanguageService,
        PluginService,
+        LoggerService,
+        NlpEntityService,
+        NlpEntityRepository,
+        NlpSampleEntityRepository,
+        NlpValueRepository,
+        NlpValueService,
        {
          provide: I18nService,
          useValue: {
--- a/api/src/chat/schemas/types/pattern.ts
+++ b/api/src/chat/schemas/types/pattern.ts
@ -8,6 +8,8 @@

 import { z } from 'zod';

+import { BlockFull } from '../block.schema';
+
 import { PayloadType } from './button';

 export const payloadPatternSchema = z.object({
@ -57,3 +59,19 @@ export const patternSchema = z.union([
 ]);

 export type Pattern = z.infer<typeof patternSchema>;
+
+export type NlpPatternMatchResult = {
+  block: BlockFull;
+  matchedPattern: NlpPattern[];
+};
+
+export function isNlpPattern(pattern: NlpPattern) {
+  return (
+    (typeof pattern === 'object' &&
+      pattern !== null &&
+      'entity' in pattern &&
+      'match' in pattern &&
+      pattern.match === 'entity') ||
+    pattern.match === 'value'
+  );
+}
--- a/api/src/chat/services/block.service.spec.ts
+++ b/api/src/chat/services/block.service.spec.ts
@ -31,6 +31,14 @@ import { LanguageRepository } from '@/i18n/repositories/language.repository';
 import { LanguageModel } from '@/i18n/schemas/language.schema';
 import { I18nService } from '@/i18n/services/i18n.service';
 import { LanguageService } from '@/i18n/services/language.service';
+import { NlpEntityRepository } from '@/nlp/repositories/nlp-entity.repository';
+import { NlpSampleEntityRepository } from '@/nlp/repositories/nlp-sample-entity.repository';
+import { NlpValueRepository } from '@/nlp/repositories/nlp-value.repository';
+import { NlpEntityModel } from '@/nlp/schemas/nlp-entity.schema';
+import { NlpSampleEntityModel } from '@/nlp/schemas/nlp-sample-entity.schema';
+import { NlpValueModel } from '@/nlp/schemas/nlp-value.schema';
+import { NlpEntityService } from '@/nlp/services/nlp-entity.service';
+import { NlpValueService } from '@/nlp/services/nlp-value.service';
 import { PluginService } from '@/plugins/plugins.service';
 import { SettingService } from '@/setting/services/setting.service';
 import {
@ -43,12 +51,23 @@ import {
  blockGetStarted,
  blockProductListMock,
  blocks,
+  mockModifiedNlpBlock,
+  mockModifiedNlpBlockOne,
+  mockModifiedNlpBlockTwo,
+  mockNlpBlock,
+  mockNlpPatternsSetOne,
+  mockNlpPatternsSetThree,
+  mockNlpPatternsSetTwo,
 } from '@/utils/test/mocks/block';
 import {
  contextBlankInstance,
  subscriberContextBlankInstance,
 } from '@/utils/test/mocks/conversation';
-import { nlpEntitiesGreeting } from '@/utils/test/mocks/nlp';
+import {
+  mockNlpCacheMap,
+  mockNlpEntitiesSetOne,
+  nlpEntitiesGreeting,
+} from '@/utils/test/mocks/nlp';
 import {
  closeInMongodConnection,
  rootMongooseTestModule,
@ -56,7 +75,7 @@ import {
 import { buildTestingMocks } from '@/utils/test/utils';

 import { BlockRepository } from '../repositories/block.repository';
-import { Block, BlockModel } from '../schemas/block.schema';
+import { Block, BlockFull, BlockModel } from '../schemas/block.schema';
 import { Category, CategoryModel } from '../schemas/category.schema';
 import { LabelModel } from '../schemas/label.schema';
 import { FileType } from '../schemas/types/attachment';
@ -75,6 +94,7 @@ describe('BlockService', () => {
  let hasPreviousBlocks: Block;
  let contentService: ContentService;
  let contentTypeService: ContentTypeService;
+  let nlpEntityService: NlpEntityService;

  beforeAll(async () => {
    const { getMocks } = await buildTestingMocks({
@ -91,6 +111,9 @@ describe('BlockService', () => {
          AttachmentModel,
          LabelModel,
          LanguageModel,
+          NlpEntityModel,
+          NlpSampleEntityModel,
+          NlpValueModel,
        ]),
      ],
      providers: [
@ -106,6 +129,14 @@ describe('BlockService', () => {
        ContentService,
        AttachmentService,
        LanguageService,
+        NlpEntityRepository,
+        NlpValueRepository,
+        NlpSampleEntityRepository,
+        NlpEntityService,
+        {
+          provide: NlpValueService,
+          useValue: {},
+        },
        {
          provide: PluginService,
          useValue: {},
@ -145,12 +176,14 @@ describe('BlockService', () => {
      contentTypeService,
      categoryRepository,
      blockRepository,
+      nlpEntityService,
    ] = await getMocks([
      BlockService,
      ContentService,
      ContentTypeService,
      CategoryRepository,
      BlockRepository,
+      NlpEntityService,
    ]);
    category = (await categoryRepository.findOne({ label: 'default' }))!;
    hasPreviousBlocks = (await blockRepository.findOne({
@ -291,6 +324,7 @@ describe('BlockService', () => {
        blockGetStarted,
      );
      expect(result).toEqual([
+        [
          {
            entity: 'intent',
            match: 'value',
@ -300,23 +334,153 @@ describe('BlockService', () => {
            entity: 'firstname',
            match: 'entity',
          },
+        ],
      ]);
    });

-    it('should return undefined when it does not match nlp patterns', () => {
+    it('should return empty array when it does not match nlp patterns', () => {
      const result = blockService.matchNLP(nlpEntitiesGreeting, {
        ...blockGetStarted,
        patterns: [[{ entity: 'lastname', match: 'value', value: 'Belakhel' }]],
      });
-      expect(result).toEqual(undefined);
+      expect(result).toEqual([]);
    });

