ABCDEFGHIJKLMNOPQRSTUVWXYZAAABACADAEAFAGAHAIAJ
1
2
3
RASA PROJECT PLAN TEMPLATE
4
5
6
* = an automatically calculated cell
7
TASK NAMESTART DATEDAY OF MONTH*END DATEDURATION* (WORK DAYS)DAYS COMPLETE*DAYS REMAINING*RESPONSIBLE ROLEPERCENT COMPLETEWEEK 1WEEK 2WEEK 3WEEK 4
8
9
1. Concept and Design
10
1.1 Explore and collect existing data for use cases1/111/870.76.3Product/Content10%
11
1.2 Select initial use cases1/111/875.61.4Product80%
12
1.3 Design the conversations around each use case1/881/1242.41.6Product/Content60%
13
1.3.1 Write sample conversations1/881/1241.62.4Content40%
14
1.3.2 Identify domain items (intents, entities, etc.) within sample conversations1/881/1240.83.2Content/Dev20%
15
1.4 Determine architecture and hosting environment1/881/1461.24.8Dev20%
16
1.4.1 Channel integrations1/991/1340.83.2Dev20%
17
1.4.2 Hosting infrastructure1/991/1340.83.2Dev20%
18
1.4.3 Deployment methods1/12121/1420.41.6Dev20%
19
2. Installation and Integration
20
2.1 Set up the Development Environment1/15151/24990Dev100%
21
2.2.1 Rasa Open Source/Rasa X installation1/15151/1943.20.8Dev80%
22
2.1.2 Network interfaces & SSL1/17171/2142.41.6Dev60%
23
2.1.3 Databases for storing user conversations1/17171/2472.84.2Dev40%
24
2.1.4 SSO configuration1/17171/2472.84.2Dev40%
25
2.1.5 Git repository for managing training data1/18181/2462.43.6Dev40%
26
2.1.6 Connect messaging channel(s)1/18181/2462.43.6Dev40%
27
2.1.7 CI/CD pipeline1/18181/2462.43.6Dev40%
28
2.2 Integrate with existing backend systems1/15151/1831.21.8Dev40%
29
2.2.1 Define methods for connecting to backend API services1/15151/1830.62.4Dev20%
30
2.2.2 Build custom action server1/15151/1610.20.8Dev20%
31
3. Domain Creation
32
3.1 Intents - what your assistant can understand1/12121/1970.76.3Content10%
33
3.2 Entities - pieces of information your assistant recognizes1/12121/1970.76.3Content10%
34
3.3 Slots - information that your assistant remembers about the user1/12121/1970.76.3Content/Dev10%
35
3.4 Actions/responses - things your assistant can say and do1/12121/1970.76.3Content/Dev10%
36
3.5 Forms - information collection tasks your assistant can perform1/12121/1970.76.3Content/Dev10%
37
4. Cold Start
38
4.1 Create an initial NLU training data set1/18181/2130.32.7Content10%
39
4.1.2 Source from historical user data, if available1/18181/2130.32.7Content10%
40
4.1.3 Create your own training examples, if no data available - 20 examples per intent1/18181/2130.62.4Content20%
41
4.2 Create stories to train dialogue engine1/21211/2320.41.6Content20%
42
4.3 Deploy MVA to development environment1/21211/2320.41.6Dev20%
43
5. Internal Training
44
5.1 Collect conversation data through internal testing and crowdsourcing1/25251/2720.21.8Product10%
45
5.1.1 Share assistant with internal testers1/25251/2720.21.8Product10%
46
5.1.2 Analyze data to understand where users are diverging from the happy paths1/26261/2820.21.8Product10%
47
5.1.3 Use the data to make annotations, fix bugs, re-train models1/26261/2820.21.8Content/Dev10%
48
5.2 Create test cases from collected conversations1/26261/2820.21.8Content/Dev10%
49
5.2.1 Integrate tests into CI/CD pipeline1/26261/2820.21.8Dev10%
50
6. Production Training
51
6.1 Roll out the AI assistant to a production environment1/31312/1212012Dev0%
52
6.1.1 Use load testing and performance monitoring to evaluate compute resources1/31312/1212012Dev0%
53
6.1.2 Limited rollout1/31312/3303Dev0%
54
6.1.3 Increase traffic as assistant improves2/332/12909Dev0%
55
6.2 Make a plan for team to review incoming messages (weekly basis)2/332/5202Product/Content0%
56
6.2.1 Annotate user messages and incorporate into training data2/332/12909Product/Content0%
57
6.2.2 Analyze user conversations, use to inform new feature development2/332/12909Product0%
58
6.2.3 Re-train and test model2/332/12909Dev0%
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100