-    it('should return undefined when unknown nlp patterns', () => {
+    it('should return empty array when unknown nlp patterns', () => {
      const result = blockService.matchNLP(nlpEntitiesGreeting, {
        ...blockGetStarted,
        patterns: [[{ entity: 'product', match: 'value', value: 'pizza' }]],
      });
-      expect(result).toEqual(undefined);
+      expect(result).toEqual([]);
+    });
+  });
+
+  describe('matchBestNLP', () => {
+    it('should return the block with the highest NLP score', async () => {
+      jest
+        .spyOn(nlpEntityService, 'getNlpMap')
+        .mockResolvedValue(mockNlpCacheMap);
+      const blocks = [mockNlpBlock, blockGetStarted]; // You can add more blocks with different patterns and scores
+      const nlp = mockNlpEntitiesSetOne;
+      // Spy on calculateBlockScore to check if it's called
+      const calculateBlockScoreSpy = jest.spyOn(
+        blockService,
+        'calculateBlockScore',
+      );
+      const bestBlock = await blockService.matchBestNLP(blocks, nlp);
+
+      // Ensure calculateBlockScore was called at least once for each block
+      expect(calculateBlockScoreSpy).toHaveBeenCalledTimes(2); // Called for each block
+
+      // Restore the spy after the test
+      calculateBlockScoreSpy.mockRestore();
+      // Assert that the block with the highest NLP score is selected
+      expect(bestBlock).toEqual(mockNlpBlock);
+    });
+
+    it('should return the block with the highest NLP score applying penalties', async () => {
+      jest
+        .spyOn(nlpEntityService, 'getNlpMap')
+        .mockResolvedValue(mockNlpCacheMap);
+      const blocks = [mockNlpBlock, mockModifiedNlpBlock]; // You can add more blocks with different patterns and scores
+      const nlp = mockNlpEntitiesSetOne;
+      // Spy on calculateBlockScore to check if it's called
+      const calculateBlockScoreSpy = jest.spyOn(
+        blockService,
+        'calculateBlockScore',
+      );
+      const bestBlock = await blockService.matchBestNLP(blocks, nlp);
+
+      // Ensure calculateBlockScore was called at least once for each block
+      expect(calculateBlockScoreSpy).toHaveBeenCalledTimes(2); // Called for each block
+
+      // Restore the spy after the test
+      calculateBlockScoreSpy.mockRestore();
+      // Assert that the block with the highest NLP score is selected
+      expect(bestBlock).toEqual(mockNlpBlock);
+    });
+
+    it('another case where it should return the block with the highest NLP score applying penalties', async () => {
+      jest
+        .spyOn(nlpEntityService, 'getNlpMap')
+        .mockResolvedValue(mockNlpCacheMap);
+      const blocks = [mockModifiedNlpBlockOne, mockModifiedNlpBlockTwo]; // You can add more blocks with different patterns and scores
+      const nlp = mockNlpEntitiesSetOne;
+      // Spy on calculateBlockScore to check if it's called
+      const calculateBlockScoreSpy = jest.spyOn(
+        blockService,
+        'calculateBlockScore',
+      );
+      const bestBlock = await blockService.matchBestNLP(blocks, nlp);
+
+      // Ensure calculateBlockScore was called at least once for each block
+      expect(calculateBlockScoreSpy).toHaveBeenCalledTimes(3); // Called for each block
+
+      // Restore the spy after the test
+      calculateBlockScoreSpy.mockRestore();
+      // Assert that the block with the highest NLP score is selected
+      expect(bestBlock).toEqual(mockModifiedNlpBlockTwo);
+    });
+
+    it('should return undefined if no blocks match or the list is empty', async () => {
+      jest
+        .spyOn(nlpEntityService, 'getNlpMap')
+        .mockResolvedValue(mockNlpCacheMap);
+      const blocks: BlockFull[] = []; // Empty block array
+      const nlp = mockNlpEntitiesSetOne;
+
+      const bestBlock = await blockService.matchBestNLP(blocks, nlp);
+
+      // Assert that undefined is returned when no blocks are available
+      expect(bestBlock).toBeUndefined();
+    });
+  });
+
+  describe('calculateBlockScore', () => {
+    it('should calculate the correct NLP score for a block', async () => {
+      jest
+        .spyOn(nlpEntityService, 'getNlpMap')
+        .mockResolvedValue(mockNlpCacheMap);
+      const score = await blockService.calculateBlockScore(
+        mockNlpPatternsSetOne,
+        mockNlpEntitiesSetOne,
+      );
+      const score2 = await blockService.calculateBlockScore(
+        mockNlpPatternsSetTwo,
+        mockNlpEntitiesSetOne,
+      );
+
+      expect(score).toBeGreaterThan(0);
+      expect(score2).toBe(0);
+      expect(score).toBeGreaterThan(score2);
+    });
+
+    it('should calculate the correct NLP score for a block and apply penalties ', async () => {
+      jest
+        .spyOn(nlpEntityService, 'getNlpMap')
+        .mockResolvedValue(mockNlpCacheMap);
+      const score = await blockService.calculateBlockScore(
+        mockNlpPatternsSetOne,
+        mockNlpEntitiesSetOne,
+      );
+      const score2 = await blockService.calculateBlockScore(
+        mockNlpPatternsSetThree,
+        mockNlpEntitiesSetOne,
+      );
+
+      expect(score).toBeGreaterThan(0);
+      expect(score2).toBeGreaterThan(0);
+      expect(score).toBeGreaterThan(score2);
+    });
+
+    it('should return 0 if no matching entities are found', async () => {
+      jest.spyOn(nlpEntityService, 'getNlpMap').mockResolvedValue(new Map());
+      const score = await blockService.calculateBlockScore(
+        mockNlpPatternsSetTwo,
+        mockNlpEntitiesSetOne,
+      );
+
+      expect(score).toBe(0); // No matching entity, so score should be 0
    });
  });

--- a/api/src/chat/services/block.service.ts
+++ b/api/src/chat/services/block.service.ts
@ -16,6 +16,8 @@ import { CONSOLE_CHANNEL_NAME } from '@/extensions/channels/console/settings';
 import { NLU } from '@/helper/types';
 import { I18nService } from '@/i18n/services/i18n.service';
 import { LanguageService } from '@/i18n/services/language.service';
+import { NlpCacheMapValues } from '@/nlp/schemas/types';
+import { NlpEntityService } from '@/nlp/services/nlp-entity.service';
 import { PluginService } from '@/plugins/plugins.service';
 import { PluginType } from '@/plugins/types';
 import { SettingService } from '@/setting/services/setting.service';
@ -53,6 +55,7 @@ export class BlockService extends BaseService<
    private readonly pluginService: PluginService,
    protected readonly i18n: I18nService,
    protected readonly languageService: LanguageService,
+    protected readonly entityService: NlpEntityService,
  ) {
    super(repository);
  }
@ -181,20 +184,21 @@ export class BlockService extends BaseService<
        .shift();

      // Perform an NLP Match
+
      if (!block && nlp) {
-        // Find block pattern having the best match of nlp entities
-        let nlpBest = 0;
-        filteredBlocks.forEach((b, index, self) => {
-          const nlpPattern = this.matchNLP(nlp, b);
-          if (nlpPattern && nlpPattern.length > nlpBest) {
-            nlpBest = nlpPattern.length;
-            block = self[index];
-          }
-        });
+        // Use the `reduce` function to iterate over `filteredBlocks` and accumulate a new array `matchesWithPatterns`.
+        // This approach combines the matching of NLP patterns and filtering of blocks with empty or invalid matches
+        // into a single operation. This avoids the need for a separate mapping and filtering step, improving performance.
+        // For each block in `filteredBlocks`, we call `matchNLP` to find patterns that match the NLP data.
+        // If `matchNLP` returns a non-empty list of matched patterns, the block and its matched patterns are added
+        // to the accumulator array `acc`, which is returned as the final result.
+        // This ensures that only blocks with valid matches are kept, and blocks with no matches are excluded,
+        // all while iterating through the list only once.
+
+        block = await this.matchBestNLP(filteredBlocks, nlp);
      }
    }
-    // Uknown event type => return false;
-    // this.logger.error('Unable to recognize event type while matching', event);
+
    return block;
  }

@ -304,7 +308,7 @@ export class BlockService extends BaseService<
  matchNLP(
    nlp: NLU.ParseEntities,
    block: Block | BlockFull,
-  ): NlpPattern[] | undefined {
+  ): NlpPattern[][] | undefined {
    // No nlp entities to check against
    if (nlp.entities.length === 0) {
      return undefined;
@ -313,14 +317,13 @@ export class BlockService extends BaseService<
    const nlpPatterns = block.patterns?.filter((p) => {
      return Array.isArray(p);
    }) as NlpPattern[][];
-
    // No nlp patterns found
    if (nlpPatterns.length === 0) {
      return undefined;
    }

    // Find NLP pattern match based on best guessed entities
-    return nlpPatterns.find((entities: NlpPattern[]) => {
+    return nlpPatterns.filter((entities: NlpPattern[]) => {
      return entities.every((ev: NlpPattern) => {
        if (ev.match === 'value') {
          return nlp.entities.find((e) => {
@ -338,6 +341,139 @@ export class BlockService extends BaseService<
    });
  }

+  /**
+   * Matches the provided NLU parsed entities with patterns in a set of blocks and returns
+   * the block with the highest matching score.
+   *
+   * For each block, it checks the patterns against the NLU parsed entities, calculates
+   * a score for each match, and selects the block with the highest score.
+   *
+   * @param {BlockFull[]} blocks - An array of BlockFull objects representing potential matches.
+   * @param {NLU.ParseEntities} nlp - The NLU parsed entities used for pattern matching.
+   *
+   * @returns {Promise<BlockFull | undefined>} - A promise that resolves to the BlockFull
+   * with the highest match score, or undefined if no matches are found.
+   */
+  async matchBestNLP(
+    blocks: BlockFull[],
+    nlp: NLU.ParseEntities,
+  ): Promise<BlockFull | undefined> {
+    const scoredBlocks = await Promise.all(
+      blocks.map(async (block) => {
+        const matchedPatterns = this.matchNLP(nlp, block) || [];
+
+        const scores = await Promise.all(
+          matchedPatterns.map((pattern) =>
+            this.calculateBlockScore(pattern, nlp),
+          ),
+        );
+
+        const maxScore = scores.length > 0 ? Math.max(...scores) : 0;
+
+        return { block, score: maxScore };
+      }),
+    );
+
+    const best = scoredBlocks.reduce(
+      (acc, curr) => (curr.score > acc.score ? curr : acc),
+      { block: undefined, score: 0 },
+    );
+
+    return best.block;
+  }
+
+  /**
+   * Computes the NLP score for a given block using its matched NLP patterns and parsed NLP entities.
+   *
+   * Each pattern is evaluated against the parsed NLP entities to determine matches based on entity name,
+   * value, and confidence. A score is computed using the entity's weight and the confidence level of the match.
+   * A penalty factor is optionally applied for entity-level matches to adjust the scoring.
+   *
+   * The function uses a cache (`nlpCacheMap`) to avoid redundant database lookups for entity metadata.
+   *
+   * @param patterns - The NLP patterns associated with the block.
+   * @param nlp - The parsed NLP entities from the user input.
+   * @param nlpCacheMap - A cache to reuse fetched entity metadata (e.g., weights and valid values).
+   * @param nlpPenaltyFactor - A multiplier applied to scores when the pattern match type is 'entity'.
+   * @returns A numeric score representing how well the block matches the given NLP context.
+   */
+  async calculateBlockScore(
+    patterns: NlpPattern[],
+    nlp: NLU.ParseEntities,
+  ): Promise<number> {
+    if (!patterns.length) return 0;
+
+    const nlpCacheMap = await this.entityService.getNlpMap();
+    // @TODO Make nluPenaltyFactor configurable in UI settings
+    const nluPenaltyFactor = 0.95;
+    // Compute individual pattern scores using the cache
+    const patternScores: number[] = patterns.map((pattern) => {
+      const entityData = nlpCacheMap.get(pattern.entity);
+      if (!entityData) return 0;
+
+      const matchedEntity: NLU.ParseEntity | undefined = nlp.entities.find(
+        (e) => this.matchesEntityData(e, pattern, entityData),
+      );
+
+      return this.computePatternScore(
+        matchedEntity,
+        pattern,
+        entityData,
+        nluPenaltyFactor,
+      );
+    });
+
+    // Sum the scores
+    return patternScores.reduce((sum, score) => sum + score, 0);
+  }
+
+  /**
+   * Checks if a given `ParseEntity` from the NLP model matches the specified pattern
+   * and if its value exists within the values provided in the cache for the specified entity.
+   *
+   * @param e - The `ParseEntity` object from the NLP model, containing information about the entity and its value.
+   * @param pattern - The `NlpPattern` object representing the entity and value pattern to be matched.
+   * @param entityData - The `NlpCacheMapValues` object containing cached data, including entity values and weight, for the entity being matched.
+   *
+   * @returns A boolean indicating whether the `ParseEntity` matches the pattern and entity data from the cache.
+   *
+   * - The function compares the entity type between the `ParseEntity` and the `NlpPattern`.
+   * - If the pattern's match type is not `'value'`, it checks if the entity's value is present in the cache's `values` array.
+   * - If the pattern's match type is `'value'`, it further ensures that the entity's value matches the specified value in the pattern.
+   * - Returns `true` if all conditions are met, otherwise `false`.
+   */
+  private matchesEntityData(
+    e: NLU.ParseEntity,
+    pattern: NlpPattern,
+    entityData: NlpCacheMapValues,
+  ): boolean {
+    return (
+      e.entity === pattern.entity &&
+      entityData?.values.some((v) => v === e.value) &&
+      (pattern.match !== 'value' || e.value === pattern.value)
+    );
+  }
+
+  /**
+   * Computes the score for a given entity based on its confidence, weight, and penalty factor.
+   *
+   * @param entity - The `ParseEntity` to check, which may be `undefined` if no match is found.
+   * @param pattern - The `NlpPattern` object that specifies how to match the entity and its value.
+   * @param entityData - The cached data for the given entity, including `weight` and `values`.
+   * @param nlpPenaltyFactor - The penalty factor applied when the pattern's match type is 'entity'.
+   * @returns The computed score based on the entity's confidence, the cached weight, and the penalty factor.
+   */
+  private computePatternScore(
+    entity: NLU.ParseEntity | undefined,
+    pattern: NlpPattern,
+    entityData: NlpCacheMapValues,
+    nlpPenaltyFactor: number,
+  ): number {
+    if (!entity || !entity.confidence) return 0;
+    const penalty = pattern.match === 'entity' ? nlpPenaltyFactor : 1;
+    return entity.confidence * entityData.weight * penalty;
+  }
+
  /**
   * Matches an outcome-based block from a list of available blocks
   * based on the outcome of a system message.
--- a/api/src/chat/services/bot.service.spec.ts
+++ b/api/src/chat/services/bot.service.spec.ts
@ -33,6 +33,14 @@ import { LanguageRepository } from '@/i18n/repositories/language.repository';
 import { LanguageModel } from '@/i18n/schemas/language.schema';
 import { I18nService } from '@/i18n/services/i18n.service';
 import { LanguageService } from '@/i18n/services/language.service';
+import { NlpEntityRepository } from '@/nlp/repositories/nlp-entity.repository';
+import { NlpSampleEntityRepository } from '@/nlp/repositories/nlp-sample-entity.repository';
+import { NlpValueRepository } from '@/nlp/repositories/nlp-value.repository';
+import { NlpEntityModel } from '@/nlp/schemas/nlp-entity.schema';
+import { NlpSampleEntityModel } from '@/nlp/schemas/nlp-sample-entity.schema';
+import { NlpValueModel } from '@/nlp/schemas/nlp-value.schema';
+import { NlpEntityService } from '@/nlp/services/nlp-entity.service';
+import { NlpValueService } from '@/nlp/services/nlp-value.service';
 import { PluginService } from '@/plugins/plugins.service';
 import { SettingService } from '@/setting/services/setting.service';
 import { installBlockFixtures } from '@/utils/test/fixtures/block';
@ -100,6 +108,9 @@ describe('BlockService', () => {
          MenuModel,
          ContextVarModel,
          LanguageModel,
+          NlpEntityModel,
+          NlpSampleEntityModel,
+          NlpValueModel,
        ]),
        JwtModule,
      ],
@ -131,6 +142,11 @@ describe('BlockService', () => {
        ContextVarService,
        ContextVarRepository,
        LanguageService,
+        NlpEntityService,
+        NlpEntityRepository,
+        NlpSampleEntityRepository,
+        NlpValueRepository,
+        NlpValueService,
        {
          provide: HelperService,
          useValue: {},
--- a/api/src/helper/lib/test/base-nlp-helper.spec.ts
+++ b/api/src/helper/lib/test/base-nlp-helper.spec.ts
@ -139,6 +139,7 @@ describe('BaseNlpHelper', () => {
          updatedAt: new Date(),
          builtin: false,
          lookups: [],
+          weight: 1,
        },
        entity2: {
          id: new ObjectId().toString(),
@ -147,6 +148,7 @@ describe('BaseNlpHelper', () => {
          updatedAt: new Date(),
          builtin: false,
          lookups: [],
+          weight: 1,
        },
      });
      jest.spyOn(NlpValue, 'getValueMap').mockReturnValue({
@ -207,6 +209,7 @@ describe('BaseNlpHelper', () => {
          updatedAt: new Date(),
          builtin: false,
          lookups: [],
+          weight: 1,
        },
      });

--- a/api/src/i18n/controllers/translation.controller.spec.ts
+++ b/api/src/i18n/controllers/translation.controller.spec.ts
@ -30,6 +30,14 @@ import { MenuModel } from '@/cms/schemas/menu.schema';
 import { ContentService } from '@/cms/services/content.service';
 import { MenuService } from '@/cms/services/menu.service';
 import { I18nService } from '@/i18n/services/i18n.service';
+import { NlpEntityRepository } from '@/nlp/repositories/nlp-entity.repository';
+import { NlpSampleEntityRepository } from '@/nlp/repositories/nlp-sample-entity.repository';
+import { NlpValueRepository } from '@/nlp/repositories/nlp-value.repository';
+import { NlpEntityModel } from '@/nlp/schemas/nlp-entity.schema';
+import { NlpSampleEntityModel } from '@/nlp/schemas/nlp-sample-entity.schema';
+import { NlpValueModel } from '@/nlp/schemas/nlp-value.schema';
+import { NlpEntityService } from '@/nlp/services/nlp-entity.service';
+import { NlpValueService } from '@/nlp/services/nlp-value.service';
 import { NlpService } from '@/nlp/services/nlp.service';
 import { PluginService } from '@/plugins/plugins.service';
 import { SettingService } from '@/setting/services/setting.service';
@ -75,6 +83,9 @@ describe('TranslationController', () => {
          BlockModel,
          ContentModel,
          LanguageModel,
+          NlpEntityModel,
+          NlpSampleEntityModel,
+          NlpValueModel,
        ]),
      ],
      providers: [
@ -130,6 +141,11 @@ describe('TranslationController', () => {
        },
        LanguageService,
        LanguageRepository,
+        NlpEntityRepository,
+        NlpEntityService,
+        NlpValueRepository,
+        NlpValueService,
+        NlpSampleEntityRepository,
      ],
    });
    [translationService, translationController] = await getMocks([
--- a/api/src/nlp/controllers/nlp-entity.controller.spec.ts
+++ b/api/src/nlp/controllers/nlp-entity.controller.spec.ts
@ -6,6 +6,7 @@
 * 2. All derivative works must include clear attribution to the original creator and software, Hexastack and Hexabot, in a prominent location (e.g., in the software's "About" section, documentation, and README file).
 */

+import { CACHE_MANAGER } from '@nestjs/cache-manager';
 import {
  BadRequestException,
  MethodNotAllowedException,
@ -67,6 +68,12 @@ describe('NlpEntityController', () => {
        NlpValueService,
        NlpSampleEntityRepository,
        NlpValueRepository,
+        {
+          provide: CACHE_MANAGER,
+          useValue: {
+            del: jest.fn(),
+          },
+        },
      ],
    });
    [nlpEntityController, nlpValueService, nlpEntityService] = await getMocks([
@ -109,6 +116,7 @@ describe('NlpEntityController', () => {
            ) as NlpEntityFull['values'],
            lookups: curr.lookups!,
            builtin: curr.builtin!,
+            weight: curr.weight!,
          });
          return acc;
        },
@ -163,6 +171,7 @@ describe('NlpEntityController', () => {
        name: 'sentiment',
        lookups: ['trait'],
        builtin: false,
+        weight: 1,
      };
      const result = await nlpEntityController.create(sentimentEntity);
      expect(result).toEqualPayload(sentimentEntity);
@ -214,6 +223,7 @@ describe('NlpEntityController', () => {
        updatedAt: firstNameEntity!.updatedAt,
        lookups: firstNameEntity!.lookups,
        builtin: firstNameEntity!.builtin,
+        weight: firstNameEntity!.weight,
      };
      const result = await nlpEntityController.findOne(firstNameEntity!.id, [
        'values',
@ -238,6 +248,7 @@ describe('NlpEntityController', () => {
        doc: '',
        lookups: ['trait'],
        builtin: false,
+        weight: 1,
      };
      const result = await nlpEntityController.updateOne(
        firstNameEntity!.id,
@ -258,7 +269,7 @@ describe('NlpEntityController', () => {
      ).rejects.toThrow(NotFoundException);
    });

-    it('should throw exception when nlp entity is builtin', async () => {
+    it('should throw an exception if entity is builtin but weight not provided', async () => {
      const updateNlpEntity: NlpEntityCreateDto = {
        name: 'updated',
        doc: '',
@ -269,6 +280,57 @@ describe('NlpEntityController', () => {
        nlpEntityController.updateOne(buitInEntityId!, updateNlpEntity),
      ).rejects.toThrow(MethodNotAllowedException);
    });
+
+    it('should update weight if entity is builtin and weight is provided', async () => {
+      const updatedNlpEntity: NlpEntityCreateDto = {
+        name: 'updated',
+        doc: '',
+        lookups: ['trait'],
+        builtin: false,
+        weight: 4,
+      };
+      const findOneSpy = jest.spyOn(nlpEntityService, 'findOne');
+      const updateWeightSpy = jest.spyOn(nlpEntityService, 'updateWeight');
+
+      const result = await nlpEntityController.updateOne(
+        buitInEntityId!,
+        updatedNlpEntity,
+      );
+
+      expect(findOneSpy).toHaveBeenCalledWith(buitInEntityId!);
+      expect(updateWeightSpy).toHaveBeenCalledWith(
+        buitInEntityId!,
+        updatedNlpEntity.weight,
+      );
+      expect(result.weight).toBe(updatedNlpEntity.weight);
+    });
+
+    it('should update only the weight of the builtin entity', async () => {
+      const updatedNlpEntity: NlpEntityCreateDto = {
+        name: 'updated',
+        doc: '',
+        lookups: ['trait'],
+        builtin: false,
+        weight: 4,
+      };
+      const originalEntity: NlpEntity | null = await nlpEntityService.findOne(
+        buitInEntityId!,
+      );
+
+      const result: NlpEntity = await nlpEntityController.updateOne(
+        buitInEntityId!,
+        updatedNlpEntity,
+      );
+
+      // Check weight is updated
+      expect(result.weight).toBe(updatedNlpEntity.weight);
+
+      Object.entries(originalEntity!).forEach(([key, value]) => {
+        if (key !== 'weight' && key !== 'updatedAt') {
+          expect(result[key as keyof typeof result]).toEqual(value);
+        }
+      });
+    });
  });
  describe('deleteMany', () => {
    it('should delete multiple nlp entities', async () => {
--- a/api/src/nlp/controllers/nlp-entity.controller.ts
+++ b/api/src/nlp/controllers/nlp-entity.controller.ts
@ -157,10 +157,19 @@ export class NlpEntityController extends BaseController<
      this.logger.warn(`Unable to update NLP Entity by id ${id}`);
      throw new NotFoundException(`NLP Entity with ID ${id} not found`);
    }
+
    if (nlpEntity.builtin) {
-      throw new MethodNotAllowedException(
-        `Cannot update builtin NLP Entity ${nlpEntity.name}`,
+      // Only allow weight update for builtin entities
+      if (updateNlpEntityDto.weight) {
+        return await this.nlpEntityService.updateWeight(
+          id,
+          updateNlpEntityDto.weight,
        );
+      } else {
+        throw new MethodNotAllowedException(
+          `Cannot update builtin NLP Entity ${nlpEntity.name} except for weight`,
+        );
+      }
    }

    return await this.nlpEntityService.updateOne(id, updateNlpEntityDto);
--- a/api/src/nlp/controllers/nlp-sample.controller.spec.ts
+++ b/api/src/nlp/controllers/nlp-sample.controller.spec.ts
@ -372,6 +372,7 @@ describe('NlpSampleController', () => {
        lookups: ['trait'],
        doc: '',
        builtin: false,
+        weight: 1,
      };
      const priceValueEntity = await nlpEntityService.findOne({
        name: 'intent',
--- a/api/src/nlp/controllers/nlp-value.controller.spec.ts
+++ b/api/src/nlp/controllers/nlp-value.controller.spec.ts
@ -6,6 +6,7 @@
 * 2. All derivative works must include clear attribution to the original creator and software, Hexastack and Hexabot, in a prominent location (e.g., in the software's "About" section, documentation, and README file).
 */

+import { CACHE_MANAGER } from '@nestjs/cache-manager';
 import { BadRequestException, NotFoundException } from '@nestjs/common';
 import { MongooseModule } from '@nestjs/mongoose';

@ -57,6 +58,12 @@ describe('NlpValueController', () => {
        NlpSampleEntityRepository,
        NlpEntityService,
        NlpEntityRepository,
+        {
+          provide: CACHE_MANAGER,
+          useValue: {
+            del: jest.fn(),
+          },
+        },
      ],
    });
    [nlpValueController, nlpValueService, nlpEntityService] = await getMocks([
--- a/api/src/nlp/dto/nlp-entity.dto.ts
+++ b/api/src/nlp/dto/nlp-entity.dto.ts
@ -1,5 +1,5 @@
 /*
- * Copyright © 2024 Hexastack. All rights reserved.
+ * Copyright © 2025 Hexastack. All rights reserved.
 *
 * Licensed under the GNU Affero General Public License v3.0 (AGPLv3) with the following additional terms:
 * 1. The name "Hexabot" is a trademark of Hexastack. You may not use this name in derivative works without express written permission.
@ -11,10 +11,13 @@ import {
  IsArray,
  IsBoolean,
  IsIn,
+  IsInt,
  IsNotEmpty,
+  IsNumber,
  IsOptional,
  IsString,
  Matches,
+  Min,
 } from 'class-validator';

 import { DtoConfig } from '@/utils/types/dto.types';
@ -47,6 +50,17 @@ export class NlpEntityCreateDto {
  @IsBoolean()
  @IsOptional()
  builtin?: boolean;
+
+  @ApiPropertyOptional({
+    description: 'Nlp entity associated weight for next block triggering',
+    type: Number,
+    minimum: 1,
+  })
+  @IsNumber()
+  @IsOptional()
+  @Min(1, { message: 'Weight must be a positive integer' })
+  @IsInt({ message: 'Weight must be an integer' })
+  weight?: number;
 }

 export type NlpEntityDto = DtoConfig<{
--- a/api/src/nlp/nlp.module.ts
+++ b/api/src/nlp/nlp.module.ts
@ -1,5 +1,5 @@
 /*
- * Copyright © 2024 Hexastack. All rights reserved.
+ * Copyright © 2025 Hexastack. All rights reserved.
 *
 * Licensed under the GNU Affero General Public License v3.0 (AGPLv3) with the following additional terms:
 * 1. The name "Hexabot" is a trademark of Hexastack. You may not use this name in derivative works without express written permission.
--- a/api/src/nlp/schemas/nlp-entity.schema.ts
+++ b/api/src/nlp/schemas/nlp-entity.schema.ts
@ -1,5 +1,5 @@
 /*
- * Copyright © 2024 Hexastack. All rights reserved.
+ * Copyright © 2025 Hexastack. All rights reserved.
 *
 * Licensed under the GNU Affero General Public License v3.0 (AGPLv3) with the following additional terms:
 * 1. The name "Hexabot" is a trademark of Hexastack. You may not use this name in derivative works without express written permission.
@ -58,6 +58,12 @@ export class NlpEntityStub extends BaseSchema {
  @Prop({ type: Boolean, default: false })
  builtin: boolean;

+  /**
+   * Entity's weight used to determine the next block to trigger in the conversational flow.
+   */
+  @Prop({ type: Number, default: 1, min: 0 })
+  weight: number;
+
  /**
   * Returns a map object for entities
   * @param entities - Array of entities
--- a/api/src/nlp/schemas/types.ts
+++ b/api/src/nlp/schemas/types.ts
@ -1,5 +1,5 @@
 /*
- * Copyright © 2024 Hexastack. All rights reserved.
+ * Copyright © 2025 Hexastack. All rights reserved.
 *
 * Licensed under the GNU Affero General Public License v3.0 (AGPLv3) with the following additional terms:
 * 1. The name "Hexabot" is a trademark of Hexastack. You may not use this name in derivative works without express written permission.
@ -25,3 +25,11 @@ export enum NlpSampleState {
  test = 'test',
  inbox = 'inbox',
 }
+
+export type NlpCacheMap = Map<string, NlpCacheMapValues>;
+
+export type NlpCacheMapValues = {
+  id: string;
+  weight: number;
+  values: string[];
+};
--- a/api/src/nlp/services/nlp-entity.service.spec.ts
+++ b/api/src/nlp/services/nlp-entity.service.spec.ts
@ -6,6 +6,7 @@
 * 2. All derivative works must include clear attribution to the original creator and software, Hexastack and Hexabot, in a prominent location (e.g., in the software's "About" section, documentation, and README file).
 */

+import { CACHE_MANAGER } from '@nestjs/cache-manager';
 import { MongooseModule } from '@nestjs/mongoose';

 import { nlpEntityFixtures } from '@/utils/test/fixtures/nlpentity';
@ -20,7 +21,11 @@ import { buildTestingMocks } from '@/utils/test/utils';
 import { NlpEntityRepository } from '../repositories/nlp-entity.repository';
 import { NlpSampleEntityRepository } from '../repositories/nlp-sample-entity.repository';
 import { NlpValueRepository } from '../repositories/nlp-value.repository';
-import { NlpEntity, NlpEntityModel } from '../schemas/nlp-entity.schema';
+import {
+  NlpEntity,
+  NlpEntityFull,
+  NlpEntityModel,
+} from '../schemas/nlp-entity.schema';
 import { NlpSampleEntityModel } from '../schemas/nlp-sample-entity.schema';
 import { NlpValueModel } from '../schemas/nlp-value.schema';

@ -48,6 +53,12 @@ describe('nlpEntityService', () => {
        NlpValueService,
        NlpValueRepository,
        NlpSampleEntityRepository,
+        {
+          provide: CACHE_MANAGER,
+          useValue: {
+            del: jest.fn(),
+          },
+        },
      ],
    });
    [nlpEntityService, nlpEntityRepository, nlpValueRepository] =
@ -117,6 +128,77 @@ describe('nlpEntityService', () => {
      expect(result).toEqualPayload(entitiesWithValues);
    });
  });
+  describe('NlpEntityService - updateWeight', () => {
+    let createdEntity: NlpEntity;
+    beforeEach(async () => {
+      createdEntity = await nlpEntityRepository.create({
+        name: 'testentity',
+        builtin: false,
+        weight: 3,
+      });
+    });
+
+    it('should update the weight of an NLP entity', async () => {
+      const newWeight = 8;
+
+      const updatedEntity = await nlpEntityService.updateWeight(
+        createdEntity.id,
+        newWeight,
+      );
+
+      expect(updatedEntity.weight).toBe(newWeight);
+    });
+
+    it('should handle updating weight of non-existent entity', async () => {
+      const nonExistentId = '507f1f77bcf86cd799439011'; // Example MongoDB ObjectId
+
+      try {
+        await nlpEntityService.updateWeight(nonExistentId, 5);
+        fail('Expected error was not thrown');
+      } catch (error) {
+        expect(error).toBeDefined();
+      }
+    });
+
+    it('should use default weight of 1 when creating entity without weight', async () => {
+      const createdEntity = await nlpEntityRepository.create({
+        name: 'entityWithoutWeight',
+        builtin: true,
+        // weight not specified
+      });
+
+      expect(createdEntity.weight).toBe(1);
+    });
+
+    it('should throw an error if weight is less than 1', async () => {
+      const invalidWeight = 0;
+
+      await expect(
+        nlpEntityService.updateWeight(createdEntity.id, invalidWeight),
+      ).rejects.toThrow('Weight must be a positive integer');
+    });
+
+    it('should throw an error if weight is a decimal', async () => {
+      const invalidWeight = 2.5;
+
+      await expect(
+        nlpEntityService.updateWeight(createdEntity.id, invalidWeight),
+      ).rejects.toThrow('Weight must be a positive integer');
+    });
+
+    it('should throw an error if weight is negative', async () => {
+      const invalidWeight = -3;
+
+      await expect(
+        nlpEntityService.updateWeight(createdEntity.id, invalidWeight),
+      ).rejects.toThrow('Weight must be a positive integer');
+    });
+
+    afterEach(async () => {
+      // Clean the collection after each test
+      await nlpEntityRepository.deleteOne(createdEntity.id);
+    });
+  });

  describe('storeNewEntities', () => {
    it('should store new entities', async () => {
@ -150,4 +232,47 @@ describe('nlpEntityService', () => {
      expect(result).toEqualPayload(storedEntites);
    });
  });
+  describe('getNlpMap', () => {
+    it('should return a NlpCacheMap with the correct structure', async () => {
+      // Arrange
+      const firstMockValues = {
+        id: '1',
+        weight: 1,
+      };
+      const firstMockEntity = {
+        name: 'intent',
+        ...firstMockValues,
+        values: [{ value: 'buy' }, { value: 'sell' }],
+      } as unknown as Partial<NlpEntityFull>;
+      const secondMockValues = {
+        id: '2',
+        weight: 5,
+      };
+      const secondMockEntity = {
+        name: 'subject',
+        ...secondMockValues,
+        values: [{ value: 'product' }],
+      } as unknown as Partial<NlpEntityFull>;
+      const mockEntities = [firstMockEntity, secondMockEntity];
+
+      // Mock findAndPopulate
+      jest
+        .spyOn(nlpEntityService, 'findAllAndPopulate')
+        .mockResolvedValue(mockEntities as unknown as NlpEntityFull[]);
+
+      // Act
+      const result = await nlpEntityService.getNlpMap();
+
+      expect(result).toBeInstanceOf(Map);
+      expect(result.size).toBe(2);
+      expect(result.get('intent')).toEqual({
+        name: 'intent',
+        ...firstMockEntity,
+      });
+      expect(result.get('subject')).toEqual({
+        name: 'subject',
+        ...secondMockEntity,
+      });
+    });
+  });
 });
--- a/api/src/nlp/services/nlp-entity.service.ts
+++ b/api/src/nlp/services/nlp-entity.service.ts
@ -1,13 +1,18 @@
 /*
- * Copyright © 2024 Hexastack. All rights reserved.
+ * Copyright © 2025 Hexastack. All rights reserved.
 *
 * Licensed under the GNU Affero General Public License v3.0 (AGPLv3) with the following additional terms:
 * 1. The name "Hexabot" is a trademark of Hexastack. You may not use this name in derivative works without express written permission.
 * 2. All derivative works must include clear attribution to the original creator and software, Hexastack and Hexabot, in a prominent location (e.g., in the software's "About" section, documentation, and README file).
 */

-import { Injectable } from '@nestjs/common';
+import { CACHE_MANAGER } from '@nestjs/cache-manager';
+import { Inject, Injectable } from '@nestjs/common';
+import { OnEvent } from '@nestjs/event-emitter';
+import { Cache } from 'cache-manager';

+import { NLP_MAP_CACHE_KEY } from '@/utils/constants/cache';
+import { Cacheable } from '@/utils/decorators/cacheable.decorator';
 import { BaseService } from '@/utils/generics/base-service';

 import { Lookup, NlpEntityDto } from '../dto/nlp-entity.dto';
@ -17,7 +22,7 @@ import {
  NlpEntityFull,
  NlpEntityPopulate,
 } from '../schemas/nlp-entity.schema';
-import { NlpSampleEntityValue } from '../schemas/types';
+import { NlpCacheMap, NlpSampleEntityValue } from '../schemas/types';

 import { NlpValueService } from './nlp-value.service';

@ -30,6 +35,7 @@ export class NlpEntityService extends BaseService<
 > {
  constructor(
    readonly repository: NlpEntityRepository,
+    @Inject(CACHE_MANAGER) private readonly cacheManager: Cache,
    private readonly nlpValueService: NlpValueService,
  ) {
    super(repository);
@ -46,6 +52,28 @@ export class NlpEntityService extends BaseService<
    return await this.repository.deleteOne(id);
  }

+  /**
+   * Updates the `weight` field of a specific NLP entity by its ID.
+   *
+   * This method is part of the NLP-based blocks prioritization strategy.
+   * The weight influences the scoring of blocks when multiple blocks match a user's input.
+   * @param id - The unique identifier of the entity to update.
+   * @param updatedWeight - The new weight to assign. Must be a positive integer.
+   * @throws Error if the weight is not a positive integer.
+   * @returns A promise that resolves to the updated entity.
+   */
+  async updateWeight(id: string, updatedWeight: number): Promise<NlpEntity> {
+    if (!Number.isInteger(updatedWeight) || updatedWeight < 1) {
+      throw new Error('Weight must be a positive integer');
+    }
+
+    return await this.repository.updateOne(
+      id,
+      { weight: updatedWeight },
+      { new: true },
+    );
+  }
+
  /**
   * Stores new entities based on the sample text and sample entities.
   * Deletes all values relative to this entity before deleting the entity itself.
@ -97,4 +125,49 @@ export class NlpEntityService extends BaseService<
    );
    return Promise.all(findOrCreate);
  }
+
+  /**
+   * Clears the NLP map cache
+   */
+  async clearCache() {
+    await this.cacheManager.del(NLP_MAP_CACHE_KEY);
+  }
+
+  /**
+   * Event handler for Nlp Entity updates. Listens to 'hook:nlpEntity:*' events
+   * and invalidates the cache for nlp entities when triggered.
+   */
+  @OnEvent('hook:nlpEntity:*')
+  async handleNlpEntityUpdateEvent() {
+    this.clearCache();
+  }
+
+  /**
+   * Event handler for Nlp Value updates. Listens to 'hook:nlpValue:*' events
+   * and invalidates the cache for nlp values when triggered.
+   */
+  @OnEvent('hook:nlpValue:*')
+  async handleNlpValueUpdateEvent() {
+    this.clearCache();
+  }
+
+  /**
+   * Retrieves NLP entity lookup information for the given list of entity names.
+   *
+   * This method queries the database for lookups that match any of the provided
+   * entity names, transforms the result into a map structure where each key is
+   * the entity name and each value contains metadata (id, weight, and list of values),
+   * and caches the result using the configured cache key.
+   *
+   * @param entityNames - Array of entity names to retrieve lookup data for.
+   * @returns A Promise that resolves to a map of entity name to its corresponding lookup metadata.
+   */
+  @Cacheable(NLP_MAP_CACHE_KEY)
+  async getNlpMap(): Promise<NlpCacheMap> {
+    const entities = await this.findAllAndPopulate();
+    return entities.reduce((acc, curr) => {
+      acc.set(curr.name, curr);
+      return acc;
+    }, new Map());
+  }
 }
--- a/api/src/nlp/services/nlp-sample-entity.service.spec.ts
+++ b/api/src/nlp/services/nlp-sample-entity.service.spec.ts
@ -6,6 +6,7 @@
 * 2. All derivative works must include clear attribution to the original creator and software, Hexastack and Hexabot, in a prominent location (e.g., in the software's "About" section, documentation, and README file).
 */

+import { CACHE_MANAGER } from '@nestjs/cache-manager';
 import { MongooseModule } from '@nestjs/mongoose';

 import { LanguageRepository } from '@/i18n/repositories/language.repository';
@ -76,6 +77,12 @@ describe('NlpSampleEntityService', () => {
        NlpSampleEntityService,
        NlpEntityService,
        NlpValueService,
+        {
+          provide: CACHE_MANAGER,
+          useValue: {
+            del: jest.fn(),
+          },
+        },
      ],
    });
    [
--- a/api/src/nlp/services/nlp-value.service.spec.ts
+++ b/api/src/nlp/services/nlp-value.service.spec.ts
@ -6,6 +6,7 @@
 * 2. All derivative works must include clear attribution to the original creator and software, Hexastack and Hexabot, in a prominent location (e.g., in the software's "About" section, documentation, and README file).
 */

+import { CACHE_MANAGER } from '@nestjs/cache-manager';
 import { MongooseModule } from '@nestjs/mongoose';

 import { BaseSchema } from '@/utils/generics/base-schema';
@ -58,6 +59,12 @@ describe('NlpValueService', () => {
        NlpEntityRepository,
        NlpValueService,
        NlpEntityService,
+        {
+          provide: CACHE_MANAGER,
+          useValue: {
+            del: jest.fn(),
+          },
+        },
      ],
    });
    [
--- a/api/src/utils/constants/cache.ts
+++ b/api/src/utils/constants/cache.ts
@ -18,3 +18,5 @@ export const LANGUAGES_CACHE_KEY = 'languages';
 export const DEFAULT_LANGUAGE_CACHE_KEY = 'default_language';

 export const ALLOWED_ORIGINS_CACHE_KEY = 'allowed_origins';
+
+export const NLP_MAP_CACHE_KEY = 'nlp_map';
--- a/api/src/utils/test/fixtures/nlpentity.ts
+++ b/api/src/utils/test/fixtures/nlpentity.ts
@ -1,5 +1,5 @@
 /*
- * Copyright © 2024 Hexastack. All rights reserved.
+ * Copyright © 2025 Hexastack. All rights reserved.
 *
 * Licensed under the GNU Affero General Public License v3.0 (AGPLv3) with the following additional terms:
 * 1. The name "Hexabot" is a trademark of Hexastack. You may not use this name in derivative works without express written permission.
@ -17,18 +17,21 @@ export const nlpEntityFixtures: NlpEntityCreateDto[] = [
    lookups: ['trait'],
    doc: '',
    builtin: false,
+    weight: 1,
  },
  {
    name: 'first_name',
    lookups: ['keywords'],
    doc: '',
    builtin: false,
+    weight: 1,
  },
  {
    name: 'built_in',
    lookups: ['trait'],
    doc: '',
    builtin: true,
+    weight: 1,
  },
 ];

--- a/api/src/utils/test/mocks/block.ts
+++ b/api/src/utils/test/mocks/block.ts
@ -16,7 +16,7 @@ import { ButtonType, PayloadType } from '@/chat/schemas/types/button';
 import { CaptureVar } from '@/chat/schemas/types/capture-var';
 import { OutgoingMessageFormat } from '@/chat/schemas/types/message';
 import { BlockOptions, ContentOptions } from '@/chat/schemas/types/options';
-import { Pattern } from '@/chat/schemas/types/pattern';
+import { NlpPattern, Pattern } from '@/chat/schemas/types/pattern';
 import { QuickReplyType } from '@/chat/schemas/types/quick-reply';

 import { modelInstance } from './misc';
@ -246,6 +246,121 @@ export const blockGetStarted = {
  message: ['Welcome! How are you ? '],
 } as unknown as BlockFull;

+export const mockNlpPatternsSetOne: NlpPattern[] = [
+  {
+    entity: 'intent',
+    match: 'value',
+    value: 'greeting',
+  },
+  {
+    entity: 'firstname',
+    match: 'value',
+    value: 'jhon',
+  },
+];
+
+export const mockNlpPatternsSetTwo: NlpPattern[] = [
+  {
+    entity: 'intent',
+    match: 'value',
+    value: 'affirmation',
+  },
+  {
+    entity: 'firstname',
+    match: 'value',
+    value: 'mark',
+  },
+];
+
+export const mockNlpPatternsSetThree: NlpPattern[] = [
+  {
+    entity: 'intent',
+    match: 'value',
+    value: 'greeting',
+  },
+  {
+    entity: 'firstname',
+    match: 'entity',
+  },
+];
+
+export const mockNlpBlock: BlockFull = {
+  ...baseBlockInstance,
+  name: 'Mock Nlp',
+  patterns: [
+    'Hello',
+    '/we*lcome/',
+    { label: 'Mock Nlp', value: 'MOCK_NLP' },
+
+    mockNlpPatternsSetOne,
+    [
+      {
+        entity: 'intent',
+        match: 'value',
+        value: 'greeting',
+      },
+      {
+        entity: 'firstname',
+        match: 'value',
+        value: 'doe',
+      },
+    ],
+  ],
+
+  trigger_labels: customerLabelsMock,
+  message: ['Good to see you again '],
+} as unknown as BlockFull;
+
+export const mockModifiedNlpBlock: BlockFull = {
+  ...baseBlockInstance,
+  name: 'Modified Mock Nlp',
+  patterns: [
+    'Hello',
+    '/we*lcome/',
+    { label: 'Modified Mock Nlp', value: 'MODIFIED_MOCK_NLP' },
+    mockNlpPatternsSetThree,
+  ],
+  trigger_labels: customerLabelsMock,
+  message: ['Hello there'],
+} as unknown as BlockFull;
+
+export const mockModifiedNlpBlockOne: BlockFull = {
+  ...baseBlockInstance,
+  name: 'Modified Mock Nlp One',
+  patterns: [
+    'Hello',
+    '/we*lcome/',
+    { label: 'Modified Mock Nlp One', value: 'MODIFIED_MOCK_NLP_ONE' },
+    mockNlpPatternsSetTwo,
+    [
+      {
+        entity: 'firstname',
+        match: 'entity',
+      },
+    ],
+  ],
+  trigger_labels: customerLabelsMock,
+  message: ['Hello Sir'],
+} as unknown as BlockFull;
+
+export const mockModifiedNlpBlockTwo: BlockFull = {
+  ...baseBlockInstance,
+  name: 'Modified Mock Nlp Two',
+  patterns: [
+    'Hello',
+    '/we*lcome/',
+    { label: 'Modified Mock Nlp Two', value: 'MODIFIED_MOCK_NLP_TWO' },
+    [
+      {
+        entity: 'firstname',
+        match: 'entity',
+      },
+    ],
+    mockNlpPatternsSetThree,
+  ],
+  trigger_labels: customerLabelsMock,
+  message: ['Hello Madam'],
+} as unknown as BlockFull;
 const patternsProduct: Pattern[] = [
  'produit',
  [
@ -285,3 +400,5 @@ export const blockCarouselMock = {
 } as unknown as BlockFull;

 export const blocks: BlockFull[] = [blockGetStarted, blockEmpty];
+
+export const nlpBlocks: BlockFull[] = [blockGetStarted, mockNlpBlock];
--- a/api/src/utils/test/mocks/nlp.ts
+++ b/api/src/utils/test/mocks/nlp.ts
@ -1,5 +1,5 @@
 /*
- * Copyright © 2024 Hexastack. All rights reserved.
+ * Copyright © 2025 Hexastack. All rights reserved.
 *
 * Licensed under the GNU Affero General Public License v3.0 (AGPLv3) with the following additional terms:
 * 1. The name "Hexabot" is a trademark of Hexastack. You may not use this name in derivative works without express written permission.
@ -7,6 +7,7 @@
 */

 import { NLU } from '@/helper/types';
+import { NlpCacheMap } from '@/nlp/schemas/types';

 export const nlpEntitiesGreeting: NLU.ParseEntities = {
  entities: [
@ -27,3 +28,52 @@ export const nlpEntitiesGreeting: NLU.ParseEntities = {
    },
  ],
 };
+
+export const mockNlpEntitiesSetOne: NLU.ParseEntities = {
+  entities: [
+    {
+      entity: 'intent',
+      value: 'greeting',
+      confidence: 0.999,
+    },
+    {
+      entity: 'firstname',
+      value: 'jhon',
+      confidence: 0.5,
+    },
+  ],
+};
+
+export const mockNlpEntitiesSetTwo: NLU.ParseEntities = {
+  entities: [
+    {
+      entity: 'intent',
+      value: 'greeting',
+      confidence: 0.94,
+    },
+    {
+      entity: 'firstname',
+      value: 'doe',
+      confidence: 0.33,
+    },
+  ],
+};
+
+export const mockNlpCacheMap: NlpCacheMap = new Map([
+  [
+    'intent',
+    {
+      id: '67e3e41eff551ca5be70559c',
+      weight: 1,
+      values: ['greeting', 'affirmation'],
+    },
+  ],
+  [
+    'firstname',
+    {
+      id: '67e3e41eff551ca5be70559d',
+      weight: 1,
+      values: ['jhon', 'doe'],
+    },
+  ],
+]);
--- a/frontend/public/locales/en/translation.json
+++ b/frontend/public/locales/en/translation.json
@ -121,7 +121,9 @@
    "file_error": "File not found",
    "audio_error": "Audio not found",
    "video_error": "Video not found",
-    "missing_fields_error": "Please make sure that all required fields are filled"
+    "missing_fields_error": "Please make sure that all required fields are filled",
+    "weight_required_error": "Weight is required or invalid",
+    "weight_positive_integer_error": "Weight must be a positive integer" 
  },
  "menu": {
    "terms": "Terms of Use",
@ -348,6 +350,7 @@
    "nlp_lookup_trait": "Trait",
    "doc": "Documentation",
    "builtin": "Built-in?",
+    "weight": "Weight",
    "dataset": "Dataset",
    "yes": "Yes",
    "no": "No",
--- a/frontend/public/locales/fr/translation.json
+++ b/frontend/public/locales/fr/translation.json
@ -121,7 +121,9 @@
    "file_error": "Fichier introuvable",
    "audio_error": "Audio introuvable",
    "video_error": "Vidéo introuvable",
-    "missing_fields_error": "Veuillez vous assurer que tous les champs sont remplis correctement"
+    "missing_fields_error": "Veuillez vous assurer que tous les champs sont remplis correctement",
+    "weight_positive_integer_error": "Le poids doit être un nombre entier positif",
+    "weight_required_error": "Le poids est requis ou bien invalide"
  },
  "menu": {
    "terms": "Conditions d'utilisation",
@ -347,6 +349,7 @@
    "nlp_lookup_trait": "Trait",
    "synonyms": "Synonymes",
    "doc": "Documentation",
+    "weight": "Poids",
    "builtin": "Intégré?",
    "dataset": "Données",
    "yes": "Oui",
--- a/frontend/src/app-components/tables/columns/getColumns.tsx
+++ b/frontend/src/app-components/tables/columns/getColumns.tsx
@ -156,8 +156,7 @@ function StackComponent<T extends GridValidRowModel>({
            disabled={
              (isDisabled && isDisabled(params.row)) ||
              (params.row.builtin &&
-                (requires.includes(PermissionAction.UPDATE) ||
-                  requires.includes(PermissionAction.DELETE)))
+                  requires.includes(PermissionAction.DELETE))
            }
            onClick={() => {
              action && action(params.row);
--- a/frontend/src/components/nlp/components/NlpEntity.tsx
+++ b/frontend/src/components/nlp/components/NlpEntity.tsx
@ -167,6 +167,16 @@ const NlpEntity = () => {
      resizable: false,
      renderHeader,
    },
+    {
+      maxWidth: 210,
+      field: "weight",
+      headerName: t("label.weight"),
+      renderCell: (val) => <Chip label={val.value} variant="title" />,
+      sortable: true,
+      disableColumnMenu: true,
+      resizable: false,
+      renderHeader,
+    },
    {
      maxWidth: 90,
      field: "builtin",
--- a/frontend/src/components/nlp/components/NlpEntityForm.tsx
+++ b/frontend/src/components/nlp/components/NlpEntityForm.tsx
@ -60,6 +60,7 @@ export const NlpEntityVarForm: FC<ComponentFormProps<INlpEntity>> = ({
      name: nlpEntity?.name || "",
      doc: nlpEntity?.doc || "",
      lookups: nlpEntity?.lookups || ["keywords"],
+      weight: nlpEntity?.weight || 1,
    },
  });
  const validationRules = {
@ -82,6 +83,7 @@ export const NlpEntityVarForm: FC<ComponentFormProps<INlpEntity>> = ({
      reset({
        name: nlpEntity.name,
        doc: nlpEntity.doc,
+        weight: nlpEntity.weight,
      });
    } else {
      reset();
@ -121,6 +123,7 @@ export const NlpEntityVarForm: FC<ComponentFormProps<INlpEntity>> = ({
              required
              autoFocus
              helperText={errors.name ? errors.name.message : null}
+              disabled={nlpEntity?.builtin}
            />
          </ContentItem>
          <ContentItem>
@ -128,6 +131,33 @@ export const NlpEntityVarForm: FC<ComponentFormProps<INlpEntity>> = ({
              label={t("label.doc")}
              {...register("doc")}
              multiline={true}
+              disabled={nlpEntity?.builtin}
+            />
+          </ContentItem>
+          <ContentItem>
+            <Input
+              label={t("label.weight")}
+              {...register("weight", {
+                valueAsNumber: true,
+                required: t("message.weight_required_error"),
+                min: {
+                  value: 1,
+                  message: t("message.weight_positive_integer_error"),
+                },
+                validate: (value) =>
+                  value && Number.isInteger(value) && value! > 0
+                    ? true
+                    : t("message.weight_positive_integer_error"),
+              })}
+              type="number"
+              inputProps={{
+                min: 1,
+                step: 1,
+                inputMode: "numeric",
+                pattern: "[1-9][0-9]*",
+              }}
+              error={!!errors.weight}
+              helperText={errors.weight?.message}
          />
        </ContentItem>
        </ContentContainer>
--- a/frontend/src/types/nlp-entity.types.ts
+++ b/frontend/src/types/nlp-entity.types.ts
@ -1,5 +1,5 @@
 /*
- * Copyright © 2024 Hexastack. All rights reserved.
+ * Copyright © 2025 Hexastack. All rights reserved.
 *
 * Licensed under the GNU Affero General Public License v3.0 (AGPLv3) with the following additional terms:
 * 1. The name "Hexabot" is a trademark of Hexastack. You may not use this name in derivative works without express written permission.
@ -19,6 +19,7 @@ export interface INlpEntityAttributes {
  lookups: Lookup[];
  doc?: string;
  builtin?: boolean;
+  weight?: number;
 }

 export enum NlpLookups {